Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Aggregating happiness: seeking and identifying a single plausible unifying theory
(USC Thesis Other)
Aggregating happiness: seeking and identifying a single plausible unifying theory
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Author: Michael Pressman Dissertation title: Aggregating Happiness: Seeking and Identifying a Single Plausible Unifying Theory Department: Philosophy Degree being conferred: Ph.D. University: University of Southern California Committee: Ralph Wedgwood (chair), Jake Ross, Jon Quong, Greg Keating Date of dissertation defense: August 31, 2018 Date of submission of manuscript to university: September 7, 2018 Degree conferral date: December 2018 Table of Contents: Executive Summary of the Dissertation page 1 Part I: Introduction, preliminary points, the theoretical problem and the landscape of the literature, the broad importance of the puzzle, practical implications, and a prima facie argument in favor of AU page 8 Part I Chapter 1: Introduction, preliminary points, the theoretical problem and the landscape of the literature, the broad importance of the puzzle, and practical implications page 9 Part I Chapter 2: The prima facie appeal of AU page 54 Part II: Reasons to believe that AMU is the most plausible version of AU page 82 Part II Chapter 1: A positive argument for why AMU is the most plausible version of AU (and for why the conscious moment is the “unit relevant for ethical inquiry”) page 85 Part II Chapter 2: Considering counterarguments: AMU and the Intrapersonal Maximizing Intuition (IMI) page 152 Part II Chapter 3 (only a chapter summary): The debate about the badness of death for the person who dies, and the underlying accounts of what is good for a person page 201 Part III: A defense of average utilitarianism (the umbrella theory): Handling three main objections to AU (as well as answering some additional objections and providing additional points and discussions) page 221 Part III Chapter 1: Parfit’s bottom up objection, a defense of AU with a temporally- neutral reference class, and an explanation of intuitions to the contrary page 224 Part III Chapter 2: Two Hells page 303 Part III Chapter 3: The Top Down Objection to AU page 319 Part IV: Zero points of value and how they play in to the AU / TU Debate page 419 Part IV Chapter 1: Introduction to zero points page 423 Part IV Chapter 2: A more detailed explanation of what a zero point is, and a preview of the exploration of the various frameworks through which the notion of a zero point can be understood and will be explored in Part IV page 451 Part IV Chapter 3 (only a chapter summary): The two ways of interpreting the claim that goodness is reducible to betterness, and why only one of the two should be focused on by the reductionist page 466 Part IV Chapter 4: Why Broome is not clearly right that a property that’s reducible to its comparative does not have absolute degrees (or, thus, a zero point): Laying the foundation for how a reductionist might be able to derive a zero point page 475 Part IV Chapter 5: (1) Further exploration of whether, and, if so, how, a reductionist might be able to derive a zero point, and (2) key discoveries about how the differences between AU and TU ultimately boil down to the respective theories’ different positions regarding whether an empty world can be compared to a non-empty world page 496 Part IV Chapter 6: Why a reductionist is unable, after all, to derive a zero point page 524 Part IV Chapter 7 (only a chapter summary): Is intrinsic goodness reducible to intrinsic betterness or is intrinsic betterness reducible to intrinsic goodness? page 553 Part IV Chapter 8 (only a chapter summary): Assessing whether or not there are absolute degrees of happiness (a question which ultimately boils down to the question of whether reductionism of happiness is true or whether non-reductionism of happiness is true) page 558 Conclusion of the Dissertation page 567 1 Executive Summary of the Dissertation The overall project of this dissertation is to articulate a plausible candidate for what Derek Parfit calls “Theory X”—a theory of welfare-based goods. The dissertation proceeds in a generally utilitarian framework, but, as I will explain, even if one is not a utilitarian, my findings will constitute important pieces of almost all moral theories—even if they do not constitute the full picture. My overall thesis is that we should espouse what I call AMU (average per conscious moment utilitarianism). This states that the best state of affairs is the one which has the highest average per moment happiness, as averaged across all conscious moments of sentient life— regardless of which moments are bundled together into different persons (or other beings). * * * * Articulating a plausible theory of population ethics is no easy task. Seemingly every theory that is introduced is confronted with unpalatable implications that suggest that the theory should be revised. As a result, population ethics is very much in turmoil. While there have been various approaches to articulating a plausible theory of population ethics, including non-consequentialist approaches, the subject has been dominated by consequentialist—and, in particular, utilitarian—thought. Perhaps the most prima facie attractive views are called “person-affecting views” (which I will define and address below), but these views are intractable because they are unable to provide plausible accounts when confronted with the “non-identity problem.” As a result of the difficulties afflicting “person-affecting views,” the two best-known approaches to population ethics have been total utilitarianism (“TU”) and average utilitarianism (“AU”). According to TU, we should seek to bring about the state of affairs that maximizes the total amount of happiness. According to AU, we should seek to bring about the state of affairs that maximizes the average amount of happiness per person. Seemingly strong objections have been made against both positions, though, and as a result, numerous variations of the accounts and hybrid accounts have been introduced. These variations and hybrid accounts, themselves, however, also confront strong objections. As a result, seemingly every theory of population ethics is confronted by unpalatable implications. Interestingly, despite the fact that there is widespread disagreement in the field, a near consensus has developed on one point—rejecting AU as an absurd view. Despite the fact that AU is such a well-known view, most philosophers who write on the topic reject it as absurd, and they focus their efforts on TU. Perhaps the main difficulty for a proponent of TU is what Parfit has called the Repugnant Conclusion (“RC”)—an alleged reductio against TU—and the philosophical landscape of population ethics is largely a function of the different ways in which philosophers handle the RC obstacle. One camp of philosophers endorses TU and “accepts” the RC, arguing that the RC does not serve as a reductio against TU because the RC is not in fact repugnant. Philosophers in another camp think that the RC is repugnant and that it does follow from TU, and they thus propose theories that are variations of TU that avoid implying the RC. Other philosophers still—though not utilitarians—argue that the RC can be avoided by denying the transitivity of “better than.” I think that all of these approaches are misguided, because I think that the main tenet of TU (and of most variations of TU) is wrong: It is not the case that, all else equal, a larger population of happy people is better than a smaller one. While this conviction has surely been shared by others, I think that others prematurely abandon it when the going seemingly gets tough. In this dissertation, I will go against the grain and argue that, if we are going to adopt a utilitarian approach to population ethics, then AU is the theory we should endorse. In sum, this dissertation will show how (and why), despite its maligned status, a properly formulated AU account is not only a plausible account of population ethics, but the most plausible account of population ethics. 2 * * * * The dissertation is broken up into four parts. Part I: Part I, which contains Chapters 1 and 2, aims to (1) lay the foundation for this dissertation’s project, and (2) provide a prima facie argument for average utilitarianism (AU). More specifically: Part I Chapter 1 will make a number of preliminary points and it will lay out the assumptions underlying the dissertation’s discussion—including, among other things, an explanation of how I understand the term “welfare-based goods.” Part I Chapter 1 will also set the stage for the arguments in the dissertation by explaining what the current state of the literature is. Part I Chapter 2 aims to explain the appeal of AU, and it describes a few possible AU accounts. First, however, it begins by explaining that person-affecting accounts are not plausible, and thus that Theory X must be an impersonal account, and that Theory X cannot be a hybrid account. Having done this, Part I Chapter 2 then explains the appeal of AU accounts. Although the chapter will flag a number of important objections to AU, it will defer discussions of these objections until later chapters. Part I Chapter 2 also explains that AU is an umbrella theory with various different possible versions. Although it mentions that my preferred account will be AMU, it defers arguments for which version of AU is the most plausible until Part II of the dissertation. Part II: Having argued elsewhere for the plausibility of average utilitarianism (having provided initial appeal for it in Part I Chapter 2, and defending it against objections in Part III), we are confronted with the question of what the relevant units are that should be averaged over. While the analogous question could theoretically arise for other theories, such as TU (where the question would be which units to take the additive sum of), in practice it would not, because the question would be a moot one for a theory like TU. The different sub-theories of TU would be coextensive. This, however, is not the case with AU, and thus there is a further question of which sub-version of AU is the most plausible. Part II addresses this question. A quick exploration into the matter at the end of Part I Chapter 2 will have suggested that ATLU (average total life value utilitarianism) is a possible candidate. (According to this view, the goodness of an entire state of affairs is equal to the average happiness of the lives it contains, and the happiness of a life is equal to the total (sum-aggregative) amount of happiness in the person’s life.) It seemingly allows us to combine two core intuitions: ATLU maintains the population-size anti-more-is-better intuition (AMBP) (that, all else equal, it is not better that more people exist than fewer people) and it also maintains the intrapersonal maximizing intuition (IMI) (that, all else equal, it is better agent-neutrally that a happy person live longer). As such, ATLU relies on the claim that, all else equal, it is not better for there to be more good conscious moments if they exist in an additional person, but it is better for there to be more good conscious moments if they exist in an already existing person. A question that confronts this view, however, is why the different location for additional good conscious moments is relevant. Part II will argue that the location for additional good conscious moments is not relevant, and thus, taking AU as given, it will argue that AMU—and not ATLU—is the most plausible version of AU. Part II Chapter 1 will offer a positive argument for why AMU is the most plausible version of AU. Part II Chapter 2 will then consider various objections to AMU. Most importantly, Part II Chapter 2 will explain why IMI, as stated, is false, and thus why AMU is true despite our IMI- friendly intuitions. Part II Chapter 3 will consider and address the question of whether death is bad for the person who dies, and, more broadly, it will consider and address what the most plausible account is of goodness for a person (i.e., agent-relative goodness). In so doing, Part II Chapter 3 will provide significant further support for the main conclusion of Part II Chapter 2—that IMI is 3 false. Thus, to the extent that Part II Chapter 3 provides support for the claim that IMI is false, Part II Chapter 3 will provide strong support for AMU. Lastly, Part II Chapter 3 will also consider and address some additional objections to AMU. Part III: Having argued in Part II that AMU is the most plausible version of AU, Part III will provide a defense of all AU theories, more generally. In so doing, it picks up where Part I left off. In Part I Chapter 2, I offered a prima facie argument in favor of AU. First, I argued that Theory X had to be impersonal (and not person-affecting), and then I argued that it had to be a pure theory (and not a hybrid theory). This then left us with AU and TU as possible candidates for Theory X. I then argued that AMBP (the population-based anti-more-is-better intuition) is a very attractive intuition that provides prima facie plausibility of AU, and I argued that it is an intuition that many readers will share (though it certainly is the case that some will not share this intuition at all and some, while sharing it most of the time, will have some situations where they do not share it). I argued that this intuition, especially when taking into consideration various prima facie problems with TU that do not afflict AU (such as the Repugnant Conclusion), constitutes a strong prima facie reason to espouse (or to at least strongly consider espousing) AU. Even though many people would find this prima facie case I provided in favor of AU (in Part I Chapter 2) to be appealing or persuasive, seemingly everyone rejects AU, and this is the case even if they would, prima facie, find AMBP plausible. Why is this? Why is it that a seemingly coherent view that seemingly plausibly rejects the “creating more happy people, all else equal, is good” view is rejected by most philosophers? In short, seemingly everyone rejects AU because of some well-known objections that people think are devastating objections against AU. Part III’s goal will be to answer these objections. If I am able to successfully answer all of these strong objections against AU, I think that I will be able to convince most readers that AU is correct. After all, as I stated, it seems to be that it is these objections that are what are preventing most people from espousing AU. If not for these objections, we would be left with my prima facie argument in favor of AU from Part I, and I think that if a reader were provided with this argument and if the objections against AU discussed in this Part had been disposed of, the vast majority of readers would espouse AU. While various objections to AU are made, despite their differences, almost all of them seem to be at least somewhat related to a single concern: Assuming that a particular action (e.g. bringing into existence a person with happiness level five) has no effects on others, why should the goodness of the action depend on anything extrinsic to those affected by the action (e.g. the happiness of the rest of society)? AU seems to dictate this conclusion, because the creation of an individual at “happiness level five” would seemingly be bad if the prior population’s average happiness were “happiness level ten,” but it would be good if the prior population’s average happiness were “happiness level three.” This concern is manifested—with varying degrees of directness—in three different types of objections to AU. Most directly, it is manifested in what I’ll call the “bottom up” objection. Somewhat less directly, it is manifested in what I’ll call the “top down” objection. Lastly, although the Two Hells objection has various features, it, too, manifests the above concern to some degree. First, the bottom up approach (which I address in Part III Chapter 1): One might point out that it is not at all clear which pool of individuals one should strive to maximize the happiness of: All sentient beings to ever exist? All people to ever exist? All people who currently exist? All people who exist currently or will exist in the future? It is unclear whether a non-arbitrary selection of any of these can be made. Further, and more importantly, according to this characterization of the concern, no choice of a relevant temporal population will yield plausible 4 results for the AU theorist. Thus, according to the bottom up argument against AU, AU is implausible because there isn’t a single “reference class” of individuals to average over that will render AU a plausible theory. The top down approach to objecting to AU (which I address in Part III Chapter 3) is to say that AU fails because it doesn’t always maximize what is of intrinsic value. The proponent of the top down objection says that since, according to utilitarianism, happiness is of intrinsic value and AU seeks to maximize something that often does not maximize the amount of happiness, AU fails to correspond with our intuitions. Parfit, McMahan, and Huemer are all examples of authors who provide a top down argument of this sort. Lastly, the objection to any version of AU that is perhaps regarded as the strongest is one that Parfit raised and termed Two Hells. Among other things, it addresses AU’s treatment of suffering. The objection goes as follows: Consider two possible hells. In Hell 1, ten people exist and suffer great agony. In Hell 2, there are ten million people who suffer ever so slightly less agony. AU will say that Hell 2 is better. Parfit and countless others, however, find it obvious that Hell 2 is far and away the worse of the two states of affairs. While the three types of objections, just discussed, constitute the main objections against AU, this Part also addresses a handful of other objections to AU (both well-known and non-well- known objections). Two of these other objections that I address are Parfit’s Adam and Eve objection and Arrhenius’s Sadistic Conclusion objection. In sum, Part III will consider the main objections that are typically considered the strongest objections to AU theories—objections responsible for the general consensus that AU theories are implausible. My view is that, in Part III, I will successfully respond to each of these objections, and that I thus defray each of these objections against AU. Thus, when this Part is taken in conjunction with the prima facie argument that I make for AU in Part I Chapter 2, the result is that the reader should be persuaded that AU is the most plausible candidate for Theory X. Thus, having argued in Part II that AMU is the most plausible version of AU, and since Part III provides a defense, more generally, of all AU theories (thus supporting the claim in Part I Chapter 2 that the most plausible Theory X is a version of AU), the conclusion after the first three parts of the dissertation will be that AMU is the most plausible candidate for Theory X. Part IV: Part IV provides an independent positive argument for AU (the umbrella theory). If successful, though, it not only provides a positive argument for AU, but it also provides a positive argument for AMU, in particular. Part IV argues that TU, and all other theories that employ any form of sum-aggregation (including ATLU), require the existence of a “zero point” (a term that will be defined in Part IV), and the bulk of the Part is then spent arguing that a zero point does not exist. In light of this, TU, and, importantly, all other theories that employ any form of sum- aggregation (including ATLU), are false. Part IV splits up the landscape of theorists into those who are reductionists of goodness (i.e. those who think that goodness is reducible to betterness) and those who are non-reductionists of goodness (i.e. those who think that goodness is not reducible to betterness—and who instead think that betterness is reducible to goodness), and it argues that reductionists are unable to get a zero point, and it argues that, while non-reductionists can get a zero point, non-reductionism ultimately is implausible. Thus, we must reject TU and all other theories that employ any form of sum-aggregation. Further, given my arguments in Part I (for the position that we must espouse an impersonal theory and the position that we must espouse a non-hybrid theory), we now see that we must espouse AU, and, in particular, AMU. As for the chapters within Part IV: Part IV Chapter 1 explains what a zero point is and it explains why TU requires a zero point. It also explains how this topic has been addressed in the literature and how the way it has 5 been addressed has not been the most fruitful way of addressing it. I then explain what the key points really are. Part IV Chapter 2 explains what reductionism of goodness is and it shows how the conceptual landscape is divided between those who espouse reductionism and those who espouse non-reductionism. This chapter then describes the structure of the arguments in Part IV. Part IV Chapters 3-6 address the question of whether a reductionist can allow for / derive the existence of zero points. Part IV Chapter 3 sets the stage for the inquiry into whether a reductionist can allow for zero points. This chapter distinguishes between two versions of reductionism, and it explains why only one of them can be and should be focused on by the reductionist. Part IV Chapters 4 and 5 argue that, contra common beliefs, it is far from clear that a reductionist is unable to get a zero point. These chapters provide what I think constitutes the best attempt that a reductionist can make to derive zero points. 1 Part IV Chapter 6 explains why the arguments in Chapters 4 and 5 ultimately fail, and, thus, why reductionism is unable to get a zero point. Part IV Chapters 7 and 8 explore whether non-reductionism or reductionism is a more plausible theory. Part IV Chapter 7 argues that, despite the common belief that reductionism is more plausible than non-reductionism, the various arguments made in support of this claim fail to support this conclusion, and we remain in an impasse between the reductionist and non- reductionist. Thus, I argue that the matter is much more nuanced and complicated than it has been recognized to be. Then, however, I argue that there is a new avenue that might prove fruitful in arguing for reductionism (of goodness). I argue that if we can show that the notion of “happiness” is a relativist concept and that it does not admit of absolute degrees, then we will get the result that goodness is a relative concept too, and thus that reductionism of goodness is more plausible than non-reductionism of goodness. Part IV Chapter 8 argues that there are no absolute degrees of happiness. If these arguments succeed, and if they are combined with the arguments of Part IV Chapter 7, I get the result that reductionism of goodness is more plausible than non-reductionism of goodness—albeit for reasons other than those offered by other authors (whose arguments, in fact, fail to support reductionism of goodness). Lastly, I close Part IV with a brief conclusion. In short, Part IV’s conclusion is that TU (and all other theories that employ any form of sum-aggregation—including ATLU) require zero points and cannot get zero points, and thus TU (and all other theories that employ any form of sum-aggregation) are false. Further, for the reasons discussed elsewhere in this introduction (and elsewhere in the dissertation), this provides strong support for the claim that AU (and, more specifically, AMU) is the most plausible theory. The Dissertation’s Conclusion: The dissertation’s Conclusion very briefly concludes. * * * * Ultimately, I will argue that many of the tensions in population ethics can be traced to two views that are in some tension with one another: The first is what I will call the “anti-more-is- better intuition”—the intuition that, when it comes to society at large, all else equal, it is not better for there to more people than fewer people. Second, when it comes to oneself, a person seeks to maximize the amount of accrued happiness going forward—one follows an analogue of total 1 Also, in the process of carrying out this attempt, I show that we can learn a lot about the relationship between AU and TU and their various commitments. In particular, it turns out that (within a reductionist framework) TU is committed to being able to derive a zero point whereas AU is not, and it turns out that the debate about whether or not a zero point can be derived boils down to whether one can compare cases of existence to cases of non-existence (for example, whether a betterness comparison can be made between a world with nothing conscious in it and a world with something conscious in it, or whether a betterness comparison can be made between the existence of a conscious moment at a time and its non-existence). 6 utilitarianism, and one does so with what I will call a “current / forward-looking reference class.” In other words, one seeks to maximize the total happiness that one will accrue between the present and one’s death. I will argue that many of the difficulties we encounter in the search for Theory X are due to intuitions about one’s own life infiltrating assessments for society at large. Three further points are in order before continuing. First, it is important to note that it will become clear at various junctures that two main Parts of the dissertation—Parts II and III— reinforce each other. The defense of AU in Part III is strengthened by the choice in Part II to endorse AMU as the most plausible version of AU. While the arguments in Part III apply to ATLU as well, AMU is even better equipped to respond to the three objections than is ATLU. It has additional resources in this regard. Similarly, the flip side of this coin is that these aspects of Part III provide additional support for the choice, in Part II, of AMU as the most plausible version of AU. These considerations are not among the points addressed in Part II, but they supplement the reasons for thinking that AMU is the most plausible version of AU. Second, it’s crucial to keep in mind throughout this dissertation, that articulating a plausible Theory X is difficult. There are many deep-seated intuitions that conflict. There is no answer that is unproblematic. It should be expected that there will need to be significant tweaking to get to reflective equilibrium. Thus, there being difficulties with a view is not reason enough to reject it in this domain. Referring to a similar topic, Parfit writes, “Some of my claims may seem implausible. This is what we should expect, even if these claims are correct.” 2 Similarly, Temkin says that in this domain, one’s state is “akin to that of a juggler who finds himself juggling too many balls, each of which is precious and important. He knows that he can’t keep juggling all of the balls indefinitely, but he can’t seriously imagine letting any of them drop.” 3 But he must. And we must. Any theory in this domain is going to have aspects that are prima facie counterintuitive. Third, it will be helpful to keep in mind that AMU, in practice, will turn out to not be as counterintuitive as it might seem to some readers prima facie, and this is for a few reasons. First of all, most cases are not “different-numbers cases” (a term I will introduce and define below), and, since all the theories discussed in the dissertation give us the same results in “same-numbers cases” (also a term I will introduce and define below), they will all give us the same result in most cases (and in typical cases). Second of all, even in the vast majority of different-numbers cases, AMU will, in practice, provide the same assessments and prescriptions as the theory that one prima facie might find the most plausible—be it, perhaps ATLU or a person-affecting morality. AMU often, even in different-numbers cases, will, in practice, provide the same assessments and prescriptions as ATLU (which prima facie might track our intuitions more than does AMU), and both AMU and ATLU will also in most cases provide the same assessments and prescriptions as person-affecting intuitions (which prima facie might track our intuitions even more closely than AMU or ATLU). The reason for this is that even if AMU is correct, if our intuitions are still aligned in many cases with these other theories, a result will be that doing things that are in accordance with the prescriptions of these theories might increase happiness, and this would thus be a good thing according to AMU. Thus, even if, say, death (even if caused by a murder) were not itself intrinsically bad according to AMU, it would typically be bad (according to AMU) in practice because of its effects on the happiness level of conscious moments that do exist, which would indeed be taken into account by AMU. The foregoing is akin to the way utilitarianism might often provide assessments and prescriptions as, say, deontological theories, even though utilitarianism would provide a prescription because of an effect typically brought about by something like murder, rather than by, 2 Parfit (1984), p.130. 3 Temkin, p.10. 7 say, a rule against murder. In cases in which the prescriptions between (what I argue is) the correct theory and the incorrect theory diverge, however, as is the case between utilitarianism and deontological theories (or between AMU and ATLU), my view is that these are cases where the incorrect theory must be abandoned. The incorrect theory, whose prescriptions and assessments often converge with those of the correct theory, often operate as successful heuristics and often bring bout the correct results. However, the cases where the correct theory and incorrect theory diverge are the “extreme” cases where the heuristics do not get the correct result and where we must stick to the assessments and prescriptions of the theory that we deem to be the correct theory. In sum, the third point is that even if AMU, in theory, appears prima facie to be contrary to one’s intuitions, in the vast majority of cases, AMU, in practice, will prescribe the result that is indeed consistent with one’s intuitions—despite this not having been immediately apparent. I think that keeping this point in mind throughout the dissertation will drastically lessen the prima facie counterintuitiveness of AMU, and thus drastically weaken potential objections to my position. Notwithstanding these points, however, there will indeed be cases of divergence between AMU and one’s prima facie intuitions, and, in these cases, it is AMU that (in my view) provides the correct assessment or prescription, and it is any other theory that one espouses (or intuition that one holds) that must be rejected in these situations. * * * * Parfit, the author who coined the term, “Theory X,” and who introduced us to the problem of articulating Theory X and its various difficulties, its counterintuitivities, its paradoxes, and its tensions, did not think that he had satisfactorily solved the riddle of Theory X (i.e. articulated a plausible account of Theory X). I argue that this dissertation accomplishes what Parfit was unable to accomplish: In my view, this dissertation solves the riddle and articulates a plausible account of Theory X. In short, the answer is AMU. If Parfit were still with us, and if he were able (and willing!) to read my dissertation, I don’t know if he would agree that I have found a plausible solution to the problem of Theory X, but I would like to think that he would. In my view, to riff on Sidgwick’s observation that the first word of his book, The Methods of Ethics, was “ethics” and the last word “failure,” it is my view that inquiries in population ethics attempting to articulate Theory X need not be failures (as some philosophers have explicitly argued and others have seemingly implicitly believed), 4 and it is my (admittedly bold) view that, even if all attempts to date to articulate Theory X have been failures, this dissertation’s attempt to articulate Theory X is not a failure! 5 One final comment: In carrying out my search for and defense of a plausible candidate for Theory X, I hope to also, along the way, shed light on a number of diverse and wide-ranging areas in philosophy that one might not have realized were connected to the population ethics inquiry 4 As discussed later in the dissertation, Gustaf Arrhenius provides an “Impossibility Theorem,” which he argues that it is not even possible for a satisfactory account of Theory X to be articulated. If, however, one does not agree with Arrhenius about the importance of each of his conditions, however, then one likely will not be moved by his conclusions. Notwithstanding this, it is notable that there are authors such as Arrhenius who indeed argue that indeed is impossible to articulate a plausible account of Theory X. Further, although Arrhenius states this position explicitly, it seems to be implicitly stated and believed by many, if not most, philosophers that it is impossible to satisfactorily articulate a candidate for Theory X—i.e. to articulate a single theory that one can fully throw one’s weight behind and fully sign on to. 5 Parfit felt that he had failed to solve the riddle of Theory X, but he, unlike Arrhenius and others, thought that the riddle was solvable. He wrote: “When he was asked about his book, Sidgwick said that its first word was Ethics, and its last failure. This could have been the last word of my Part Four. As I argued, we need a new theory about beneficence. This must solve the Non-Identity Problem, avoid the Repugnant and Absurd Conclusions, and solve the Mere Addition Paradox. I failed to find a theory that can meet these four requirements. Though I failed to find such a theory, I believe that, if they tried, others could succeed.” (Parfit, 1984, p. 443.) 8 into articulating a plausible account of Theory X. Among these areas are: weak will, the self, metaphysics of personhood, causation, philosophy of time, philosophy of consciousness, higher- order desires, utilitarianism’s treatment of animals, questions regarding the relativity of happiness, relativism versus absolutism in physics, measurement theory, and the question of whether a theory in value theory can be hybrid. Not only do I hope to shed some light on these areas in this dissertation, but I hope, in this dissertation, to provide the beginnings of theories in all of these areas that, if further developed and expanded upon, could constitute important contributions to these respective sub-fields of philosophy. Thus, while I view this dissertation as an attempt to provide the first full plausible solution to the riddle of Theory X, I also view it as an attempt to offer the beginnings of contributions on a wide range of related topics. At long last, I now proceed to carry out my project and to explain why AMU is the most plausible candidate for Theory X. ====================================================================== Part I: Introduction, preliminary points, the theoretical problem and the landscape of the literature, the broad importance of the puzzle, practical implications, and a prima facie argument in favor of AU Part I, which contains Chapters 1 and 2, aims to (1) lay the foundation for this dissertation’s project, and (2) provide a prima facie argument for average utilitarianism (AU). More specifically: Chapter 1 will make a number of preliminary points and it will lay out the assumptions underlying the dissertation’s discussion—including, among other things, an explanation of how I understand the term “welfare-based goods.” Chapter 1 will also set the stage for the arguments in the dissertation by explaining what the current state of the literature is. Chapter 2 aims to explain the appeal of average utilitarianism AU, and it describes a few possible AU accounts. First, however, it begins by explaining that person-affecting accounts are not plausible, and thus that Theory X must be an impersonal account, and that Theory X cannot be a hybrid account. Having done this, Chapter 2 then explains the appeal of AU accounts. Although the chapter will flag a number of important objections to AU, it will defer discussions of these objections until later chapters. Chapter 2 also explains that AU is an umbrella theory with various different possible versions. Although it mentions that my preferred account will be AMU, it defers arguments for which version of AU is the most plausible until Part II of the dissertation. 9 Chapter 1: Introduction, preliminary points, the theoretical problem and the landscape of the literature, the broad importance of the puzzle, and practical implications 1.A. Introduction to the Theoretical Puzzle and the Turmoil that Afflicts the Population Ethics Literature Articulating a plausible theory of population ethics is no easy task. Seemingly every theory that is introduced is confronted with unpalatable implications that suggest that the theory should be revised. As a result, population ethics is very much in turmoil. While there have been various approaches to articulating a plausible theory of population ethics, including non-consequentialist approaches, the subject has been dominated by consequentialist—and, in particular, utilitarian—thought. Perhaps the most prima facie attractive utilitarian 6 views are called “person-affecting views” (which I will define and address below), but these views are intractable because they are unable to provide plausible accounts when confronted with the non-identity problem. 7 As a result of the difficulties afflicting “person-affecting views,” the two best-known approaches to population ethics have been total utilitarianism (“TU”) and average utilitarianism (“AU”). According to TU, we should seek to bring about the state of affairs that maximizes the total amount of happiness. According to AU, we should seek to bring about the state of affairs that maximizes the average amount of happiness per person. Seemingly strong objections have been made against both positions, though, and as a result, numerous variations of the accounts and hybrid accounts have been introduced. These variations and hybrid accounts, themselves, however, also confront strong objections. As a result, seemingly every theory of 6 Person-affecting views really aren’t utilitarian views at all, and indeed they are inconsistent with utilitarianism. As will be clear in my later discussion, this is because utilitarian views are impersonal views and, unlike person-affecting views, they in many ways ignore questions of which specific persons are at which happiness levels. The sense in which I’m referring to person-affecting views as being able to be utilitarian views, though, is just in the sense that a person-affecting view could be a view that only assesses goodness for a particular person by looking at the person’s happiness level. A person-affecting view, could, however, on the other hand, be one that values things other than happiness / welfare. In light of the simplifying assumptions that I address below, in Section 1.B, discussions of person-affecting views in this dissertation will focus exclusively on “utilitarian” versions of person-affecting views. As will become clear, though, the problems that afflict utilitarian person-affecting views also (equally powerfully) afflict non-utilitarian person-affecting views. 7 See below, in this section, for a definition of the non-identity problem and a discussion of how person-affecting views are unable to provide plausible accounts when confronted with the non-identity problem. 10 population ethics is confronted by unpalatable implications. Interestingly, despite the fact that there is widespread disagreement in the field, a near consensus has developed on one point—rejecting AU as an absurd view. Despite the fact that AU is such a well-known view, most philosophers who write on the topic reject it as absurd, and they focus their efforts on TU. Perhaps the main difficulty for a proponent of TU is what Derek Parfit has called the Repugnant Conclusion (“RC”) 8 —an alleged reductio against TU—and the philosophical landscape of population ethics is largely a function of the different ways in which philosophers handle the RC obstacle. One camp of philosophers endorses TU and “accepts” the RC, arguing that the RC does not serve as a reductio against TU because the RC is not in fact repugnant. 9 Philosophers in another camp think that the RC is repugnant and that it does follow from TU, and they thus propose theories that are variations of TU that avoid implying the RC. 10 Other philosophers still—though not utilitarians—argue that the RC can be avoided by denying the transitivity of “better than.” 11 I think that all of these approaches are misguided, because I think that the main tenet of TU (and of most variations of TU) is wrong: It is not the case that, all else equal, a larger population of happy people is better than a smaller one. While this conviction has surely been shared by others, I think that others prematurely abandon it when the going seemingly gets tough. In this dissertation, I will go against the grain and argue that, if we are going to adopt a utilitarian approach to population ethics, then AU is the theory we should endorse. Before getting to my own arguments, however, I provide this initial chapter, which carries out various stage-setting tasks. After this introductory sub-section, Section 1.A, I proceed to Section 1.B, in which I lay out various assumptions that the dissertation will be operating under, 8 Parfit (1984). 9 See Huemer (2008); Ryberg (1996); Ryberg (2004); Tännsjö (1992); Tännsjö (2002). 10 See Hurka (1983). 11 See Temkin (1987). 11 and I make handful of other similar preliminary points. Next, in Section 1.C, I provide an overview of the theoretical puzzle that is population ethics. I begin by defining and discussing “different-numbers” problems, explaining how and why they give rise to the theoretical puzzle that this dissertation sets out to solve. After thus explaining what the puzzle is and what creates the puzzle, I canvas the various approaches that have been taken to population ethics. I explain not only what these various approaches are, but I will also explain the various difficulties that seemingly afflict the various accounts provided. Thus, this section will illustrate the turmoil in the population ethics literature, a literature that seemingly still has not identified a plausible account of Theory X. Indeed, in this section, I also address the views of one philosopher, Gustaf Arrhenius, who offers an “impossibility theorem,” arguing that not only has no plausible account of Theory X been articulated, but no plausible account of Theory X even can be articulated. Next, in Section 1.D, I proceed to explain that it is the case (and how and why it is the case) that articulating a plausible account of Theory X is not a task that is merely important for making sense of population ethics, but, rather, that articulating a plausible account of Theory X constitutes a task of foundational theoretical significance that permeates throughout value theory and reaches core questions that underlie value theory as a whole. Then, in Section 1.E, I describe the many and wide-ranging practical implications of the theoretical exploration into articulating a plausible account of Theory X. Our account of Theory X has great implications both for individual decisions as well as a wide range of policy decisions made by the government. In both contexts, these implications involve questions of how to value not only population size but also questions of how to value the lengths of individual lives. Only after carrying out the various tasks just described, and thus completing Chapter 1, will I turn to my own arguments—specifically, my arguments in favor of AU. I begin my discussion of AU in Chapter 2, which provides an argument in favor of the prima facie plausibility 12 of AU. Chapter 2 begins by describing the desiderata of Theory X and re-establishes the need for Theory X to be an impersonal theory. It then proceeds to explain in depth why we should reject hybrid theories, thus leaving the two “pure” theories—AU and TU—as the only potentially plausible candidates for Theory X. Having established this, Chapter 2 then provides a prima facie case in favor of AU. I begin making this prima facie case in favor of AU by providing a discussion of the main intuition motivating AU, and I then explain why AU is the position we must adopt if we want to retain this intuition. This will, by itself, constitute a prima facie argument for AU. In other words, this will explain why one might—at least prima facie—find AU to be an appealing theory. Of course, there are various important objections that are raised against AU, but a discussion of these objections (and why they all miss the mark) will be deferred to later chapters (in Part III). Also deferred to later chapters will be all sorts of other related discussions. In sum, this dissertation will show how (and why), despite its maligned status, a properly formulated AU account is not only a plausible account of population ethics, but the most plausible account of population ethics. Having described the route ahead, I now proceed with the first task at hand: describing some of the background assumptions of the dissertation, after which I will canvas the theoretical landscape of the population ethics literature. 1.B. Preliminary points: Assumptions and clarifications regarding metaethics, axiology vs deontology, the nature of value and goodness, the importance of this topic for almost all theorists (and not just for utilitarians) etc. This short sub-section will make a number of preliminary points that clarify what I will be doing, what I will not be doing, and what assumptions I will be making. I will address metaethical preliminaries and also some issues regarding definitions of utilitarianism. I will also explain the broad importance of this project: It is not relevant only to utilitarians, but it is relevant to almost 13 all theorists, albeit to different degrees. Conclusions about welfare-based goods will be of importance to most theorists, even if one does not think that welfare is the sole matter of importance. Welfare-based goods are at least a component of most accounts. Point 1: Metaethics This dissertation addresses a puzzle that sits, fairly squarely, in the domain of normative ethics. Lying beneath issues in normative ethics, however, are various issues in metaethics, many of which cast doubt on various aspects of the normative ethics at issue in this dissertation. Among other things, some deny that there can be an overall goodness of a state of affairs; others doubt the authority of intuitions in providing ethical knowledge. 12 In this dissertation, however, I will bracket these issues and assume that there is such a thing as overall goodness and that our intuitions can provide ethical knowledge. Point 2: Assuming a utilitarian account (or welfare-based goods account) of axiology Next, I will be assuming a utilitarian account for the assessment of the goodness of states of affairs, according to which happiness (or, more generally, welfare-based goods) are what are of value. It’s absolutely crucial, though, that the reader recognize why I’m doing this. I’m doing this to fix ideas and to simplify my inquiry. I’m not making this assumption because it’s so obviously right, or because those who disagree are obviously wrong and not worth engaging with on this point. The point here is simply to fix ideas and to narrow and simplify the scope of my dissertation’s inquiry. The goal, despite making this assumption, is to have the dissertation be relevant for almost all theorists. 12 Mackie (1977). 14 (a) This utilitarian account (or welfare-based goods account) of axiology is consistent with many types of accounts It is important here that I explain and address the question of how I understand the term “welfare-based goods.” After all, welfare is generally thought of as a property of a person’s life, but my theory holds that person-data is not directly relevant to the goodness of an outcome, and that the goodness of an outcome is a function of the goodness of the conscious moments it contains. Thus, one might wonder whether I can even speak of the “welfare of a conscious moment.” Further, one might wonder how I can speak in welfarist terms if I ultimately sever the link between what’s good for people and what’s good agent-neutrally. Similarly, one might think that my view, which focuses on the goodness of conscious moments, is committed to hedonism. In short, though I must admit I am a hedonist, I do not think that the view defended in the dissertation is committed to hedonism, and further, I think it can accommodate the vast majority of non-hedonistic utilitarian positions. Perhaps my account won’t be able to accommodate some Aristotelian views that focus on a full life-time, but I think that my focus on conscious moments can accommodate most other views. For example, my account would be consistent with the Millian distinction between higher and lower pleasure. This distinction could exist even at the momentary level. Similarly, my account could accommodate Moore’s account of veridical pleasures. Similarly, it could accommodate many aspects of objective list theories, and also—but perhaps more debatably—preference satisfaction accounts. Furthermore, I don’t think that there is any problem with assigning levels of welfare to conscious moments. 13 Of course, my account might not be able to accommodate all understandings of “welfare,” but I think that it will be able to accommodate many. I will assume that we can assign a level of welfare to conscious moments. 13 Although I don’t think we need to resort to the following to makes sense of this, here is a possible analogy that give additional reason to think that it makes to assign a level of wellbeing to a conscious moment: Even though nothing can move instantaneously, we can certainly make sense of its moving, and having a certain determinate velocity and direction, at a particular instant in time. 15 Going forward, however, for ease of presentation, I will use the words “happiness” and “welfare” interchangeably, and I intend for their meanings to be the same. (b) My analysis in this dissertation, operating within a utilitarian framework (or welfare-based- goods framework) of axiology, is relevant for and important to just about all theorists (even if they are not operating within a utilitarian framework or welfare-based-goods framework). This is because, even if happiness or welfare-based goods do not constitute a theorist’s whole account, almost all theorists maintain that happiness or welfare-based goods are at least one component of their account. The fact that my dissertation will be operating within a utilitarian framework (or welfare- based-goods framework) does not mean that the discussions in my dissertation will not be relevant for or of interest to a non-utilitarian. My discussions in this dissertation, while often couched in utilitarian terms (as is the case for most works on population ethics), will similarly be greatly important and relevant to the production of complete descriptions of other non-utilitarian accounts. The discussions here are relevant not only to utilitarians, but also to almost all other theorists (albeit perhaps to different degrees). Conclusions about happiness (or, to use the more general term that I will frequently employ, conclusions about welfare-based goods) will be of great importance to most theorists, even if one does not think that welfare is the sole matter of importance. Welfare-based goods are at least a component of most accounts. For example, consider a view according to which distributional equity is a part of the picture in determining the goodness of a state of affairs. My dissertation will not be addressing issues such as distributional equity, in part because of the fact that for a utilitarian (such as myself), questions of distributional equity are subsumed within an analysis of utility, and in part because, even if this were not the case, there is more than enough to discuss in the dissertation before even addressing questions of distributional equity. Despite my not addressing these questions, however, those who consider welfare a component, but not the sole component, of the good can think of the issues discussed here as merely analysis of the welfare component of 16 good—as opposed to the entire picture. Thus, the discussion here is compatible with there being further inquiries into other issues, such as those regarding distributional equity, once the inquiry into welfare is complete. A theorist who thinks that welfare is a component, but not the only component, of what’s important can implement the analysis in this dissertation in making sense of the welfare component of a state of affairs, and then turn to an analysis of whatever else he deems important, thus using the data coming out of his analysis of these other aspects of a state of affairs to supplement the data coming out of my analysis of the welfare aspect of a state of affairs (which could also become “his” analysis of the welfare aspect of a state of affairs, if he finds my analysis persuasive and adopts it). Point 3: Axiological questions versus questions relating to assessing actions: What I will be focusing on and which assumptions I will be making Next, it’s important to note that with respect to the distinction between the goodness of a state of affairs and the goodness of an action, there are two different ways of understanding utilitarianism: One might think of it as a way of making comparisons between the goodness of states of affairs, but another equally common way of thinking of it is as a way of making prescriptions about actions. These notions might not be identical, but surely there is a close connection. The connection I will assume in this dissertation is act consequentialism: that the better action, according to utilitarianism, is the action that brings about the better state of affairs. Thus, the better of two actions, according to TU, is the one that brings about the state of affairs with the greater total amount of happiness, and the better of two actions, according to AU, is the one that brings about the state of affairs with the greater average amount of happiness per person. These clarifications notwithstanding, however, my primary purpose in this paper—though not my sole purpose—will be to address axiological questions. Thus, this dissertation does not need to be committed to the view about the relation between goodness of states of affairs and the goodness of 17 actions that I said I will be assuming. Indeed, the dissertation need not be committed to any view about the relation between goodness of states of affairs and goodness of actions. Even though I will at times talk about action, this dissertation, at bottom, is about axiology. Point 4: An additional point, which relates both to Point 2 (regarding how my account is relevant to most non-utilitarian accounts since almost all accounts maintain that welfare is at least one component of the picture, even if not the only component) and Point 3 (regarding axiology versus assessment of actions), above It’s important to note that this dissertation’s inquiry (solving the puzzle of articulating a plausible Theory X) is concerned with, aimed at, and relevant for both (governmental) policy decisions and decisions of personal morality. Of course, most governmental policy decisions (e.g. what the tax rate should be) and most decisions of personal morality (e.g. whether to go, as promised, to dinner with a friend) are “same numbers cases,” 14 though, so utilitarian accounts of welfare will come out pointing in the same direction most of the time, and thus the distinctions at issue in this dissertation will not be relevant. Nevertheless, “different numbers cases” 15 do indeed appear both in governmental policy decisions and in decisions of personal morality. In both of these domains, though, this dissertation will be focusing on the question of what’s best, not on questions regarding right and wrong actions. As such, it might be that this is the full analysis for policy decisions, while, in the personal domain, it might be that after questions of what is best are answered, there is a subsequent inquiry into how demanding morality should be, and other related questions. Thus, this is another area in which one might think that my inquiry is only one of many pieces of the puzzle. 14 I will introduce and explain this term in the next sub-section. 15 I will also introduce and explain this term in the next sub-section. 18 1.C. The theoretical challenge of coming up with Theory X: What the puzzle is, why it exists, and the current landscape of the literature that attempts (currently without success) to solve the puzzle Having made the foregoing preliminary points, I now proceed to provide an overview of the theoretical puzzle that is population ethics. I begin by defining and discussing “different- numbers” problems, explaining how and why they give rise to the theoretical puzzle that this dissertation sets out to solve. After thus explaining what the puzzle is and what creates the puzzle, I canvas the various approaches that have been taken to population ethics. I explain not only what these various approaches are, but I will also explain the various difficulties that seemingly afflict the various accounts provided. Thus, this section will illustrate the turmoil in the population ethics literature, a literature that seemingly still has not identified a plausible account of Theory X. Indeed, in this section, I also address the views of one philosopher, Gustaf Arrhenius, who offers an “impossibility theorem,” arguing that not only has no plausible account of Theory X been articulated, but no plausible account of Theory X even can be articulated. 1.C.1. The theoretical challenge: Different-numbers problems 1.C.1.a. Context Why, one might wonder, could it be difficult to determine what is the best outcome in terms of happiness values if we grant ourselves that happiness is the one thing that is of value? Isn’t the state of affairs that maximizes happiness the better state of affairs? Or, to scale back the assumption that happiness is the only thing of value, another formulation of this pair of questions is: If we are only focusing on the dimension of value that is a function of happiness values, isn’t it clear what the best outcome is in terms of the value of happiness? Isn’t the state of affairs that maximizes happiness the state of affairs that is the best outcome in terms of happiness? The answers to these questions seem as though they should be simple, and in most cases 19 (and typical cases), the answers are indeed simple. There are some cases, however, where the answers are less simple, though, and these are the cases, common in population ethics but less common in other domains of ethics, that involve what I will call “different-numbers problems”—a term that I will define and explain below. It turns out that there are a number of axiological accounts of welfare-based goods, and, in typical cases, we need not choose which among the various possible specific views we should endorse, because they all would provide the same betterness-assessment or prescriptions regarding different states of affairs. As a result of the fact that so many of the specific accounts of welfare-based goods provide the same assessments or prescriptions, not only do we not need to choose among these views in typical cases, but it might not even be apparent that there are different views that are capable of providing different assessments in other cases, and, further, it might not even be apparent that there are different views that even exist. Thus, it is the cases I discuss below, “different-numbers cases,” that force us to be more precise in articulating a theory of welfare-based goods. (Incidentally, two of the reasons that I think it is so difficult to articulate a single plausible account of Theory X seemingly are that (1) these non-typical cases are not cases that we explicitly confront frequently, and thus we are less accustomed to making these assessments and our intuitions might thus be less developed than they are in cases that we confront more frequently, and (2) since, in the typical cases, the various specific theories provide the same result, we are used to having the various features of the different specific accounts get grouped together with each other, and we are used to signing onto all of these features, and thus when the different specific theories (and their respective features) are pulled apart and point in different directions, we often might be at a loss about which specific theory and feature we should give up (if we must give up one or more), since we are used to being able to be in favor of all of the specific theories and features simultaneously, and it might not be clear to us, without significant thought, which of 20 these features we find most important, plausible, or correct.) Having laid out some of the context regarding different-numbers cases, I now explain what different-numbers problems are and I explain how and when these cases arise. 1.C.1.b. “Different-Numbers Cases” (as Distinguished from “Same-Numbers Cases”) In the following description of “different-numbers cases,” I state that they force us to choose between aggregation mechanisms, and I discuss the different-numbers cases in the context of presenting a need to choose between AU and TU. Describing different-numbers problems in this context is just for the ease of exposition, however. While indeed most of the dissertation will amount to a debate about which of just these two theories—AU and TU—is the most plausible theory in different-numbers cases, because, as I argue, these two theories end up being the two most plausible contenders for Theory X, different-numbers cases make us not only consider the relative merits of AU and TU (and thus which aggregation mechanism is most plausible), but they also make us consider other specific accounts of welfare-based good that are not rooted in aggregation mechanisms at all, or at least not directly (for example, person-affecting accounts). Notwithstanding this wrinkle, the following explains what different-numbers problems are and how they force us to be more precise in articulating our theory of welfare-based goods. The types of situations that require us to choose which aggregation method to espouse are cases that I will call “different-numbers cases,” and these are to be distinguished from what I will call “same-numbers cases.” Both different-numbers cases and same-numbers cases have the following features. Both cases are cases where we are comparing two (or more than two, but for simplicity’s sake, I’ll focus on two) possible states of affairs, both of which contain a number of parts (or “components”), and both in which the goodness or value of each of the whole states of affairs is a function of the goodness or value of each of the components that exists in that state of 21 affairs. These types of cases can arise in various contexts where the components are various different types of things, but the two main types are (1) cases where a component is a person’s life and where the whole is, say, a world, and (2) cases where a component is a single conscious moment that a person experiences and where the whole is, say, a person’s whole life. 16 Further, to keep things simple (for the time being), I’ll focus in particular on the first of these two types of cases—cases where the components in question are individual lives. Thus, in the cases that I’m considering, we’re comparing the goodness of two worlds, and each world is made up of nothing other than the lives that exist in the world. 17 Further, the goodness or value (here, happiness level) of a whole world is a function of the goodness or value (here, happiness level) of its components. The foregoing features are what same-numbers cases and different-numbers cases have in common. The way in which same-numbers cases and different-numbers cases differ is (perhaps unsurprisingly in light of the names I’ve given them) that in same-numbers cases, the numbers of components in the two states of affairs being compared are the same, whereas in different- numbers cases, the numbers of components in the two states of affairs being compared are different. Taking as our two basic aggregation methods the two aggregation methods that I have introduced, TU-type aggregation and AU-type aggregation, we can now see that while same- numbers cases will not require us to choose between these two methods, the different-numbers 16 Although these two examples involve a world (the whole) that is made up of people (the components) and a person (the whole) that is made up of conscious moments (the components), all sorts of other possible same-numbers and different-numbers cases exist. For example, to name just a few other types of data that could give rise to different- numbers and same-numbers cases: They could involve things such as wins and games played for a sports team, or shots made and shots taken by a player in basketball, or numbers of pieces one has in a game of checkers and whether the pieces are kings or not. In this dissertation, however, I will focus on (1) cases where the whole is a person’s life (or a portion thereof), and where the components are his moments of conscious experience, and (2) where the whole is a world (i.e. an entire state of affairs), and where the components are the lives that exist in that world. 17 Or, more specifically, the two worlds are made up of nothing that’s relevant for our inquiry other than the lives that exist in the worlds we are considering. I’m not denying that other things exist in the worlds, such as trees, air, buildings, etc. 22 cases will often require us to do so. Let’s first consider a same-numbers case. Take, for example, a case where a world, call it w1, is a world contains ten lives. Let’s suppose that each life in w1 is a life that we assess as being a life with a happiness value assessment of seven. Suppose that a different world, w2, also includes ten lives, but all of the lives in w2 have a happiness value assessment of five. In assessing the happiness levels (or levels of goodness or value) of w1 and w2, and in assessing whether w1 or w2 is the world with a greater happiness-value assessment (or a higher level of goodness or value), we need not choose between AU-type aggregation and TU-type aggregation. According to both aggregation methods, the happiness-value assessment of w1 is higher than that of w2, and by the same ratio: According to AU-type aggregation, the happiness-value assessment of w1 is 7 and the happiness-value assessment of w2 is 5, and according to TU-type aggregation, the happiness-value assessment of w1 is 70 and the happiness-value assessment of w2 is 50. Further, in same-numbers cases, it will always be the case that AU-type aggregation and TU-type aggregation provide the same determination of which state of affairs is better and provide the same ratio by which this is the case. The reason for this is simple: The AU-type aggregation value for a state of affairs is equal to the total sum of values divided by the number of components, and since the same number of components exist in both states of affairs in a same-numbers case, the AU-type aggregation value for each state of affairs will be arrived at by taking the TU-type aggregation sum in each case and dividing it by the same number in each case. Thus, if a TU-type happiness-value assessment for w1 is higher than a TU-type happiness-value assessment for w2, and by a ratio of, say, seven to five, dividing the TU-type happiness-value assessments for w1 and w2 by the same number 18 preserves both the fact that the w1 number is bigger than the w2 number, and the ratio 18 This holds true because the number of components will always be a positive number. Also, though the following feature is not required for the statements in question to hold true, it’s worth noting that the number of components will also always be a whole number. (For TU, there are in principle some cases where the number of components is zero. 23 by which this is so. In light of this, in same-numbers cases, we need not make a choice between espousing a TU-type aggregation mechanism or an AU-type aggregation mechanism. In different-numbers cases, however, we generally do have to choose between espousing TU-type aggregation and AU-type aggregation, and this is because the two aggregation methods do not always provide the same assessments of states of affairs. Consider, for example, two possible worlds, w3 and w4. W3, which is identical to w1 (above), contains ten lives, each of which has a happiness-value assessment of seven. W4, however, contains fewer lives. In w4, however, the lives that do exist have a higher happiness-value assessment than do the lives in w3. In w4, there are only five lives, but each one has a happiness-value assessment of ten. Given these two possible states of affairs, which one has a higher happiness-value assessment? (Or, stated slightly differently, which state of affairs has more happiness? Or, stated in another way, which of these two states of affairs is better?) One’s answer to this question will depend, crucially, on whether one espouses TU-type aggregation or AU-type aggregation, because they will not say the same thing. According to TU-type aggregation, the happiness-value assessment of w3 is 70, and the happiness-value assessment of w4 is 50. According to AU-type aggregation, however, the happiness-value assessment of w3 is 7, and the happiness-value assessment of w4 is 10. Thus, TU- type aggregation says that w3 has a higher happiness-value assessment than w4, but AU-type aggregation says that w4 has a higher happiness-value assessment than w3. While in this case TU- type aggregation and AU-type aggregation have differing assessments of which state of affairs has a higher happiness-value assessment, this, however, is not always the case. Even in many cases where they do not have differing assessments of which state of affairs has a higher happiness- So, for TU (but not AU), the only constraint is that the number of component must be non-negative, not that it must be positive. If we were considering a case where the number of components was zero, though, this wouldn’t be a counterexample to my point in the text, because the happiness-value ratio seemingly might be undefined, so it isn’t the case that the ratio “wouldn’t be preserved.” Questions concerning worlds with zero components, i.e. “empty worlds,” will be addressed at length in Part IV of the dissertation.) 24 value assessment, though, they still may provide different assessments, because even if both say that one state of affairs is the better one, one of the mechanisms might say that the happier state of affairs only has slightly more happiness whereas the other mechanism might say that the gap in happiness level between the two states of affairs is huge. These examples illustrate that different-numbers cases require us to choose between TU- type aggregation and AU-type aggregation. It’s important to note, though, that these different- numbers cases need not be comparisons between states of affairs where the components are all of the individual lives in a world. Instead, perhaps the numbers in the examples above could be applied to an example where the numbers represent the lives in some subset of a world’s population and where the whole represents that subset. Further, leaving aside the context where a state of affairs consists of components that are individual lives, similar examples can be provided where the whole in question is a single life and where the components (of which there are different numbers in the two states of affairs) are the conscious moments of the person’s life (or the number of years of conscious moments in the person’s life). Or, similarly, the numbers in the examples above could represent years of life in a person’s future (which, of course, could also be considered to be shorthand for, say, one billion conscious moments). Similarly, in different-numbers cases in these contexts, we need to choose between AU-type aggregation and TU-type aggregation. As I’ve shown, different-numbers cases (unlike same-numbers cases, which are the more typical cases we confront) require us to be more precise in articulating our account of Theory X— our theory of everything, or the theory that gives us the correct (or most plausible) answer in every case. In other words, different-numbers cases (unlike same-numbers cases, which are the more typical cases we confront) require us to choose among various possible specific accounts of Theory X. As we will see in this dissertation, and as I stated a few pages above, these choices are not easy ones to make. This is because seemingly every theory, even if seemingly plausible in 25 some cases, encounters some cases where, if the theory is applied, it yields seemingly unpalatable implications—implications so unpalatable that they seemingly provide strong reasons not to espouse the theory. The problem, however, is that seemingly every theory has some cases in which its implications are unpalatable, and seemingly no theory is able to avoid unpalatable implications altogether. Having now explained what different-numbers problems are, and in what types of situations they arise, I turn now to the various possible accounts that one could offer as a possible candidate for Theory X, as well as the various problems that afflict each of these accounts. 1.C.2. Non-consequentialist approaches As a preliminary matter, it is worth mentioning, very briefly, that there indeed are non- consequentialist approaches that are taken to population ethics and to articulating candidates for Theory X. Among those considering non-consequentialist approaches are also authors, such as James Woodward 19 and Thomas Hurka, 20 who are optimistic about the prospects for theories that are a mix of consequentialist and non-consequentialist principles. Although attempts to articulate a Theory X using non-consequentialist principles do exist (be these theories either fully non- consequentialist or part consequentialist and part non-consequentialist), I will not consider these attempts in this dissertation. As I explained in Section 1.B, above, this dissertation will be focusing primarily on axiology, and, to the extent it considers questions about action, it will assume a consequentialist framework. Thus, for these reasons, I will not further consider views that employ non-consequentialist approaches, regardless of whether these approaches are fully non-consequentialist or only partly non-consequentialist. 19 Woodward, James. (1986) “The Non-Identity Problem, Ethics, 96, pp. 804-831. 20 Hurka, Thomas. (1983) “Value and Population Size,” Ethics, 93, pp. 496-507. 26 1.C.3. Person-affecting accounts Perhaps the prima facie most plausible type of account is what are called “person-affecting accounts.” As Parfit describes these theories, a “person-affecting” morality is a moral theory according to which the part of morality concerned with human well-being “should be explained entirely in terms of what would be good or bad for those people whom our acts affect.” 21 Person-affecting theories can come in various forms. Two of the various sub-versions of the theory could be called the “sufficient” version and the “necessary” sub-version. According to the “sufficient” version, state of affairs x is better than state of affairs y if x is better than y for some actual person and worse for no one. According to the “necessary” version, state of affairs x is better than state of affairs y only if x is better than y for some actual person. There are also various other versions of person-affecting theories that include both of the above formulations and others that include neither of these two formulations, and there are various other related formulations that could be among the possible definitions of theories that are based on “person- affecting” principles. It’s also important to note that, in addition to the various ways in which person-affecting theories can be articulated along the dimensions described in the past paragraph, there are also other ways in which these theories can differ from each other. For one, different versions of these theories can have different accounts of what it means for a person to be “affected.” Secondly, different versions of these theories can also have different accounts of what specifically is meant by the word “person.” Despite the fact that numerous variations of “person-affecting theories” exist, however, I will not at this point elaborate further on the details of the different versions of these accounts. Despite the initial plausibility of “person-affecting” accounts, however, the vast majority 21 Parfit (1984, p. 394). 27 of the population ethics literature rejects even utilitarian person-affecting theories and instead focuses on “impersonal” utilitarian theories. As Parfit and others argue, person-affecting accounts must be rejected because they are unable to satisfactorily address or handle the “non-identity problem.” 22 The non-identity problem is illustrated by Parfit’s “14-Year-Old Girl” case: A young girl has a child, who, due to her mother’s youth, has a life that is difficult, yet worth living. If, however, the girl had waited a number of years to have a child, she would have had a different child, but one that would have had a better life than that of her actual child. The problem is that it seems clear that it would be better if the girl waited before having a child, but we cannot explain this by saying the girl’s decision was worse for her child, because if the girl had waited, a different child would have been born. 2324 Thus, despite its initial plausibility and strong attempts to defend it, 25 a person-affecting morality seems ill-equipped to handle questions of population ethics. 26 Instead, and for good reason, Parfit and others turn to impersonal moralities—moralities that are not limited to a consideration of what is good or bad for those affected by an act—with the hopes of constructing a plausible account. As I address below, both AU and TU are impersonal moralities, as are most hybrid accounts that include some features of AU and some features of TU. Before turning to impersonal accounts, however, it is worth noting that some authors do attempt to defend person-affecting accounts, and these authors argue that person-affecting 22 Parfit (1984, p. 359). 23 Some would deny that this case gives reason to reject person-affecting moralities, arguing that the second child would be worse off in the state of affairs where the girl does not wait to have a child. However, it seems odd to say that this child is harmed by not being brought into existence, and maintaining this position would require one to accept additional implausible positions as well. For further discussion, see Broome (2004). 24 If our intuition is that it is better for the girl to wait, then this is a counterexample to the "necessary condition" version of the person-affecting view. But it's not yet a counterexample to the "sufficient condition" version. The “sufficient condition” version can easily be shown to be implausible and does not even need elaborate cases. One only need to think of, say, a population of ten people who are happy and then the possible addition of a suffering person. According to a “sufficient condition” version of a person-affecting morality, it would be sufficient (for the purposes of concluding that the addition of the person was not bad) that no person who is in both states of affairs has been negatively affected. But that seems like an implausible result, and thus the “sufficient condition” version seems implausible. 25 See Roberts (2004). 26 See Holtug (2004) for a discussion of the difficulty—in fact, the impossibility—of constructing a plausible person- affecting morality. 28 accounts can indeed satisfactorily handle objections to person-affecting accounts that arise out of the non-identity problem. Among the authors who defend person-affecting accounts are, for example, Melinda Roberts, 27 David Boonin, 28 and Caspar Hare. 29 Despite the valiant attempts made by these authors, however, in my view, the resulting accounts that they espouse either (1) do not successfully handle the non-identity problem, or (2) successfully handle the non-identity problem but, in so doing, remove from their accounts the features of person-affecting accounts that are prima facie appealing and worth trying to preserve in the first place. Before continuing, however, I also add a few additional notes about person-affecting accounts and person-affecting intuitions (i.e. intuitions in favor of person-affecting accounts). First of all, it is unquestionably the case that most people do have person-affecting intuitions at least to some degree, and, for some people, these intuitions are very strong. I explore these intuitions in considerable depth in later chapters in the dissertation, in particular in Part III Chapter 1. In Part III Chapter 1, I offer and explain various hypotheses of mine regarding person-affecting intuitions. For one, I explain how our temptations to be influenced by person-affecting intuitions explain why, in certain cases, some might find AU implausible and TU more plausible, and I explain how this explanation helps support AU in the debate between AU and TU. Secondly, in Part III Chapter 1, I further explain and support my position that person-affecting intuitions, while both pervasive and powerful, are errors and are, at best, heuristics. I then also offer and explain my hypothesis about how and why we have developed person-affecting intuitions, and I explain that this is because person-affecting intuitions indeed are heuristics—heuristics for (impersonal) utilitarian goals. Not only are person-affecting intuitions heuristics for impersonal utilitarian goals, but I suggest that in typical cases (i.e. typical same-numbers cases (i.e. cases unlike the same 27 Roberts. 28 Boonin. 29 Hare. 29 number-case in Parfit’s 14-year old girl example)), they work extremely well as heuristics. It is only in the types of cases arising in the context of debates about population ethics (e.g., different- numbers cases and also some same-numbers cases, such as the one in Parfit’s 14-year old girl example) that person-affecting intuitions perform extremely poorly and fail to give us correct (or plausible) assessments or results. See Part III Chapter 1 for considerably greater detail regarding these points about person-affecting intuitions. Now returning to the discussion of the current section, however, at this point I have raised and explained a serious problem that afflicts person-affecting accounts. Thus, at this point, having set aside person-affecting accounts, I now turn exclusively to impersonal theories that have been offered as possible candidates for Theory X. Among the impersonal theories that I will discuss will be total utilitarianism (TU), average utilitarianism (AU), and a variety of impersonal views that are hybrids of TU and AU. 1.C.4. Total utilitarianism (TU) Total utilitarianism (TU) is an impersonal view, and, according to TU, the goodness of a state of affairs is equal to the total amount of happiness included in the state of affairs, aggregated summatively. Thus, the more total units (whatever this “unit” of measure ends up being) of happiness a state of affairs has, the better the state of affairs is, and, in comparing two or more states of affairs, the state of affairs that has the largest number of “units” of happiness, measured sum-aggregatively, is the best among the states of affairs being considered. It’s worth noting that TU is generally also considered to be “classical utilitarianism,” and it was the version of utilitarianism seemingly espoused by the various founding fathers of utilitarianism—e.g., Bentham and Mill. While, in most cases of the writings of these early utilitarians, it is not clear whether they indeed espoused TU over AU, because their writings do 30 not directly address cases where TU and AU in tension, their formulations of utilitarianism are generally made in ways that sound more like TU and which lead a reader to believe that these early utilitarians indeed would espouse TU if they were presented with different-numbers problems—i.e. cases requiring that they be more precise regarding whether they espouse TU, AU, or some other theory. There also are some additional indications that these early authors were proponents of TU. 30 In sum, while it’s not completely clear whether these early utilitarians were indeed proponents of TU—because of the fact that they did not consider situations involving different-numbers cases—there is some reason to think that they might have been proponents of TU. And, regardless of whether this is true or not, TU is generally referred to as classical utilitarianism. While there are various ways in which TU seems to be an appealing theory (all of which— to the extent I’m aware of them—will be discussed in this dissertation), there are a number of common objections raised against TU. I’ll briefly mention three of these here. (These three objections do not also afflict AU). What these objections, and seemingly most 31 objections against TU, have in common is that they show that we have strong intuitions in favor of the position that, at least in some cases, a person’s quality of life, 32 even if not the only thing that is relevant to the goodness of a state of affairs, is at least somewhat relevant to this determination. Thus, TU, which does not (in and of itself—i.e. all else equal) place any value on quality of life, does not seem to accord fully with our intuitions. 30 For one of the first known discussions of the difference between TU and AU, see Sidgwick's Methods of Ethics, Book IV, Chap. 1, §2 (esp. pp. 415-6). Incidentally, another crucial early discussion, which, along with Sidgwick, must have influenced Parfit deeply, was in Rawls's Theory of Justice (pp. 162f.). 31 Although most objections to TU ultimately boil down to showing (albeit in various different ways) that we have strong intuitions in favor of the position that, at least in some cases, a person’s quality of life is at least somewhat relevant to the determination of the goodness of a state of affairs, one exception to this is the argument that is the entire subject of Part IV of this dissertation: that TU is false because (1) TU, as a theory, requires the existence of non- arbitrary zero points, and (2) non-arbitrary zero points do not exist. 32 By “quality of life,” I mean the person’s level of welfare. 31 The Repugnant Conclusion: The Repugnant Conclusion was originally formulated by Parfit, and, in Parfit’s words, the Repugnant Conclusion is as follows: “For any possible population of at least ten billion people, all with a very high quality of life, there must be some much larger imaginable population whose existence, if other things are equal, would be better even though its members have lives that are barely worth living.” 33 As Parfit says, the Repugnant Conclusion is a conclusion that seemingly follows from the acceptance of TU. If one is a proponent of TU, one will seemingly be committed to accepting the Repugnant Conclusion. According to Parfit, however, the Repugnant Conclusion seems unpalatable or repugnant (hence the name that Parfit gives it), it seems implausible, and it is hard to accept to accept it. Thus, if the Repugnant Conclusion indeed is implausible, and if it indeed follows from TU, then the Repugnant Conclusion serves as a reductio against TU. Even if Parfit’s Repugnant Conclusion does not show that TU is false, it, at the very least, constitutes an important hurdle that a TU theorist must address and explain in his attempt to defend TU as a candidate for Theory X. In sum, the Repugnant Conclusion is typically thought to be one of the strongest reasons against espousing TU as the most plausible candidate for Theory X. Additionally, as stated earlier, much of the landscape in the population ethics literature can be organized and described in terms of what type of a response one gives to the Repugnant Conclusion and the issues that it raises. The Mere Addition Paradox: In addition to it being suggested and argued that the Repugnant Conclusion is a result that follows from TU, and thus which the TU theorist is seemingly committed to, there are also other, more general, arguments given for why a series of seemingly reasonable assumptions and assessments result in our being committed to the Repugnant Conclusion being true. One of these 33 Parfit, 1984, p.388. 32 more general arguments, introduced by Parfit, is entitled the “Mere Addition Paradox,” and I will describe it below. This paradox does not explicitly take TU (or any other theory) as a starting point, and it simply suggests that a progression of reasonable steps leads one to the Repugnant Conclusion. (This is unlike the Repugnant Conclusion argument itself, which is couched as a reductio against TU, and which thus takes the truth (or plausibility) of TU as its premise.) 34 Without further ado, Parfit’s Mere Addition Paradox is as follows: Consider three populations, A, A+, and B: In population A, everybody enjoys a very high quality of life. In population A+ there is one group of people as large as the group in A and with the same high quality of life. But A+ also contains a number of people with a somewhat lower quality of life. In Parfit’s terminology A+ is generated from A by “mere addition”. Comparing A and A+ it is reasonable to hold that A+ is better than A or, at least, not worse. The idea is that an addition of lives worth living cannot make a population worse. Consider the next population B with the same number of people as A+, all leading lives worth living and at an average welfare level slightly above the average in A+, but lower than the average in A. It is hard to deny that B is better than A+ since it is better in regard to both average welfare (and thus also total welfare) and equality. However, if A+ is at least not worse than A, and if B is better than A+, then B is also better than A given full comparability among populations (i.e., setting aside possible incomparabilities among populations). By parity of reasoning (scenario B+ and C, C+ etc.), we end up with a population Z in which all lives have a very low positive welfare. Thus, the final conclusion is that Z is better than A, which is the Repugnant Conclusion. By what apparently constitute sound steps of reasoning we have arrived at an absurd conclusion. 35 Thus, according to Parfit, the Mere Addition Paradox provides a strong reason for one to believe (or determine) that one is committed to the Repugnant Conclusion—regardless of whether one is a committed TU theorist (who seemingly would already be committed to the Repugnant 34 Thus, it’s important to note that, in considering the Mere Addition Paradox, as the reader will see, although the TU theorist seemingly is committed to accepting the progression in the Mere Addition Paradox and committed to accepting its conclusion, the Repugnant Conclusion, the AU theorist will simply deny various steps and will thus not accept the progression to the Repugnant Conclusion or be committed to the Repugnant Conclusion. Further, not only is the AU theorist not committed to the Repugnant Conclusion, but his view explicitly denies the Repugnant Conclusion, and it is clear that there is no progression that shows that an AU theorist is committed to Repugnant Conclusion. All that the Mere Addition Paradox arguably does to harm the AU theorist is provide some intuitive pull in favor of the incremental steps in the paradox plausibly being good whereas the AU theorist would say they are bad. Whether the AU theorist has a plausible account in denying that these steps in the Mere Addition paradox are good, though, is a more general question for AU and its plausibility that will be taken up elsewhere. The Mere Addition Paradox and its progression is not an argument that is meant to take aim at AU. It is important, however, to note this relationship between AU and the mere addition paradox. 35 Arrhenius, Gustaf, Ryberg, Jesper and Tännsjö, Torbjörn, “The Repugnant Conclusion,” The Stanford Encyclopedia of Philosophy (Spring 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2017/entries/repugnant-conclusion/>. 33 Conclusion) or whether one is not a committed TU theorist but someone who generally finds oneself persuaded by the reasoning in the Mere Addition Paradox. Further, since the Repugnant Conclusion is seemingly a reductio against TU, the Mere Addition Paradox also constitutes a seemingly strong argument against TU. Thus, the Mere Addition Paradox constitutes a strong argument against TU even if one was persuaded by its reasoning not because of a prior commitment to TU, but, rather, because one simply generally and independently found oneself persuaded by the reasoning of the Mere Addition Paradox. Even this latter route to the Repugnant Conclusion would provide reasons to reject TU. The Negative Repugnant Conclusion: Regardless of whether one feels committed to the Repugnant Conclusion as a result of being a committed proponent of TU, as a result of the Mere Addition Paradox, or for any other reason, someone who feels committed to the Repugnant Conclusion but who simultaneously finds it repugnant might seek ways in which he can show that he is not actually committed to the Repugnant Conclusion after all—despite it seeming prima facie that he was. One way in which one might try to avoid the Repugnant Conclusion is by setting the “neutral level for existence”— “lives above which increase the value of a distribution, and lives below which decrease this value—at a level that does not sound so terribly bleak, but, rather, at the level of a reasonably good life. 36 What to make of this type of a manoeuver, the issues it involves, and, even, the definitions of the terms and concepts just employed, will be deferred to Part IV of the dissertation, but the purpose of raising this maneuver here is that, even if this strategy were carried out, it would lead to a particularly severe version of what Broome calls the Negative Repugnant 36 The topics and issues described briefly here, regarding zero points and what constitutes a life of a “neutral level of existence”—“lives above which increase the value of a distribution, and lives below which decrease this value—will be addressed in much greater depth in Part IV. Indeed, the topic of zero points will be the entire subject of Part IV of the dissertation. Not only will the general topics addressed in the main text here be addressed in greater depth in Part IV, but, so too, the specific topics addressed in the main text here will also be addressed in greater depth in the context of the discussions in Part IV. The discussion in the main text here in Part I only briefly alludes to these issues and does not purport to delve into a discussion of them at this point. 34 Conclusion. 37 The Negative Repugnant Conclusion is just like the Repugnant Conclusion, but with all negative values. Thus, it states: “For any possible population of at least ten billion people, all with a very terrible quality of life (i.e. a strongly negative quality of life, for which it would be much much better to not be alive at all than to have this life), there must be some much larger imaginable population whose existence, if other things are equal, would be worse even though its members have lives that are only a hair under the level at which it is equally good to live or not live.” However, when one employs the strategy, discussed above, of raising the “neutral level for existence” to a level that is not so terribly bleak but to a level of a reasonably good life, one is then confronted with a much more severe version of the Negative Repugnant Conclusion, namely one that states: “For any distribution in which everyone leads a horrendous life, there is a worse distribution involving a much larger population in which everyone’s life is reasonably good. 38 This follows because the smaller population will still be living horrendous lives (though somewhat less horrendous than before the neutral level was moved up), but the larger population that consists of people just below the neutral level for existence will now be living lives just a hair worse than a life that is reasonably good—and thus these lives will themselves be reasonably good. The thought then is that this is especially repugnant, because it seemingly wrong that it could be worse to have a population consisting of reasonably good lives than to have a population consisting of horrendous lives. Thus, the idea then is that not only is the conclusion of the original Negative Repugnant Conclusion repugnant, but if one employs the strategy of raising the neutral level in an attempt to weaken the Repugnant Conclusion, this will lead to an especially severe 37 Depending how certain terms are defined, it’s also possible that the manoeuver described here could lead to what Arrhenius has called the Sadistic Conclusion—a “conclusion” that is typically raised as an objection against AU and against hybrid views that have at least some features of AU. I will introduce the Sadistic Conclusion below, in the section introducing some of the objections against AU. It will then get addressed in depth in Part III of the dissertation. 38 Broome, Weighing Lives, p. 213. 35 version of the Negative Repugnant Conclusion. Since, however, the Negative Repugnant Conclusion, in either of its forms, is something that a TU theorist seemingly is committed to endorsing (because of TU’s sum-aggregation feature), the Negative Repugnant Conclusion (in both of its forms) is a strong objection against TU. * * * * In sum, the foregoing objections (namely, the Repugnant Conclusion and its variants) constitute some of the most common reasons for doubting the plausibility of TU. (Further, these objections do not afflict AU. On the other hand, they seemingly provide support for AU.) 1.C.5. Average utilitarianism (AU) Another impersonal view that one might offer as a possible candidate for Theory X is average utilitarianism (AU). According to the common definition of AU, the goodness of a state of affairs is determined by the average happiness of the lives included in the state of affairs. I will address at length, though, both in section 2.C, and throughout this dissertation, that the above formulation of AU is not a good formulation of the AU theory, and this is for two reasons. First of all, the formulation above is too broad of a brush, and this is because there are a large number of different ways in which the above formulation can be precisified. Secondly, the above formulation is too narrow, because it excludes a number of theories that would appropriately be called average utilitarianism. I address both of these mistakes in a bit more detail in what follows, and, in so doing, I give a brief preview of the various types of AU accounts that exist. The following only mentions some of the various types of AU that exist, so as to just give the reader a flavor of the various versions of AU that exist. The following does not attempt to canvas all possibilities, but it mentions only a few. For a more in depth discussion of the various possible accounts of AU that exist, see Section 2.C. 36 First, with respect to the first mistake (i.e. that the above formulation of AU is stated with too broad of a brush): According to this formulation of AU, the goodness of a state of affairs is determined by the average happiness of the lives included in the state of affairs. This formulation is painted with an overbroad brush, because there are various different ways in which this formulation can be precisified. Further, and importantly, the further precision in articulating the above formulation is indeed crucial in many cases, because different formulations will, in many cases, provide different comparative assessments of states of affairs. The fact that the above formulation needs to be precisified is, in this way, very much akin to the way in which Theory X needs to be precisified when we are confronted with different-numbers cases. With same-numbers cases, various theories come out the same way, but when we encounter different-numbers cases, we realize that we must be more precise in our formulations. As for what the lack of precision is in the above formulation of AU, the lack of precision is that it doesn’t state how to determine the value of each individual life in the state of affairs. The literature on AU has in just about every case assumed one particular method for determining the value of each individual life for the purposes of AU, and this account is that the value of a life is to be determined by TU-type (i.e. sum-aggregative) aggregation. Thus, the value of a life is determined by adding up all of the units of happiness that the life includes. Thus, the resulting version of AU (which people seem to assume is the only version of AU) is what, in this dissertation, I will call Average Total Lifetime Value Utilitarianism (ATLU). ATLU is what almost all authors have in mind when they refer to AU, and it is thus this version of AU that almost all objections against AU are raised as objections against. (As I ultimately show, some of these objections only have their force against this version of AU or against a group of AU views (but not against all AU views). Other objections against AU, however, seemingly afflict all versions of AU. Making sense of how different versions of AU fare against different objections to 37 AU will be a topic of discussion throughout large portions of the dissertation (e.g., Part III in particular, but also Part II), and I will not further address this topic here. Notwithstanding the fact that I’m deferring this discussion, however, it’s useful to flag this topic as a topic that will be of key importance later in this dissertation.) While ATLU is one way of precisifying the AU formula that I provided above, it is not the only way of doing so. There are various ways that could be endorsed for determining the value of an individual life. One other such way would be to state that the value of an individual life is equal to that average happiness of each of the life’s component moments. Another way would be to use some account that is a hybrid between AU and TU (perhaps one of the ones discussed below, in Section 1.C.6, though Section 1.C.6 discusses aggregating in the population-wide context) to determine the value of an individual life. Regardless of which of these two options (or which of many other options) is chosen, the general formula of AU that I provided at the beginning of this subsection could then be used, giving the result that the goodness of the overall state of affairs is equal to the average of the goodness of each life that the state of affairs contains—regardless of how the value of an individual life is determined. The first problem with the formulation of AU that I gave was that it was overbroad and required further precisifying. The second problem is that the formulation is too narrow because it excludes a number of theories that would appropriately be called AU. The formulation of AU that I gave was that, according to AU, the goodness of a state of affairs is determined by the average happiness of the lives included in the state of affairs. There are other possible accounts that one could articulate, though, for which the goodness of the state of affairs is determined by averaging over values but for which these values are not the values of the individual lives included in the state of affairs. Numerous possibilities exist, but here I’ll mention just two. First, the goodness of the overall state of affairs could be arrived at by averaging over the 38 goodness of each conscious moment (i.e. each conscious moment that exists as part of a person’s life where that life is included in the overall state of affairs) that is included in the overall state of affairs. This possible view is a view that I will later consider and discuss at length (and which I ultimately will endorse as the most plausible account of Theory X), and I will call this view Average per Moment Utilitarianism (AMU). Second, one could articulate an account according to which the goodness of the overall state of affairs could be arrived at by averaging over the goodness of each temporal period (including all individual conscious moments occurring across all spatial locations at that moment in time) in that state of affairs. For example, one could maintain that the goodness of the overall state of affairs is equal to the average of the goodness of all temporal parts of an overall state of affairs, such as t1, t2, t3 etc. There could then be various different types of functions used for determining the goodness of that temporal part, given the various values of individual conscious moments that occur at that time. Regardless of how that question is resolved, the idea of averaging across temporal parts of a state of affairs is yet another way in which one could articulate an AU view. Having (1) described the way in which AU is typically formulated, as well as the more precise way in which it is typically understood (i.e. ATLU), and (2) described various other accounts of AU that can be articulated, I now proceed to mention five of the main objections raised against AU (i.e. raised against AU as a theory in general, though typically raised as objections aimed at ATLU, because, as I have stated, it is typically assumed that ATLU is the only version of AU (or at least the only version of AU worth considering etc.)). The five objections I describe here are (1) what I’ll call the “bottom up” objection, (2) the Sadistic Conclusion objection, (3) what I’ll call the “top down” objection, (4) the Two Hells objection, and (5) the Adam and Eve objection. 39 Despite the differences among these objections to AU, they all seem to be somewhat related to a single concern: Assuming that a particular action (e.g. bringing into existence a person with happiness level five) has no effects on others, why should the goodness of the action depend on anything extrinsic to those affected by the action (e.g. the happiness of the rest of society)? AU seems to dictate this conclusion, because the creation of an individual at level five would seemingly be bad if the prior population’s average happiness were level ten, but it would be good if the prior population’s average happiness were level three. This concern is manifested—with varying degrees of directness—in all five of the objections to AU listed above. Most directly, it is manifested in the bottom up objection and the Sadistic Conclusion objection. Additionally, although the Two Hells objection and the Adam and Eve objection have various features (some of which are somewhat unrelated to the above concern) they, too, manifest the above concern and it is one of the various driving factors in these objections. Lastly, the above concern is also manifested in the top down objection—albeit somewhat less directly than in the case of the other four objections. (1) The bottom up objection Versions of the bottom up argument are offered by Parfit and McMahan, among others. The following lays out the bottom up objection: In determining how to implement one’s theory of choice, it seems that one must decide which pool of individuals one should strive to maximize the happiness of: All sentient beings to ever exist? All people to ever exist? All people who currently exist? All people who exist currently or will exist in the future? It is unclear whether a non- arbitrary selection of any of these can be made. This is not a problem for TU, because with TU, it seems that any category of individuals—as long as it includes those who are “affected” by an action—will get the same result. In other words, no matter how many “unaffected” individuals one includes in the relevant pool of individuals that one is considering in one’s happiness calculation, 40 the same betterness assessments and prescriptions will result. 39 If one is employing AU, however, different choice of which population (of the various possible populations that include the “affected” individuals as well as somewhere from zero to all “unaffected” individuals) is the relevant population can lead to diverging betterness assessments and prescriptions. Additionally, it’s unclear whether a non-arbitrary selection of any of these possible populations can be made. Further, and the following is the key point for the bottom up argument against AU: It seems that there is no choice of a relevant temporal population that will yield plausible results for the AU theorist. The TU theorist, however, seemingly does not confront this problem. Thus, according to proponents of the bottom up objection, AU’s seeming inability to articulate a plausible account of a relevant population results in the determination that AU is an implausible theory. (2) The Sadistic Conclusion objection The Sadistic Conclusion objection, introduced by Arrhenius, is an objection that can be raised against AU and against hybrid theories that have at least some features of AU. The Sadistic Conclusion is as follows: “For any number of lives with any level of negative welfare, there are situations in which it would be better to add these lives to a population rather than to add some number of lives with positive welfare.” To see why this might be the case according to AU, consider the following: Suppose w1 contains ten people with a happiness level of ten. According to AU, the value of w1 is ten. Now, compare the results brought about by adding either one hundred people at level two versus adding one person at level negative ten. If we add to w1 the 39 As I say in Part III Chapter 1, those in the literature who articulate versions of this bottom up objection against AU do not seem to recognize the point that I just made: that even a TU theorist must articulate an account of what the relevant pool of individuals is that we are considering the happiness of in our calculations. The fact that TU confronts this question too (in addition to AU doing so) is certainly due to the fact that all populations, as long as they include the affected population, will bring about the same result and seemingly not bring about counterintuitive results. I address these topics in depth in Part III Chapter 1. It is important to point out there, though, that the very fact that a similar choice-of-population question confronts TU is something that those who argue for the bottom up objection do not recognize, and my recognizing that it the choice of the relevant population is a question confronting TU as a well (and thus a question that is distinct from and orthogonal to the question of what the appropriate aggregation function is for the relevant population) is a key first step in my articulation of a defense to the bottom up objection on behalf of the AU theorist. Thus, it is important to note that the point I make here is not one that is made in the literature and this point is not one that should be taken for granted as an obvious one. 41 one hundred people at level two, we will get a world, w2, the value of which, according to AU, is 2.73. If, however, we add to w1 the one person at level negative ten, we will get a world, w3, the value of which, according to AU, is 8.18. Thus, according to AU, this is a case where a world that is brought about by adding people with negative welfare is better than a world that is brought about by adding people with positive welfare. This is an implication of AU, but, according to the proponent of the Sadistic Conclusion objection, this is sadistic. Thus, the proponent of the objection argues that this is a reason to reject AU. 40 (3) The top down objection against AU and its implausible account of intrinsic value (if an AU account of intrinsic value even exists) There are a group of similar objections to AU that I call, collectively, the top down objection against AU. The top down objection characterizes the problem with AU as follows: While, according to utilitarianism, happiness is intrinsically valuable, AU seeks to maximize something else (namely average happiness), which is not intrinsically valuable, and often at the expense of maximizing total happiness, which is intrinsically valuable. Thus, according to this objection, AU fails because it doesn’t always maximize what is of intrinsic value for utilitarianism (namely happiness), and thus it is not a plausible utilitarian account (and, similarly, it will in various cases conflict with a utilitarian’s intuitions). Various authors, including Parfit, offer forms of this top down objection. Huemer, too, raises this concern with AU, although he frames it slightly differently. 41 He seeks to show that if happiness is what is of intrinsic value, then it follows that average happiness is irrelevant, and maximizing average happiness does not necessarily lead to maximizing total value. Thus, Huemer takes a step-by-step approach to show why TU follows if happiness is of intrinsic value. 40 Because the Sadistic Conclusion objection involves issues similar to those involved in the bottom up objection, I address the Sadistic Conclusion objection in the same chapter in which I address the bottom up objection—Part III Chapter 1. 41 Huemer (2008). 42 (4) The Two Hells objection Perhaps the objection to any version of AU that is regarded as the strongest is one that Parfit raised and termed Two Hells. 42 Among other things, it addresses AU’s treatment of suffering. The objection goes as follows: Consider two possible hells. In Hell 1, ten people exist and suffer great agony. In Hell 2, there are ten million people who suffer ever so slightly less agony. AU will say that Hell 2 is better. Parfit and countless others, however, find it obvious that Hell 2 is far and away the worse of the two states of affairs. (5) The Adam and Eve objection Parfit introduces the Adam and Eve objection as a reductio against AU. 43 It goes as follows: Suppose that Adam and Eve exist and have fantastic lives, say at a happiness level of 100. Suppose they can choose to either procreate or not procreate and that, if they procreate, their descendants—namely the human race—will also live fantastic lives, though their descendants lives will be ever so slightly less fantastic than the lives of Adam and Eve. Suppose that the lives of the descendants would be at a happiness level of 99. AU seems to be committed to the position that it would be better if Adam and Eve do not procreate, because doing so would lead to a state of affairs with lower average happiness than the state of affairs in which they do not procreate. According to Parfit and other proponents of the Adam and Eve objection, though, it seems that the state of affairs in which Adam and Eve procreate would be much much better than the one in which they do not. Even more strongly, though, it seems as though the state of affairs in which they do procreate certainly could not be worse than the state of affairs in which they procreate (or, to couch the objection in terms of action, it certainly could not be wrong for them to procreate when confronted with the choice between bringing about each of the two states of affairs just described. Thus, because AU seemingly is committed to a seemingly implausible position 42 Parfit (1984, p. 393). 43 Parfit, Reasons and Persons, page 401. 43 regarding the Adam and Eve examples, this example is thought to constitute a strong objection against AU. * * * * The foregoing objections are considered to be devastating objections against AU. As a result, AU is widely rejected as an implausible view. Since I will be defending a version of AU in this dissertation, I ultimately will defend AU against all five of these objections. 44 However, as I stated above in this section, these objections are raised against AU in general, but, since ATLU is the only account of AU that those raising objections against AU have in mind, these objections really can be viewed as being raised against ATLU. I return, later in the dissertation, to explore the different ways in which the different objections to AU affect different versions of AU, and, as I ultimately show, some of these objections only have their force against this version of AU or against a group of AU views (but not against all AU views). Other objections against AU, however, seemingly afflict all versions of AU. Making sense of how different versions of AU fare against different objections to AU will be a topic of discussion throughout large portions of the dissertation (e.g., Part III in particular, but also Part II), and I will not further address this topic here. Notwithstanding the fact that I’m deferring this discussion, however, it’s useful to flag this topic as a topic that will be of key importance later in this dissertation. 1.C.6. Hybrid theories In light of the fact that both AU and TU seemingly have devastating implications that seemingly render the theories untenable, many authors have espoused impersonal views that are 44 The bottom up objection is the subject of Part III Chapter 1, the Sadistic Conclusion objection also gets addressed in Part III Chapter 1 (since it turns out that it involves issues related to those that are involved in the bottom up objection), the top down objection is the subject of Part III Chapter 3, the Two Hells objection is the subject of Part III Chapter 2, and the Adam and Eve objection gets addressed both in Part III Chapter 1 and in Part I Chapter 2. 44 hybrids. These hybrid views can come in various forms, but, in many cases, they amount to views that are hybrids between AU and TU, having some features of AU and other features of TU. It’s not surprising that hybrid views of this sort have sprung up. After all, our commonsense intuitions in population cases typically do include some intuitions in favor of TU and some intuitions in favor of AU. In fact, most of the seemingly devastating objections against TU are powerful because they show that in certain situations we do indeed think that quality of life is relevant to a determination of the goodness of a state of affairs (and this is a feature that AU has but that TU does not have). Similarly, most of the seemingly devastating objections against AU are powerful because they show that in certain situations we do indeed think that the quantity of people (or time-slices of people, if we’re considering intrapersonal different-numbers cases) that exist and feel happiness is relevant to a determination of the goodness of a state of affairs. Various authors endorse hybrid views. Some of these authors, among others, are Hurka, 45 Parfit, 46 Ng, 47 Sider, 48 Blackorby, Bossert and Donaldson, 49 Fehige, 50 and Griffin. 51 Despite perhaps aligning more closely with our intuitions on the whole than do AU or TU, however, hybrid theories are still prima facie seemingly quite implausible. Hybrid views lack elegance, they seem arbitrary and ad hoc, and they leave us wondering what the underlying rationale really is for why the theory in question is true (or the most plausible theory)—even if they are perhaps more closely tailored to our actual intuitions in particular cases than are either AU or TU. This is how hybrid theories seem to me, and while others do not seem to raise these 45 Hurka, Thomas. (1983) “Value and Population Size,” Ethics, 93, pp. 496-507. 46 Parfit (1984). 47 Ng, Y-K. “What Should We Do About Future Generations? Impossibility of Parfit’s Theory X”, Economics and Philosophy 5 (2), pp. 235-253, 1989. 48 Sider, T. R. “Might Theory X Be a Theory of Diminishing Marginal Value?”, Analysis 51 (4), pp. 265-271, 1991. 49 Blackorby, C., Bossert, W., and Donaldson, D. “Intertemporal Population Ethics: Critical-Level Utilitarian Principles”, Econometrica 65, pp. 1303-1320, 1995. 50 Fehige, C. “A Pareto Principle for Possible People,” mimeo, Univ. of Pitssburg, Center for Philosophy of Science, 1992. 51 Griffin, J. Well-Being: Its Meaning, Measurement, and Moral Importance, Oxford: Clarendon Press, [1990] 1986. 45 concerns with hybrid theories, I imagine these are gripes that are doing a lot of the work (though certainly not necessarily all of the work) in implicitly preventing any non-advocate of hybrid views from endorsing a hybrid view. Here I have only briefly mentioned a potential problem that confronts hybrid views, but I return below, in Section 2.B, to discuss hybrid views and these difficulties with hybrid views, and to argue that there not only is a prima facie problem with hybrid views, but to argue affirmatively that, because of these problems, we should reject hybrid views and instead focus our efforts in searching for Theory X on impersonal views that are “pure”—i.e. non-hybrid. 1.C.7. Arrhenius and his “impossibility theorem” According to one philosopher, Gustaf Arrhenius, not only has no plausible account of Theory X been articulated, but no plausible account of Theory X even can be articulated, and he aims to show this by offering what he calls his “Impossibility Theorem.” 52 Arrhenius’s Impossibility Theorem states: “There is no population axiology which satisfies the Dominance, the Minimal Inequality Aversion, the Minimal Non-Extreme Priority, the Non-Sadism, and the Quality Addition Condition.” 53 Thus, in Arrhenius’s view, it is not possible to articulate a theory that simultaneously satisfies all of these desiderata. According to Arrhenius, his theorem leaves us with three options: “(1) to abandon some of the conditions on which the theorem is based; (2) to become moral skeptics; or (3) to try to explain away the significance of the proofs—alternatives 52 Arrhenius, G., 2000, Future Generations: A Challenge for Moral Theory, F.D. Dissertation, Uppsala: University Printers; Arrhenius, G., 2011, “The Impossibility of a Satisfactory Population Ethics,” in H. Colonius & E. Dzhafarov (eds.) 2011. 53 Arrhenius, G., “An Impossibility Theorem in Population Axiology with Weak Ordering Assumptions,” in Rysiek Sliwinski (ed.), Philosophical crumbs. Essays dedicated to Ann-Mari Henschen-Dahlquist on the occasion of her seventy-fifth birthday, Uppsala: Department of Philosophy, Uppsala University, 1999 (Uppsala philosophical studies 49), pp. 11-21. 46 which do not invite an easy choice.” 54 If, however, one does not agree with Arrhenius about the importance of each of his conditions, however, then one likely will not be moved by his conclusions. Notwithstanding this, it is notable that there are authors such as Arrhenius who indeed argue that indeed is impossible to articulate a plausible account of Theory X. Further, although Arrhenius states this position explicitly, it seems to be implicitly stated and believed by many, if not most, philosophers that it is impossible to satisfactorily articulate a candidate for Theory X—i.e. to articulate a single theory that one can fully throw one’s weight behind and fully sign on to. 1.C.8. Conclusion of sub-section 1.C In sum, at this point, it seems that the population ethics literature is in turmoil. It seems that there is currently no satisfactory solution in the literature. All attempted solutions seemingly are afflicted by serious problems that seemingly suggest that they are unsatisfactory solutions. Further, as I just said in the previous sub-section, Arrhenius argues that only is there no currently articulated account that is satisfactory, but there is no satisfactory solution that even can be articulated. Thus, this state of turmoil is where the population ethics literature currently stands. And this is the landscape in which (and the starting point from which) I begin my own exploration and my own arguments. I ultimately will show, however, that AU is the correct answer. I will show how confusions, unwarranted assumptions, and mistakes in the literature on population ethics are obstructing from our view the fact that the right answer lies right before us—in the form of AU. First, however, I provide a few more foundation-setting sections in the remainder of this 54 Arrhenius, Gustaf, Ryberg, Jesper and Tannsjo, Torbjorn, “The Repugnant Conclusion,” The Stanford Encyclopedia of Philosophy (Spring 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2017/entries/repugnant-conclusion/>. 47 chapter. 1.D. The broader importance of this philosophical problem in philosophy in general: It’s not just about “population ethics.” Population ethics cases are non-typical cases that force us to precisify our theories and get to the root of some big questions about value As I have shown, it seems as though all attempts to articulate a plausible candidate for Theory X have failed. It’s important to note, however, that Theory X and the attempts to articulate a plausible account of Theory X are of crucial importance not only to the field of population ethics, but also, and much more broadly, to normative ethics, ethics, and value theory on the whole. Even though it might seem as though the quest for Theory X is a project limited to the population ethics field, this is not the case. In seeking to articulate a plausible account of Theory X, what I am really doing is attempting to articulate a plausible account of welfare-based goods that is general in the sense that it applies not only to population ethics cases, but also to all cases. Thus, a successful articulation of a Theory X not only is a theory that can be applied in population ethics, but it is a theory that can be applied in all cases. As I’ve stated already in this chapter, what’s tricky about cases that population ethics confronts (i.e. different-numbers cases) is that they force us to precisify our prior account of welfare-based goods. It turns out that in different-numbers cases, various possible accounts of welfare-based goods (e.g., AU and TU, among others) can potentially give us different results, whereas in the more typical cases (i.e., same-numbers cases), we had not been forced to articulate precisely what we mean by “maximizing” welfare-based goods (or happiness). Thus, a successful Theory X is not just a theory of population ethics, but rather it is a complete theory of welfare-based goods (or, at least, a complete theory of the aggregation of welfare-based goods). Of course, it might only seem necessary that we have a plausible account of Theory X when we confront different-numbers cases, because in same-numbers cases different accounts of 48 theory X might have the same assessments and prescriptions. My point, though, is that even in same-numbers cases it is one’s underlying account of Theory X that explains one’s assessment or prescription in the same-numbers case. Even in same-numbers cases, it indeed is one’s account of Theory X that is doing the work in determining one’s assessment of a state of affairs, and it is not a more general notion of maximizing. It is one’s specific account of Theory X that is driving one’s assessment, even if one is not aware of that fact at that very instant. Thus, even if a particular theory (perhaps TU) is able to get the correct answer in the vast majority of cases (e.g., same-numbers cases), for a theory to be correct, it must get the correct answer in all cases—including different-numbers cases. Thus, the complete and correct theory of welfare goods will be the one that not only gets the easy cases right, but also gets the difficult cases right. Thus, in sum, a plausible Theory X is not just an account for difficult, and perhaps one might say obscure, extreme, or atypical, cases in the context of population ethics; a plausible Theory X is a full account of welfare-based goods in all contexts. In addition, as the reader will see in Part III Chapter 3, a key aspect of determining whether AU or TU is the more plausible view will involve assessing their respective accounts regarding what is of intrinsic value—or, said differently, what is the smallest component (which we could also refer to as a “unit” or “building block”) of value according to their respective accounts. This point provides further support for my point above—that the choice of a plausible account of Theory X truly is a theory that permeates value theory on the whole, and does not only apply to different-numbers problems. This point supports that earlier point because the determination of which account of Theory X one espouses ultimately results in extremely different underlying accounts of what type of things can have intrinsic value. Further, since an account of what types of things can have intrinsic value lies at the core of not only normative ethics but also ethics and value theory on the whole, it is clear that the choice of which account of Theory X to espouse has 49 far-reaching consequences for one’s whole understanding of value on the whole. Thus, this is actually an even broader point than my first point in this subsection. The first point was that one’s choice of an account of Theory X is actually a full account of welfare-based goods that is not only limited to cases in population ethics, but which also is one’s overall account of welfare-based goods. The second point is that one’s account of Theory X also cuts to the very core of value theory as a whole and the question of what type of things can have intrinsic value. This is an even broader implication than the first one. In the foregoing, I stated that what made one’s choice of one’s account of Theory X cut to the core of value theory was the fact that AU and TU have differing accounts of what types of things can have intrinsic value. Before continuing, it’s worth nothing that there is an additional way (albeit an indirectly related way) in which one’s selection of AU versus TU cuts to the core of questions in value theory generally. As we will see in Part IV of the dissertation, it turns out that AU and TU also have drastically different accounts and commitments regarding what they do say or can say about whether an empty world 55 has a defined value (and, if so, what this value can or must be), and what they do or can say about whether an empty world can be compared to a non- empty world. It turns out that whether one espouses AU or TU not only is a function of what one’s view is about what can be the bearer of intrinsic value, as discussed in the preceding paragraphs, but it ultimately is also a function of what one’s position is on these underlying issues regarding the value of the empty world 56 and vice versa (i.e. one’s position on these underlying issues might similarly be seen to be a function of one’s choice of whether to espouse AU or TU). In sum, the project of articulating a plausible account of Theory X is one that is much 55 The term “empty world” will be defined and explored in depth in Part IV, but, for the time being, it will suffice for me to say (somewhat question-beggingly) that an empty world is a world in which there are no things that are of the type that can be bearers of value (which for a utilitarian, or, more generally, for the purposes of this dissertation, are instances of moments of conscious experience). 56 And whether an empty world is comparable to a non-empty world. 50 much broader than a project in population ethics. It is a project that (1) pertains to all cases involving welfare-based goods, and (2) it also cuts to the very core of questions in value theory more generally. 1.E. The practical relevance of articulating a plausible account of Theory X An enormous number of important implications hang on what turns out to be the most plausible version of Theory X. Thus, since the dissertation’s goal is to argue for an account of Theory X, there are very important implications of the dissertation’s analysis. The implications are of various types and they arise in the context both of different-numbers problems in the population-wide (or interpersonal) context as well as in the context of single lives (the intrapersonal context). Within each of these contexts of valuing, there are, perhaps, at least three general ways in which the questions are confronted: in individual decision making, in governmental policy-making, and in legal rules—though these areas are perhaps overlapping and also perhaps do not exhaust the ways in which different-numbers problems get confronted. In this sub-section, I’ll briefly sketch what some of the implications are in each of these two domains. Furthermore, it’s also important to note that there are reasons to believe that, as technology continues to improve, the importance of determining a plausible account of Theory X will grow importance because there might be many new types of situations requiring decision making, as individuals and as a society, that involve different-numbers problems and our ability to make choices in these situations. Thus, it is all-the-more imperative that we articulate a plausible account of Theory X, and that we do sooner rather than later. Policy topics confronting different-numbers problems in population ethics abound and they are involved in questions comparing populations of smaller versus larger numbers of lives, questions comparing lives of shorter or longer lengths, and questions that simultaneously involve 51 both of these questions. Among the various areas that confront these questions are: where to put our medical and scientific research dollars (e.g., toward technologies and practices associated with saving lives of fetuses and infants, toward other life-saving technologies for adults, or perhaps, on the other hand, toward quality-of-life-enhancing technologies such as things like hip replacements); whether we should have laws or, more mildly, perhaps tax-incentives, that promote larger or smaller families; and, perhaps, questions about to what extent various resources should be conserved (assuming some type of relationship between resources and a population’s size and between resources and a population’s happiness). The various topics just listed were described in the context of questions regarding governmental policy, but analogous questions that raise these issues can also be confronted in the context of an individual person’s decision-making (be it in the context of one’s procreative decisions or in an analogue of the various other decisions discussed above that are confronted in the policy context). Further, these types of questions can also get confronted in the context of the law, where, for example, courts might need to determine if people should be compensated for having their expectable quantity of life being shortened as the result of a tort committed by another person, and, if so, how lost years should be valued, and, further, how a person should be compensated in the legal system for a loss of this sort. While the foregoing examples only scratch the surface in describing the ways we already confront decisions that implicate different-numbers problems and thus require us to employ (at least implicitly, if not explicitly) an account of Theory X in answering the questions, there are also reasons to think that technology will lead to an even greater need to have a plausible account of Theory X, because the cases in which we are confronted with decisions implicating different- numbers problems is likely to grow. One example of a new domain in which society is already starting to confront different- 52 numbers problems arises in the context of self-driving cars. The programming of self-driving cars forces us to answer questions about both how to compare states of affairs where different numbers of lives are lost (e.g., Should the car be programmed to avoid an accident that would kill two pedestrians by crashing into a wall that will kill one driver?), and also about how to compare lives of different lengths (e.g., Should the car be programmed to prioritize the lives of the elderly, middle-aged adults, or children? And how should the cars trade off accidents are likely to cause deaths versus accidents that are likely to not cause deaths but instead cause varying degrees of damages that will lead to disabilities and lower quality of life?) Thus, technologies like self-driving cars not only require us to confront different-numbers problems in the context of comparing lives of different lengths (thus forcing us, at least implicitly, to value the existence of more or fewer years of life), but also different-numbers problems in the context of populations of different sizes (thus forcing us, at least implicitly, to value the existence of more or fewer lives). Furthermore, as technology improves, we will likely continue to confront different- numbers problems more frequently, and this seemingly can be attributed to two main general reasons. First, we will develop greater capabilities in terms of predicting what the results of certain events will be (e.g., we will have a greater ability to know with more precision how many years of life (or how much happiness during the years of life the person does still have) a person will lose as a result of a tort that was inflicted on him, or, say, what the results will be of our using certain drugs to treat certain diseases). Second, we will also develop greater capabilities in terms of affecting the future. In conjunction, these two considerations likely will result in an increase in the prevalence of our confronting decisions implicating different-numbers problems—and thus our account of Theory X. In sum, even though different-numbers cases are not the typical case that we confront at every turn in personal decisions, governmental policies, or legal policies, and thus vast majority of 53 cases we confront are same-numbers cases, it nevertheless is the case that different-numbers problems pervade the questions that we confront, and this is the case in various ways. Further, the prevalence of our confronting decisions that implicate different-numbers cases is likely to grow. Thus, our quest to articulate a plausible account of Theory X is not only of key theoretical importance, but it also is of key practical importance. There is also an additional (and separate) point worth making here which is not directly related to the point about practical implications and applications, but the truth of which is evidenced by the fact that there are so many practical implications and applications of a decision regarding which account of Theory X to espouse: The question of which account of Theory X to espouse is not a definitional question. There is real meat to the question of which account of Theory X to espouse. Different answers prescribe different results. (And not only is that the case, but, as stated above, there are important real-world implications.) For any philosophical debate to be of interest, it would arguably be important that it not be a definitional debate. (I’m not saying that it would be important that it have real-world implications—I’m only saying that it would be important that it not be a merely definitional debate.) Thus, presumably in most philosophical debates that are had, there should hopefully be at least an argument for why the debate is not merely definitional. For some debates, however, there could be arguments made that the debate indeed does boil down to being merely definitional. I think that a very attractive aspect of the inquiry in this dissertation (i.e. not my work on the topic, but rather the topic itself) is that it seemingly is hard to argue that it is merely a definitional debate. It seems that there are key differences between the proponents of different accounts of Theory X, and these differences, among other things, amount to differences in views about which states of affairs are better than others, and also about which actions it would be better to carry out. Thus, the topic of inquiry in this dissertation is a meaty one that does not seem to be at risk of boiling down to a definitional 54 debate. While hopefully that is the case in most, if not all, philosophical debates, in some philosophical debates it is less clear that the debate does not boil down to a definitional debate. 1.F. Summary This chapter laid the foundation for this dissertation’s project. Having carried out all of these preliminary tasks, I now begin my quest to articulate a plausible account of Theory X. Part I Chapter 2 will be the first step in this process. ===================================================================== Part I Chapter 2: The prima facie appeal of AU 2.A. Introduction / background In Part I Chapter 1, I did a few things. I began by presenting the problem that the dissertation sets out to solve (articulating a plausible “Theory X”), and I explained how it is an extremely challenging problem that is of great theoretical interest—both because the question in population ethics is itself interesting, but also because it turns out to boil down to foundational questions underlying value theory. Thus, the importance of articulating a plausible Theory X is much broader than merely to population ethics, and it permeates value theory on the whole. Additionally, I explained the various ways in which these theoretical questions have numerous and extremely important practical implications that affect a whole range of issues governing various policy issues confronted by governments and society at large as well as a whole range of issues confronted by people in their daily personal lives. After providing these discussions, I clarified the various assumptions that I would be making in the dissertation and clarified which frameworks I would be working in. After describing and framing the problem that I’m setting out to solve, I then canvased the theoretical landscape, describing the various possible accounts that could be given as possible 55 answers to the problem of articulating a plausible Theory X. I also then described Arrhenius’s impossibility theorem and provided the prima facie reasons for rejecting each of the possible accounts of Theory X. As I stated and as I then thus showed, the population ethics literature sits in a state of turmoil, with seemingly no plausible answers to Theory X, and, at the very least, with no plausible defense currently articulated of any particular theory, even if one of the theories lying before us is indeed, after all, a plausible account of Theory X. With this context set, my dissertation’s task of articulating a plausible account of Theory X then began in earnest. Section 1.C began this task by arguing that, despite the seeming plausibility of person- affecting accounts, 57 a person-affecting account would not be a viable candidate for Theory X (in short, because of problems for person-affecting accounts posed by the non-identity problem). Thus, the previous section concluded that, in our search for a plausible account of Theory X, we must instead focus exclusively on impersonal accounts. 58 This, thus, is where we now stand in our search for Theory X: We reached one fork in the road, where one path was a “person-affecting” path and one was an “impersonal” path, and we took the “impersonal” path. The question now is which impersonal account to endorse. Many further forks lie ahead of us, however. 2.B. Why we must espouse a pure (non-hybrid) axiology and not a hybrid axiology 2.B.1. The task of this section The fork that currently confronts us is the choice of whether to adopt a “pure” theory or a “hybrid” theory, and this question is the question that is the topic of this section. I will discuss in further depth, in what follows, what these terms mean, but the short explanation is the following: A pure theory is one for which there is one thing that is of value. In the context of possible accounts of Theory X, there are two theories that appear to be pure theories: TU and AU. While it 57 See Chapter 1, Section 3, above, for definitions and explanations of what a “person-affecting account” is. 58 See Chapter 1, Section 3, above, for definitions and explanations of what an “impersonal account” is. 56 will be probed in further depth in later chapters (for example, Part II Chapter 1, and Part III Chapter 2) what exactly these theories say is of intrinsic value, TU basically says that a unit or instance of happiness is what’s of intrinsic value, and more of these are better than fewer of them (due either to a particular bearer of value having more units of value or due to there being more bearers of value—assuming that the additional bearer of value has positive value). AU, on the other hand, basically says that the average value among all bearers of value is what is of intrinsic value, and the higher this average is, the better the state of affairs. Further details certainly need to be ironed out, but both AU and TU thus seem, respectively, to articulate one thing that is of intrinsic value. For hybrid theories, however, it is not the case that there is just one thing that is of value. Hybrid theories have more than one thing that is of value, and, in almost all cases, these two things are the things that are of value, respectively, to TU and AU—total units of happiness (or value) and average happiness (or value) over all bearers of value. Different hybrid theories, however, are constructed in different ways, as will be discussed below. This section will proceed in the following way. I will begin, in Section 2.B.2, by explaining what’s at stake in Section 2.B, and by explaining how this section fits into and relates to the rest of the dissertation on the whole. Then, in Section 2.B.3, I will (a) explain why authors seemingly are driven to articulate hybrid theories, and (b) sketch out a few types of hybrid theories and a few examples of these theories offered in the recent past by prominent theorists. Then, in Section 2.B.4, I will begin my argument against the plausibility of hybrid theories, and my argument will be aimed not at any specific hybrid theory, but against hybrid theories in general. This argument will have a few different pieces and also a few different sub-pieces. Section 2.B.4 will explain why it would be ideal, or desirable (all else equal), for a candidate for Theory X to be pure and not hybrid. Section 2.B.5 will then argue that hybrid theories are not plausible. Section 57 2.B.6 will then attempt to provide a proof 59 for the claim that Theory X is a pure theory and not a hybrid theory. Lastly, in Section 2.B.7, I will briefly conclude. 2.B.2. What’s at stake in this section, and how this section fits into and relates to the rest of the dissertation What’s at stake in this section? A lot. As I have shown and will continue to show, it is very difficult to come up with a plausible account for Theory X, and this is because we simultaneously hold many conflicting intuitions. As a result of the fact that we hold so many conflicting intuitions, and the fact that we hold many of these intuitions very strongly, it might seem to many that hybrid views allow us to get the best of various worlds: Adopting the various parts of different theories that seem plausible and rejecting the various parts of different theories that seem implausible seemingly might lead to a hybrid theory that not only best characterize a person’s intuitions, but it might thus constitute the most plausible possible theory. Further, as I will show, hybrid theories not only might seem to best characterize what we typically do believe and what might seemingly constitute the most plausible theory, but it is also the case that many of the theories recently offered as possible accounts of Theory X are indeed hybrid theories. Despite these considerations and these trends, I think that hybrid theories are implausible and that they must be rejected. I think, instead, that the key resulting debate is the debate between the two pure theories, AU and TU. Thus, the entire remainder of the dissertation, after this section, which amounts to just about the entire dissertation, will focus exclusively on whether AU or TU is the most plausible account of Theory X. In other words, after this section, the only theories that will be in play are AU and TU. I ultimately defend a version of AU as the most plausible account of Theory X, but I do think that one reasonably, even after considerable and careful thought, could 59 “Proof” is arguably too strong of a word here (even if the argument that I’m referring to as a “proof” is completely successful (which it might not be)), but I will continue to refer to this attempted argument of mine as an attempt to provide a “proof.” 58 espouse TU. I do not, however, think that one could reasonably, after considerable and careful thought, espouse a view other than AU or TU. Thus, from my perspective, while I wholeheartedly will endorse AU as the most plausible Theory X, there is no competition for what is the second- most plausible theory. TU wins that prize, and by a long shot. Thus, despite the fact that the entire remainder of the dissertation is arguing against TU, it still remains the case that I think TU is the only theory that can be reasonably seen as competition to AU, and, if I were to be convinced that AU were false, I would espouse TU. The reason that the vast majority of the dissertation is devoted to arguing for AU over TU is a testament to the fact that TU is a plausible candidate for Theory X, and that the choice between AU and TU is not an easy one. Such long and intricate discussions and inquiries would not be required to compare AU’s merits against any other competitor. TU is a worthy competitor. Having said this, however, the debate between AU and TU often requires a reader to accept unpalatable implications of one’s theory of choice (be it AU or TU). In light of this, it is absolutely crucial that before one gets to all of the chapters in this dissertation about the TU- versus-AU debate, that one is firmly convinced that AU and TU are indeed the only options that we can reasonably consider. In other words, by the time we begin our debate between AU and TU, we must be firmly convinced that hybrid theories are not viable and that any other theories, be they person-affecting theories or some other option, are not viable either. In order for the debates between AU and TU to have traction and hold water, we must, by the time they start, be firmly locked into the position that the most plausible position is indeed locked in the room containing AU and TU, and that we can no longer look to other options. If we were not convinced of this, the debate between AU and TU would slip out from underneath us and other possibilities would creep back in. Further, at times during the AU versus TU debate, it will indeed be very tempting to try to let other options come back into the room for consideration, because, at times, the going will be 59 tough. Both AU and TU have various implications that are seriously counterintuitive. We cannot afford to have a possible response to these difficulties and counterintuitive implications of AU or TU, when they crop up in future chapters, be to open up again the possibility that a hybrid answer is the correct answer. We must close this possibility now. Once and for all, so that this option remains closed for good. Thus, after this section, I will not further discuss hybrid theories at all in any of the remaining portions of the dissertation. Thus, the current section’s success is absolutely imperative for the success of the dissertation. It must be shown conclusively here that hybrid theories are not tenable, and that the options remaining for us to consider are only AU and TU. After having completed this section, I will then begin the task that will occupy the remainder of the dissertation—exploring and assessing the debate between AU and TU. Before continuing, however, it is important to also be clear about what is and is not necessary to the success of this section. What is necessary is that I convince the reader that hybrid theories are implausible and that the most plausible account of Theory X is not a hybrid theory. What’s not crucial for the success of this section, though, is the success of my arguments that there is a proof that Theory X is not a hybrid theory. In this section I do make such an argument, and I do so because I believe it to be correct. But this section will be successful, for the purpose of the dissertation, even if one does not agree with what I say about there being a proof that Theory X is not a hybrid theory. What’s key for the purposes of the dissertation is not that there is a proof that Theory X is not a hybrid theory or that it would be impossible for a hybrid theory to be Theory X, but, rather, just that the reader come away believing that hybrid theories are implausible and that they are (likely) not a plausible account of Theory X, and that (likely) the correct account of Theory X is either AU or TU. 60 2.B.3. The motivation for espousing a hybrid view and examples of some hybrid views 2.B.3.a. The motivation for espousing a hybrid view As I’ve stated, it is difficult to articulate a single plausible account of Theory X. This is because we have many conflicting intuitions that make any given possible account seem implausible due to its having unpalatable implications—often when a particular theory is applied to “extreme” cases. In light of the fact that no single theory seems to satisfy all of our intuitions, many theorists have defended hybrid accounts of Theory X. A hybrid account is an account according to which more than one thing is of intrinsic value. Thus, for example, a hybrid account might be a function according to which the overall goodness of a state of affairs is a function of both AU’s assessment of a state of affairs and TU’s assessment of a state of affairs. Hybrid views of this sort could take various forms, but what they have in common is that at least some weight is given to at least two inputs, and the overall goodness is a function of these two inputs. At least prima facie, these accounts might seem attractive, because they attempt to pick and choose features from the various non-hybrid theories, thus keeping features of theories that seem attractive and getting rid of features of theories that seem unpalatable. These mix and match theories thus are viewed by many authors as providing the best of both worlds, where the worlds are AU and TU. These hybrid attempts seem to be have some prima facie plausibility, because they do attempt to better match our intuitions. As a matter of fact, it does seem as though people’s prima facie intuitions in various possible cases may, as a matter of fact, map onto hybrid theories more closely than they do onto AU or TU. As a matter of practice, it seems to be a fact that people’s commonsense intuitions do reflect a basket of intuitions including different strands of different pure theories (namely giving some weight to TU considerations (i.e. having some features of TU) and giving some weight to AU considerations (i.e. having some features of AU), though not all 61 strands of either of these pure theories). For these various reasons, it is worth exploring whether a hybrid view can be plausible. 2.B.3.b. Examples of a few types of hybrid theories, and a few specific hybrid theories that have been offered recently The vast majority of theorists who offer possible accounts of Theory X these days seem to be offering hybrid theories. These hybrid theories, even when the two inputs being used are just AU and TU, can take various forms. For example, one form that these hybrid theories can take is the form of H = AU + TU, thus giving weight to both AU and TU values. Or, similarly, it could take the form of H = AU * TU, a function that while also giving weight to both AU and TU values, would bring about a particularly low value if any one of the two inputs is low (i.e. the overall result would, to a greater extent than in the addition function, be kept lower by having the one low value). Alternatively, one could have a function that tells us to ignore certain data. For example, one could have a function for whose AU or TU component, the function says to only consider the happiness of those below a certain happiness level, or perhaps a function that says to not consider the happiness level of people at a certain level of happiness beyond the first one million people at that level. (Theories of this sort could take as inputs aspects of AU and of TU, but some might only take as inputs one or the other, with the constraints on which subset of the population is to be considered etc. While this might not technically be a hybrid theory in the sense that it takes as an input, in some form, either only AU data or only TU data, I’ll consider functions of this sort together with hybrid theories. This is because different theories of this sort are variations of AU and TU that share some of one of these theories’ features but alter the way in which this data is used (often by applying either AU or TU to some subset of the relevant population).) Yet another way in which a theory might be hybrid would be to have a “piece-wise” 62 function, which articulates different functions for different types of cases. The overall function could be thus described as conjunction of the different functions for the different circumstances. The foregoing types of cases aren’t fully distinct and the categories are overlapping. These categories do, however, provide a bit of a map of what the different types of hybrid functions are. As for who has articulated hybrid theories, some of the theorists to do so have been Temkin, Parfit, and Hurka. Hurka has articulated views, similar to the H = TU * AU view mentioned in the first category, above, which yield bad results if either input yields a very low result, thus guarding against some extreme results of AU or TU in their pure forms. 60 Parfit has considered a number of views falling into the second and third categories above (i.e. the views considering only some of the data, and the piece-wise views). One view that Parfit considers is a “critical level view.” 61 According to critical level views, which are a modified form of TU, a person’s life contributes positively to a population’s value only if the quality of the person’s life is above a particular positive critical level. The value that lives below this level contribute to the value of a population is neutral (i.e., zero). Variants of critical level views are often introduced in order to avoid the Repugnant Conclusion. Another type of view that Parfit considers is a variant of TU according to which there is an upper limit to the total value realized by members of a population experiencing positive happiness but where there is no upper limit to the total negative value realized by members of a population experiencing negative happiness. 62 (Parfit ultimately does not espouse this view, however, because it implies what he calls the “Absurd Conclusion.” 63 ) 2.B.4. Why it’s ideal to have a non-hybrid theory This sub-section does not argue that Theory X is non-hybrid, but it argues that it would be 60 Hurka, T., 1983, “Value and Population Size,” Ethics, 93: 496-507. 61 Parfit (1984); see also Broome 2004; Blackorby et al. 1997, 2004, 2005. 62 Parfit (1984). 63 Ibid. 63 ideal, or desirable, if it were non-hybrid. In other words, all else equal, it would be better if Theory X were non-hybrid. Articulating and espousing a theory in ethics typically involves a process of revising one’s theory to fit one’s intuitions and revising one’s intuitions to fit one’s theory, and continuing to do these two things until one arrives at what Rawls called a reflective equilibrium. While this process is specific to constructing ethical theories, the construction of ethical theories and the construction of scientific theories share many commonalities. Both involve various data points and then the attempt to articulate a theory (extrapolating from the data points) that both explains the data points and predicts and explains additional data points. Unlike the scientific domain, the domain of ethics seems to involve the feature, described above, where we can revise some of our intuitions (i.e. data points) during the process of coming up with an ethical theory (and, indeed, this seems to actually constitute part of the process). Despite this apparent difference, however, there are methodological questions involved in theory making that the scientific and ethical domains share. One feature of theories in both the scientific and ethical domains (and in other domains as well), is that, all else equal, theories seemingly are more plausible the simpler they are. The simpler the theory, the more likely it is to be correct. And, similarly, the more convoluted a theory is, the more likely it is to seem ad hoc, and thus be less plausible. At the extreme, a “convoluted” theory would be a conjunction of all the specific data points that one has. This would be (1) an unhelpful theory, because it would not provide any predictive information, because it wouldn’t at all generalize away from the data points that one has, and (2) an implausible theory, because it would seem ad hoc. Moving slightly away from this extreme, we might have a theory that does generalize somewhat and could thus have some predictive power regarding additional data points, but it might seemingly be ad hoc because of how closely tailored it is to the specific data points and because of a failure to offer a more general principle that truly explains the data points that are 64 observed. In short, the simpler a theory, the more plausible it tends to be. Of course, there is a tradeoff, however. A theory is more plausible the simpler or more elegant it is. But a theory is also more plausible, the more it successfully accounts for our data points. An elegant theory is not plausible if it predicts data that are contrary to the data that we observe. Thus, a trade-off is at work in determining how simple of a theory to endorse. Thus, although simplicity or elegance of a theory is not the only factor that contributes to whether a theory is plausible, it is a significant factor. Thus, at the very least, it is the case that, all else equal (all else equal in the sense of how well the theory meshes with the data points that we do in fact have), a simpler theory is more plausible than a convoluted theory. Thus, if possible, it would be good if our ethical theory were simple and elegant, and thus non-hybrid. This principle militates against hybrid views and in favor of pure views, such as AU and TU. Section 2.B.5. Continuing off of the previous points about the desirability and simplicity of theories, these points cast the plausibility of hybrid views in doubt. With hybrid views, we are left wondering why the view is what it is, and what explains this theory being a true theory. This question will always exist, because no matter how simple a theory is, there will still be the underlying question of why it is the way it is and what the underlying explanation is, but this question seems more problematic the more convoluted the theory is. Further, it seems that any proponent of hybrid theories likely would not deny these points. In other words, a proponent of a hybrid theory likely would concede that a lack of simplicity is a strike against his account. Nevertheless, proponents of hybrid theories likely would say that they espouse their hybrid theories because, while the lack of simplicity is a cost associated with their 65 theory, the benefit is that their theory is able to correctly account for more of our intuitions (i.e. their theory is able to account for more of the data points). Thus, proponents of hybrid theories endorse these theories because they think that the simpler theories do not get enough of the data points right. Notwithstanding their endorsing the hybrid theories, however, even these theorists would concede that it would be better, all else equal, if one’s theory were simpler. 64 Particularly in the context of candidates for Theory X, hybrid accounts seem implausible. It seems utterly arbitrary to suggest that, say, only the data regarding a subset of a population (and not the whole population) would be relevant. Similarly, piece-wise functions seem arbitrary and it seems unexplainable why such a function could be plausible. It’s unclear in these cases what the underlying principle would be. In physics, for example, we have a workable theory for sub-atomic particles and we have a different theory for larger objects. We are not satisfied with the existence of these two parallel theories, however, and we seek a single unifying theory—a theory of everything—that can stand as the broader principle explaining the two specific theories that we currently have. String theory, among other recent theories, is a candidate for the unifying theory. In sum, it seems that while a hybrid theory is not obviously false, it does seem as though, to be plausible, it would require at least some sort of further explanation motivating it. This type of motivation seems to be mostly absent in hybrid accounts that have been offered. Defenders of hybrid accounts would likely state, in their defense, that the motivation and explanation is clear: 64 During the course of all of the dissertation chapters about the AU versus TU debate, I will, among other things, be showing the reader that AU does indeed correctly account for our intuitions even in cases where one might have thought that it did not. Similarly, in other cases, I will show why and how intuitions that appear to not be consistent with AU are or were mistaken. Thus, these types of points I make later on will support this section’s conclusion that hybrid theories are not the most plausible solutions in our search for Theory X. This is because I will show that a simpler theory (indeed, a pure theory), once we wipe away confusions, is able to account for all of our data points. Thus, with a simpler theory available that explains the data, we need not resort to hybrid and seemingly ad hoc theories. The fact that the simpler theory correctly explains the data, I will argue, has been obscured due to philosophical confusions, and once we strip away these confusions, it will be clear that a simple theory (AU) correctly explains the data, and thus, due to its simplicity and its correct explanations, this theory is the most plausible account of Theory X. 66 namely that it matches our intuitions in cases better than do other theories. In light of the admitted trade-off that exists between simplicity of a theory and its ability to match our data, this response by proponents of hybrid theories results in a bit of an impasse, and their position at this point seems difficult to attack with a knockdown argument. Notwithstanding this, one can come away from the debate with the impression that hybrid views are implausible, and this could be enough to reject them. Section 2.B.6. Not only are hybrid theories implausible, but I think that I can show that a hybrid account cannot be the Theory X that we are seeking. Let’s, as an example, consider a hybrid view H, according to which the goodness of a state of affairs is equal to TU’s assessment multiplied by AU’s assessment. For short, we could say H = AU * TU. (Of course we could imagine all sorts of other hybrid functions where H is a function of AU and TU. Among these options could be (1) H = AU + TU, (2) H = AU(squared) + TU, or perhaps (3) (AU of the worse-off half of the population) + TU. All sorts of variants exist. But, to keep the argument simple, for the time being, I will focus just on the function H = AU * TU. Similar points would be able to be made, mutatis mutandis, for the other hybrid theories as well.) The point made in the prior sub-section was that it would seem arbitrary to endorse a theory where H = AU * TU. 65 Relatedly, the point in the section on this “arbitrariness” point is that it’s unclear how one would come to decide that the value that one seeks to maximize is AU * TU. There seems to be a need for an explanation, and the explanation is lacking. Here, however, I leave behind these “arbitrariness” / “how to decide” points. 65 With this equation, if both AU and TU were negative, we would get the result that H is positive, which seems utterly implausible. A version of this formula could be employed, though, that would avoid this odd feature but otherwise get the same results. 67 Here, I’ll take as a given that the correct account of Theory X is H = AU * TU. I think, though, that even if this is the case, we can “purify” or “non-hybridize” this function. Consider, for example, that when a person, or government, or whatever the deciding entity may be, is making a decision in accordance with this function (H = AU * TU), this person will be striving to maximize H, the overall output. This is how a pluralistic system works. Regardless of what the inputs are and how the tradeoff between the input values works, there is an output, and one seeks to maximize the output. 66 What this means, though, is that even though this theory claims that AU and TU are both of intrinsic value, what in reality is the case is that neither AU nor TU is of intrinsic value. 67 Rather, the amount of value in the state of affairs is directly tied to something else—the H value (i.e. the output value). Thus, while either the AU value or TU value might contribute to the H value, it is only the H value that itself is of intrinsic value. Only H’s changes are reflected directly in the amount of overall value according to this function. 68 In response to this, one might say that it is assuming my conclusion to say that H actually is a monistic (or pure) system, because only a monistic system could have the result that the thing of value varies directly with (is equal to) the amount of value. Thus, a pluralist / proponent of a hybrid theory would deny this and say specifically that AU and TU values can still be of intrinsic 66 Of course, one could argue that we’re operating with vague inputs and it’s not clear exactly what the function is, and thus the output is vague and it’s vague how to get from the inputs to the outputs. This, I believe, is what was said by pluralists such as the intuitionists in the monist-pluralist literature involving Ross et al. I need not further address this “vagueness” argument, though. This is because the hybrid functions I’m considering are all articulated in precise terms, and it is these that I am addressing. Further, I would argue that the “vague” theories can all be put in precise terms. Further, we can use revealed preference (of a person or of a government) to construct a precise theory even if one denies that such a precise theory exists. The theory could be made, at least at first, by constructing a conjunction of all observed data points. 67 Although I’m using the term “intrinsic value” here, I only mean it to be understood in a loose sense. Later in the dissertation (in Part III Chapter 3), I will provide an in-depth inquiry into the notion of intrinsic value. I will define and explain my understanding of the term at that point. For now (in Part I), however, I mean for “intrinsic value” to simply be understood in a loose sense. 68 The topics discussed here regarding the plausibility of hybrid functions will also be discussed in further depth (albeit in the context of the debate between two non-hybrid theories, AU and TU) in Part III Chapter 3 (the chapter on the top down argument). Additionally, this material will be closely related to discussions related to arguments made by Huemer and questions regarding the “unit relevant for ethical inquiry” and questions of whether the intrinsic value of two different conscious moments (tokens) must have the same intrinsic value if their intrinsic and non-evaluative properties are identical. 68 value even if they do not necessarily vary directly with the amount of output value. In this sense, it is true that my argument assumes its conclusion. The following, however, is another way to articulate my point: Taking as our function H = AU * TU, we can simply rename H as happiness or the thing of value that we originally were stating was being quantified by AU’s or TU’s value-determinations. The reason that this works is that when referring to the happiness (or value, more generally) that is quantified with the AU and TU metrics, we have not made it clear exactly what we are referring to or how it is measured. Thus, we can deny that that the input (i.e. that which is quantified by AU or TU) has any intrinsic value. One might object, here, that it does not make sense that I could “simply rename H as happiness or thing of value that we originally were stating was being quantified by AU’s or TU’s value-determinations.” After all, it seems as though the facts are the facts and I can’t just rename something as “happiness” or “the thing of value”; what has happiness or value seemingly should be independent of any “renaming” I carry out. Or, to be more precise, though, the objection isn’t that I can’t rename it (because there’s nothing stopping anyone from naming anything anything he or she wants), but the objection is that a renaming could only amount to a definitional manoeuver, and thus my renaming could not actually do the work that I seem to be purporting that it does. These points are well taken, but, if I clarify my point, I think I’m able to avoid this objection. What I mean to say (again, exploring the theory H = AU * TU) is not that that AU and TU were of intrinsic value and that my “renaming” makes it the case that H is what’s of intrinsic value and that AU and TU no longer are of intrinsic value. Rather, by carrying out the “renaming,” what I mean is that if (H = AU * TU) is our theory, then it never was the case that AU or TU were of intrinsic value. Rather, all along they lacked intrinsic value because it is H that actually is what is of intrinsic value. 69 Thus, it seems that when there’s an allegedly pluralistic function, such as (H = AU * TU), it must be the case that the overall output, H, is of intrinsic value. If this were not of intrinsic value, we would not be seeking to maximize this. Instead, we would be seeking to maximize something else, whatever that might be. Thus, by definition, the overall output of the function is what has intrinsic value. But, if that’s the case, then it can’t be that the inputs have intrinsic value. Thus, in sum, it seems to me that the foregoing points show that a hybrid account cannot be the Theory X that we are seeking. 69 2.B.7. Conclusion regarding my rejection of hybrid theories In sum, the conclusion of this section is that hybrid theories are not plausible candidates for Theory X, and thus, in carrying out the inquiry into candidates for Theory X, we should reject hybrid theories. It’s important to reiterate the fact that I believe that I have supported this conclusion even if the reader does not accept my argument for the position that a hybrid account cannot be the Theory X we are seeking. I do not need that argument to succeed for my section on hybrid accounts to be successful. (If my “proof” argument succeeds, that would only be icing on the 69 Various points can be made in response to my argument that a hybrid account cannot be the Theory X that we are seeking. As I’ve stated, the main response that one might have is to maintain that my argument assumes its conclusion (despite my valiant attempt to have it not assume its conclusion). There are other points that an opponent of mine might raise in response to my argument, however. One such point is as follows: Perhaps, one might say, the points in my argument prove too much (given that I ultimately will be arguing in this dissertation for AMU as the most plausible account of Theory X): It could be that my points might support not only the claim that Theory X must be a non-hybrid function, but also the claim that it might have to be a sum-aggregative function (e.g., TU). This is a very interesting point to raise. It relies on the idea that it seems that whatever it is that is of intrinsic value, more of it seemingly would have more value, and this, one might think, amounts to a tenet of TU but not AU. I address this very question (albeit not in the context of hybrid theories) in Part III Chapter 3 (the chapter on the top down argument) when I’m comparing the relative plausibilities of AU’s and TU’s accounts of intrinsic value. For the reasons I provide in that chapter, I do not think that the concern raised in this footnote comes to fruition. In other words, I do not think that my “proof” that hybrid accounts cannot be the Theory X that we are seeking also proves that Theory X must be sum-aggregative. I think that the points I make in my “proof” are also compatible with AU. My discussion in Part III Chapter 3 explains why this is. Notwithstanding this, I think that the question I raise in this footnote is an interesting one that does indeed merit being raised and considered—even though, as I just stated, the concern ultimately does not succeed. 70 cake.) I first argued why it would be ideal for Theory X to not be a hybrid account, and I then argued why hybrid accounts would be implausible as candidates for Theory X. This section will be successful, for the purpose of the dissertation, even if one does not agree with my “proof” that Theory X is not a hybrid theory and thus my claim that it is impossible for a hybrid theory to be Theory X. What’s key for the purposes of the dissertation is just that the reader come away believing that hybrid theories are implausible and that they are (likely) not a plausible account of Theory X, and that (likely) the correct account of Theory X is either AU or TU. In my view, I have reached this threshold. Having now shown that hybrid theories are not plausible candidates for Theory X, we can now, thus, limit our focus to non-hybrid theories. As a result, we are left with either AU or TU, since these are the two non-hybrid options. The rest of this dissertation proceeds to explore whether AU or TU is the more plausible theory. Also, as I will show, AU turns out to be a family of different views (with their differences, in part, being a function of their having different answers to the question of what is the “unit relevant for ethical inquiry” (a term that I coin and define in Part II Chapter 1)). Thus, assessing the relative plausibility of AU and TU will also involve articulating and considering the plausibility of the various sub-versions of AU. 2.C. The question of AU versus TU: The prima facie case for AU over TU 2.C.1. The Main Intuition Motivating Average Utilitarianism The main intuition motivating AU—and distinguishing it from TU—is the idea that while we want people to be as happy as possible, we do not necessarily have reason to bring into existence additional happy people. In other words, when it comes to happy people, more is not better. This is what Jan Narveson has in mind when he writes: “[T]here is no moral argument at 71 issue here. How large a population you like is purely a matter of taste, except in cases where a larger population would, due to indirect effects, be happier than the first . . .” 70 Further, Narveson explains that it is not good that people exist “because their lives contain happiness.” Instead, “happiness is good because it is good for people.” 71 Similarly, according to David Boonin-Vail, we should aim “to produce more happiness for people, not to produce more people for happiness.” 72 This is straightforward enough, and it might strike many as plausible. This intuition, however, is consistent with at least two different views, which I will now discuss: First, the view that endorses what John Broome calls the “neutrality intuition,” and second, AU. 73 C.3.a The Neutrality Intuition The thought that a larger population of happy people is not necessarily better than a smaller population of happy people has led some authors, including Narveson, to adopt what Broome calls the “neutrality intuition.” Narveson appeals to this intuition when he further clarifies his position described above: “We are in favor of making people happy, but neutral about making happy people.” 74 In other words, not only does Narveson think that the size of population is a non- moral matter, but he thinks that adding an additional person to a population is something that we should generally consider to be an ethically neutral event. It is neither good nor bad. Narveson himself, however, holds an asymmetrical neutrality intuition: He thinks that the neutrality does not apply to individuals whose lives are comprised of net suffering—he thinks the addition of such a person would not be neutral, but a bad thing. Broome further qualifies what he takes to be the commonly-held neutrality intuition: He suggests that there also might be a sufficiently high level 70 Narveson (1967, p. 68). 71 Parfit (1984, p. 394). 72 Boonin-Vail (1996, p. 268). 73 Broome (2004). 74 Narveson (1973, p. 73). 72 of happiness in a potential person that would result in the creation of this person being not neutral, but a good thing. Nevertheless, the thought is that there exists at least a range of happiness levels, which, if possessed by an additional person, would render the creation of that person a morally neutral event. As Broome points out, however, although the neutrality intuition might be appealing, it can be shown to imply a contradiction, 75 suggesting that it is in fact incoherent. Consider the three following possible worlds, in Example 1. (Each number represents the happiness level 76 of an individual in the world. The “x” represents the fact that there is no individual who exists in this slot.) Example 1: 77 A = (1, 1, x) B = (1, 1, 1) C = (1, 1, 2) According to the neutrality intuition, A is equally as good as B, and A is equally as good as C, yet C is better than B (since it is agreed that it is a good thing, all else equal, to maximize the happiness of members existing in a population). This appears to be a devastating implication of the neutrality intuition. One might, however, try to avoid this implication by espousing a slightly different version of the neutrality intuition—one grounded in the notion of conditional good. According to a notion of conditional good, if X is good conditional on Y, this means that (Y and X) is better than (Y and ~X), but that neither (Y and X) nor (Y and ~X) is better than, worse than, or equally as good as either (~Y and X) or (~Y and ~X). In the context of our topic, 75 “Contradiction” might be too strong of a word here. A better choice of words might be to say that there is an “intransitivity.” If, however, I assume that goodness is transitive, as I do, then perhaps the intransitivity would imply that there is a contradiction. 76 Throughout the paper, reference will be made to “happiness levels” and “quantities of value.” While there are many interesting and important issues surrounding how to determine what happiness level an individual has and what these numbers actually represent, this paper will attempt to bracket these issues. I will simply make the seemingly plausible assumption that sense can be made of numerical values representing people’s happiness. 77 This is a very slight variation of an example offered by Broome. Broome (2004, p. 146). 73 the idea is that the goodness of the third person’s happiness is conditional on his existence. This avoids the contradiction discussed above, because this account denies that A is equally as good as either B or C—in addition to claiming that A is neither better nor worse than either B or C. Unfortunately, however, despite avoiding the contradiction, conditional good does not help us much. It fails to give us a way to make comparisons between worlds with different numbers of people. While this might not be deemed particularly problematic in the case of Example 1—in fact, it illustrates the commonly held neutrality intuition—it seems more problematic if the individuals who exist in both worlds are not equally well off. In other words, it’s not clear what the conditional good account would say about a comparison between A = (1, 1, x) and D = (1, 2, 1). Further still, the inability to make comparisons between worlds with different numbers of people is particularly problematic if we attempt to compare worlds that do not have individuals assigned to particular slots—for example, if none of the individuals in one world exists in the other, and vice versa. Consider a comparison between worlds E and F, where E includes a “4” and a “5,” and F includes a “4,” a “6,” and a “1.” Depending on how the neutrality intuition account addresses the comparison between A and D, above, two equally unattractive responses could be given with respect to the comparison between E and F. Either the assessment of which world is better would seemingly depend on which value, for calculation purposes, is put in the third slot, 78 or no comparison at all can be made since there are different numbers of individuals in each world. 78 It does not seem as though the arbitrary choice of how to order individuals should have an effect on the betterness ranking of two worlds, and in cases such as the comparison between E and F, it seems that it would. If the conditional good account allows for the comparison between E and F, it seemingly would say that (4, 5) is better than (4, 1, 6), but worse than (4, 6, 1). One might think that if, among the arbitrary orderings of individuals within each world, there are orderings that provide different prescriptions, then perhaps this indicates that the two worlds are equally good. However, not only does this inference seem to not cohere with the intuitions that result in the seeming contradiction, but there would be extreme cases that could be constructed that would make this rule seem implausible. For example, consider a comparison between A: (2) and B: (9, 9, 9, 9, 9, and 1.99). One of the arbitrary orderings of B would put 1.99 in the first slot, and since this means that there would be at least one ordering for a prescription in each direction, the result, on this rule, would then be that A is equally as good as B. 74 Neither of these replies is attractive, however, and the latter can be shown to be particularly absurd in cases where all individuals in the populous world are drastically better off than those in the less populous world. There are various other ways of attempting to defend the neutrality intuition, but they all are afflicted by serious difficulties. 79 The underlying problem seems to be that the neutrality intuition is a component of a person-affecting morality, and while there are situations that make it seem like a plausible account of population ethics, it is unable to provide a theory that can satisfactorily address the vast majority of cases. C.3.b. Average Utilitarianism Despite the fact that the neutrality intuition appears to be false and that the notion of conditional good fails to vindicate it, this does not mean that we are forced to accept the claim that a greater quantity of happy people, all else equal, is better than a lesser quantity of happy people. In fact, although Narveson’s and Boonin-Vail’s quotations explicitly and implicitly endorse person-affecting moralities, and although the anti-more-is-better intuition will likely be appealing to proponents of person-affecting moralities, the anti-more-is-better intuition does not rely on a person-affecting morality. Even if we reject person-affecting moralities and the two related accounts discussed above, we can still largely maintain the core intuition that happiness is only a good conditional on a person’s existence. We can do this by appealing to AU. Consider Example 1: If we assess goodness of a state of affairs by the degree to which it maximizes average happiness, we get a similar result in the sense that the non-existence of an additional happy (third) person in World A does not necessarily count against its value. However, 79 Among the approaches that have been offered are views that (1) deny the transitivity of “equally as good as,” (2) maintain that goodness is relative and not absolute, (3) deny that betterness is fully determinate, and (4) no longer try to find an axiological account of the neutrality intuition. See Broome (2004, p. 149) for further discussion of the notion of conditional good, and of the failure of these other approaches. 75 using AU, we have guidance about which of the three worlds to choose, because there is no intransitivity or other difficulty that results during the comparison between worlds. Each world can be attributed a fixed value that can be compared with the values of other worlds. This is the case even when it’s not determined which particular person is occupying the third slot. AU is not, however, simply an improved and more coherent version of a position embracing the neutrality intuition. In fact, despite their similarities, AU rejects the neutrality intuition. 80 Consider Example 1: A position embracing the neutrality intuition would say that A is equally as good as (or incomparable with) both B and C. AU, on the other hand, will say that C is better than A because the third person raises the average happiness of C above the average happiness of A. AU will still say that A is equally as good as B because the third person does not raise or lower the average happiness of B in relation to that of A. Nevertheless, AU allows us to maintain the intuition that a greater quantity of happy people, all else equal, is not better than a lesser quantity of happy people. It is important to note, at this point, that while I think that many of those who reject AU will hold the anti-more-is-better intuition but then abandon AU because of one or more of the three objections that will be discussed in Part III of this dissertation, there are also people who do not hold the anti-more-is better intuition. Thus, to persuade someone in this camp of the truth of AU (over, perhaps, TU), one would need to do more than show that AU is the theory best able to maintain the anti-more-is-better intuition and that the three objections to AU can be met. Additional support for AU would be needed, because, at best, the above might render one agnostic between, perhaps, AU and TU. This support exists, however. While perhaps TU is the first version of utilitarianism that one might think of, it is confronted with serious objections—namely the Repugnant Conclusion—and the same is true for other theories. Thus, this alone gives us reason to 80 In fact, the rejection of the neutrality intuition is what enables AU to be coherent in the ways discussed in the previous paragraph—ways in which a position embracing the neutrality intuition would not be coherent. 76 look for a different utilitarian theory, and this is enough reason for a non-holder of the anti-more- is-better intuition to take interest in this paper’s project. 81 In addition to there being some readers who do not share the anti-more-is-better intuition (and thus, whom the previous paragraph speaks to), there are also some might be some readers who not only do not share the anti-more-is-better intuition, but who in fact hold a more-is-better intuition. What do I say to people like this? First of all, what seems most certainly to be the case is that there will be people who feel the pull of the more-is-better intuition, although while some might hold this intuition consistently across all cases, many people (and, as I will argue in Part III, most) who feel the pull of the more-is-better intuition will only feel it in a smaller number of particular cases—such as cases like the one in the Adam and Eve example. Regardless of how broadly these people will hold the more-is-better intuition, though, what is my response to these people? I have a number of different points to make in response, and I make these points in my section on the Adam and Eve objection (Section 1.F.2 of Part III Chapter 1). Before continuing, though, it is useful to clarify the dialectic here and the difference between how the Adam and Eve objection and the other arguments against AU that I address in Part III relate to the structure of my dissertation’s argument: The Adam and Eve objection elicits intuitions contrary to AU, but it does so in a way that is different from just about every single other objection to AU that I consider in this dissertation. The Adam and eve objection elicits intuitions that are contrary to AMBP (the (population-wide) anti-more-is-better intuition), and AMBP’s role in my arguments for AU is to play a large role in what I call the “prima facie case” 81 One might think that the “anti-more-is-better intuition”—the foundation of my argument for AU—begs the question in a debate with a proponent of TU. In a sense this is true, but in a harmless sense. As I say in the text, I think that the majority of proponents of TU are prima facie sympathetic to the anti-more-is-better intuition, but that they then abandon AU because they are persuaded by one or more of the main objections to AU that I will discuss. Thus, in this sense, I am not begging the question, because the target of my arguments is someone who would buy my premise— and not a committed TU proponent. As I mention in this paragraph, though, my argument does not rest on the anti- more-is-better intuition, and it could entice a reader to believe AU even if he does not share this intuition prima facie. This is because there are prima facie difficulties with other views—such as the Repugnant Conclusion, which is a difficulty for TU. 77 for AU, which I’m offering here in Part I Chapter 2. Just about all of the other objections to AU that I consider in the dissertation are not objections to the prima facie case for AU, but, rather, they are objections to AU that attempt to show why AU is implausible even if one is attracted to the prima facie case for AU. Thus, these other objections, for the most part, do not (directly) cast doubt on AMBP itself. Section 2.D. As I have alluded to above, in Section 1.C.5, the vast majority 82 of the population ethics literature seems to assume that AU is a single theory, and, in particular, in referring to AU, they seemingly assume a specific version of AU that I call ATLU. In defining AU, the literature typically states that, according to AU, the goodness of a state of affairs is determined by the average happiness of the lives included in the state of affairs. As I stated in Section 1.C.5, though, this is both too narrow and too broad of a definition: This definition excludes many theories that are appropriately called AU theories, and it also is too broad, because there are various different theories (which are very different) that would all fit the definition listed above. In particular, the literature, when giving the above definition of AU, seems to mean to refer to ATLU, but this is not the only type of AU theory that would fit the above definition. In this subsection I elaborate a bit on the possible versions of AU that exist, and I set the stage for the inquiry that will be the focus of Part II of the dissertation. Note that, in what follows, I will use various terms that I will not define precisely, but these terms will be defined precisely in Part II. For the time being, I will merely use these terms loosely to sketch a picture for the reader—a picture that will be clarified and precisified in Part II. 82 A notable exception to this is Thomas Hurka, who discusses the existence of various types of versions of AU in his articles “Average Utilitarianisms,” Analysis, Vol. 42, No. 2 (Mar. 1982), pp 65-69, and “More Average Utilitarianisms,” Analysis, Vol. 42, No. 3 (Jun. 1982), pp. 115-19. 78 In making sense of the landscape among AU views, there are various possible forks in the road, and I will address these in order. The first step in thinking about AU is to correctly state what it is that all AU views have in common, and that is the following: For something to be an AU view, it must be that the goodness of a state of affairs is equal to the average value of all of the instances, within the state of affairs, of what I call the “unit relevant for ethical inquiry”—or “the thing to be averaged over.” Various things can be the unit relevant for ethical inquiry, and, for example, for ATLU, the unit relevant for ethical inquiry is individual’s lives. Of course, if AU is to assess the value of lives, it must be that there is some way of doing so, though this method is something that can differ among different versions of AU. What AU theories have in common, though, is that the there is something that I will call a “building block,” which is the smallest thing that has value, and the value of lives are in one way or other a function of the values of all of its component building blocks. According to AU (and also according to TU), the building blocks are conscious moments (of happiness or unhappiness of varying degrees). There are various forks in the road, however, in defining one’s particular version of AU. First, one must select what one takes as one’s unit relevant for ethical inquiry. For ATLU, the relevant unit is a person’s life. Other possibilities exist, however. One option is for the building blocks themselves—individual conscious moments—to be the unit relevant for ethical inquiry. Alternatively, one might hold that the relevant unit is a moment in time, say, t1, whose value is a function of the values of all conscious moments in any spatial location that occur at that time. Various other options exist as well. One’s unit relevant for ethical inquiry could be a country, a city, babies between the age of six months and nine months, or seemingly any grouping of 79 conscious moments, regardless of how seemingly arbitrary it might be. 83 The second step in precisifying one’s AU theory is to determine the function (be it an aggregation function or some other sort of function) for determining the value of instances of the unit relevant for ethical inquiry in light of the various component building blocks that compose one of these instances of the unit relevant for ethical inquiry. Among other options, one might use TU-type aggregation or AU-type aggregation to get the value of the instance of the unit relevant for ethical inquiry, given its parts. The third step is slightly different from the first two, in that it overlays the previous steps rather than continuing the progression of digging more deeply in detail (as the previous steps progressively did). The third step asks what “reference class”—be it a temporal population or perhaps some other type of population not defined temporally—is relevant in assessing a state of affairs. In other words, are there some instances of the unit relevant for ethical inquiry to not be included? One possible account would be that we should consider, say, lives (if that is one’s unit 83 Ultimately, in this dissertation, I will be arguing for AMU, a view according to which the unit relevant for ethical inquiry is the conscious moment. In Part II, I will offer various arguments for why AMU is the most plausible version of AU, and, thus, why the conscious moment is the most plausible option among the various possibilities for what should constitute the unit relevant for ethical inquiry. Although the following is not one of my main arguments for why the conscious moment should be the unit relevant for ethical inquiry (though it is related in certain ways to my main arguments), it is important to note that it seems as though any of the other possibilities, including those just described in the previous paragraph in the main text, are seemingly indefensibly arbitrary. It seems as though any selection of an intermediate grouping of conscious moments (i.e. intermediate between, on the one hand, the whole state of affairs, and, on the other hand, the building block conscious moments) is without a solid justification. This is particularly apparent with some of the examples I gave, but my view is that even ATLU’s choice of an individual’s life as the unit relevant for ethical inquiry is arbitrary. For example, why should conscious moments be grouped into the lives of individuals but not into the group of moments that exist at any one moment of time? It seems that there are an endless number of ways in which the world’s conscious moments can be sliced into groups, and each of them are similarly just breaking the conscious moments into groups without a strong justification. Indeed it seems to me that the very fact that there are so many ways in which one can slice up the field of conscious moments into groupings of various sorts—be it across time, space, or lives, or whatever else—and the very fact that all of them are possible ways in which the conscious moments can be grouped and which can constitute a descriptive statement of some sort about the world is itself a strong reason to think that there is no non-arbitrary reason to privilege any one of these slicings over any other slicing. The view according to which the unit relevant for ethical inquiry is the conscious moment is the only view regarding the unit relevant for ethical inquiry that does not suffer from the objection that it is arbitrary. This itself is an extremely strong reason to be sympathetic to AMU. It’s unclear what the moral relevance is of any of the other slicings, and it’s unclear that a plausible explanation about the moral relevant of the slicing can be provided. The comments in this footnote are, however, of course only the beginning of an argument—and not a whole argument. This topic will be elaborated upon and addressed in depth in Part II. 80 relevant for ethical inquiry) that exist at any time, past, present, or future. 84 Another account would be to only consider those in the present. A third would be to consider all lives that either currently exist or will exist in the future. Additional forks in the road might exist as well, but these are the main three forks in the road that confront an AU theorist in his attempt to precisify his specific AU account. As a result of the various possible answers to each question and the various ways in which these answers can be mixed and matched resulting in accounts that are different from one another, there are a large number of different AU accounts that one can articulate (and thus attempt to defend as the most plausible candidate for Theory X). * * * * Not only is it a gross oversight to fail to recognize 85 that there are numerous versions of AU, but, as I will argue at various points in the dissertation, the assumption that ATLU is the only version of AU (or, more weakly, the most plausible version of AU) is in many ways responsible for many of the misconceptions regarding AU, and for what, in my view, is a severely premature rejection of AU. 86 84 Unless we are considering a context where we’re interested in evaluating actions, however, and terms like “past,” “present,” and “future,” are defined in reference to the time of an action, it’s not clear what time is being picked out by these terms. Thus, when I use these types of terms, I will be typically using them in reference to a particular action. 85 Again, I remind the reader that I recognize that those who I accuse of failing to recognize that ATLU is not the only account of AU, could, instead of failing to recognize this, instead be simply coming to the conclusion that ATLU is the only version of AU that is even worth considering because perhaps, in their view, all other versions of AU are even more implausible. This is possible, and might even be partly true, but in light of there being various ways in which AU accounts other than ATLU might have more success than does ATLU at responding to some of the objections leveled against ATLU, it seems that those who argue against AU are likely failing to recognize that other AU accounts exist, or, at the very least, are rejecting non-ATLU AU accounts very prematurely. Even if one didn’t think that other AU accounts had a fighting chance, I think that in various situations, the mere consideration of how they would fare in response to an objection to AU would be extremely illuminating and fruitful, because it would help show what, specifically is an attractive or unattractive feature of an account, rather than lumping together four or five different features (i.e. the features the conjunction of which constitute ATLU) and not knowing what it is precisely about ATLU that is leading to counterintuitive implications or the prima facie inability to provide a plausible result in a particular situation. 86 I return in Part III Chapter 1 to further address this point of mine, and, in that chapter, I also discuss various ways in which similar points can be made regarding mistaken assumptions that authors make regarding the different temporal reference classes that can be employed by AU. The mistake that I identify as being made in the context of the choice 81 * * * * At this point, having sketched out the various ways in which one can define one’s account while still remaining within the umbrella of AU accounts, it should be clear that, even though I have offered a prima facie argument in favor of AU in this chapter, and even though I have stated that I ultimately will endorse AMU, I have not yet offered an explanation of why AMU is the version of AU that I espouse. It is to this task that I now turn. The goal of Part II will be to argue that AMU is the most plausible version of AU. In carrying out this argument, I will, for the purposes of Part II, simply assume the truth of AU, and only consider which version of AU is version that one should espouse if one espouses AU. In Parts III and IV, I will then remove the stipulation that AU is true, and I will, at that point, reopen the inquiry into whether AU (the umbrella theory or one of its subversions) is the most plausible candidate for Theory X. Thus, having at this point made a prima facie argument in favor of the plausibility of AU, Part II begins its inquiry (into which version of AU is the most plausible sub-version of AU) by addressing what I called the first fork in the road in determining which version of AU one endorses: by answering the question of what one espouses as the unit relevant for ethical inquiry. 2.E. Summary Chapter 2 has provided a prima facie argument for AU. As I have stated, however, there is a big difference between a prima facie argument for AU and a complete argument for the plausibility of AU (the umbrella theory or one of its sub-versions). Notwithstanding this, this chapter constitutes the beginnings of an argument for AU, and, at the very least, it provides initial reasons to take seriously the importance of my project. After all, if there were not even a prima facie reason to find AU plausible, it likely would not seem fruitful to carry out an in-depth of reference class is somewhat different from the mistake I’m identifying here, but both categories of mistake have some features in common. 82 exploration into whether AU is plausible (and, similarly, it then would not seem fruitful to carry out an in-depth exploration into which sub-version of AU is most plausible). Thus, at this point, the stage has been set, and the dissertation proceeds to carry out what will turn out to be a defense of AMU. Part I Conclusion At this point I have (1) laid the foundation for this dissertation’s project, and (2) provided a prima facie argument for AU. I now proceed to carry out the project and to explain why AMU is the most plausible candidate for Theory X. Parfit, the author who coined the term, “Theory X,” and who introduced us to the problem of articulating Theory X and its various difficulties, counterintuitivities, its paradoxes, and its tensions, did not think that he had satisfactorily solved the riddle and articulated a plausible account of Theory X. I argue that this dissertation accomplishes what Parfit was unable to accomplish: In my view, this dissertation solves the riddle and articulates a plausible account of Theory X. In short, the answer is AMU. ===================================================================== Part II: Reasons to believe that AMU is the most plausible version of AU Introduction to Part II Having argued elsewhere for the plausibility of average utilitarianism (having provided initial appeal for it in Chapter 2, and defending it against objections in Part III), we are confronted with the question of what the relevant units are that should be averaged over. While the analogous question could theoretically arise for other theories, such as TU (where the question would be which units to take the additive sum of), in practice it would not, because the question would be a moot one for a theory like TU. The different sub-theories of TU would be coextensive. This, 83 however, is not the case with AU, and thus there is a further question of which sub-version of AU is the most plausible. Part II addresses this question. A quick exploration into the matter at the end of Part I Chapter 2 suggested that ATLU is a possible candidate. It seemingly allows us to combine two core intuitions: ATLU maintains the population-size anti-more-is-better intuition (AMBP) (that, all else equal, it is not better that more people exist than fewer people) and it also maintains the intrapersonal maximizing intuition (IMI) (that, all else equal, it is better agent-neutrally that a happy person live longer). As such, ATLU relies on the claim that, all else equal, it is not better for there to be more good conscious moments if they exist in an additional person, but it is better for there to be more good conscious moments if they exist in an already existing person. A question that confronts this view, however, is why the different location for additional good conscious moments is relevant. Part II will argue that the location for additional good conscious moments is not relevant, and thus, taking AU as given, it will argue that AMU—and not ATLU—is the most plausible version of AU. Part II Chapter 1 will offer a positive argument for why AMU is the most plausible version of AU. Part II Chapter 2 will then consider various objections to AMU. Most importantly, Part II Chapter 2 will explain why IMI, as stated, is false, and thus why AMU is true despite our IMI- friendly intuitions. Part II Chapter 3 will consider and address the question of whether death is bad for the person who dies, and, more broadly, it will consider and address what the most plausible account is of goodness for a person (i.e., agent-relative goodness). In so doing, Part II Chapter 3 will provide significant further support for the main conclusion of Part II Chapter 2—that IMI is false. Thus, to the extent that Part II Chapter 3 provides support for the claim that IMI is false, Part II Chapter 3 will provide strong support for AMU. Lastly, Part II Chapter 3 will also consider and address some additional objections to AMU. A few additional broader points to keep in mind: 84 First: In Part III of the dissertation, it is noted that while I think that person-affecting intuitions are mere heuristics for (impersonal) utilitarian goals and thus that they should be rejected when in tension with the utilitarian goals themselves, in practice they end up giving us the same prescriptions in most cases. A related point will be made and addressed in this Part: Despite the fact that AMU and ATLU appear to provide very different prescriptions, their prescriptions in practice will not diverge as much as one might think. In the vast majority of actual cases, the two will converge. This is because AMU will take as given the intuitions that people have, which are often, in many respects, aligned with ATLU, as opposed to with AMU. These ATLU-friendly intuitions would thus affect people’s actual experiences (and similarly, conscious moments’ experiences) in many cases, and thus any theory would need to take into consideration, as inputs, the empirical facts about the happiness or unhappiness that actually is brought about by certain actions. In other words, if people in fact have ATLU-friendly intuitions, then in many—if not most—cases, carrying out the prescriptions of ATLU would maximize the utility of people and of conscious moments, and, thus, carrying out the prescriptions of ATLU will be instrumentally valuable according to AMU. 87 Carrying out the prescriptions of ATLU, in most cases, will be the best instrument for carrying out the prescriptions of AMU—the theory that I actually endorse as being the most plausible. Thus, only in fairly circumscribed cases will there be a divergence between the prescriptions of AMU and ATLU. The above-described way in which AMU relates, in practice, to ATLU is akin to utilitarianism’s accounts of various non-utilitarian intuitions: A utilitarian might say that (1) his theory actually agrees with non-utilitarian intuitions in “non-extreme” cases (not only because non-utilitarian intuitions might be helpful heuristics for maximizing utility, but also because, in non-extreme cases, acting in accordance with one’s non-utilitarian intuitions might itself bring 87 The difference is, though, that ATLU interprets these intuitions as correct, while AMU interprets them as strictly false approximations to a truth that obtains partly because we have those very intuitions. 85 about utility and thus further utilitarian prescriptions), and then say that (2) in the “extreme” cases, non-utilitarians are mistakenly influenced by the heuristics that work well for the “non-extreme” cases but which (a) do not work well as heuristics and maximize utility in extreme cases, and (b) do not bring about sufficient utility from acting in accordance with these intuitions (even when we take this utility as given and accept it as a fact for the purposes of calculating happiness levels) that it would be optimific to follow the non-utilitarian prescriptions in these extreme cases. Second: The conclusions of this Part will, at first glance, be counterintuitive to many readers. After all, we are deeply accustomed to thinking of person-data 88 as of paramount importance ethically. As I argued in Part I Chapter 2, however, it seems clear that we must reject person-affecting intuitions, and these were deeply rooted intuitions as well. The move that occurs in this Part is a different move than was the rejection of person-affecting intuitions, but I think the conclusions of this Part will be less difficult to handle now that we have rejected person-affecting intuitions. In a sense, the rejecting of person-affecting intuitions was the first step of a process of which this Part is the second step. ===================================================================== Part II Chapter 1: A positive argument for why AMU is the most plausible version of AU (and for why the conscious moment is the “unit relevant for ethical inquiry”) 1.1. Introduction The task at hand in this dissertation is one of aggregation, and it’s important to be clear about what aggregation is and what it involves. As Larry Temkin explains at the beginning of his endeavor to offer a plausible account of aggregation in Rethinking the Good, Aggregation is the process or method of combining different parts into a whole. In ethics, there are principles of aggregation for assessing individual lives. Basically, these tell us how the different elements or components of individual lives combine to affect the goodness, comparative quality, or desirability of those lives. Likewise, there are principles of aggregation for assessing outcomes. 88 By “person-data,” I mean facts about which conscious moments are parts of which persons. 86 Basically, these tell us how the different elements or components of outcomes combine to affect the goodness, comparative quality, or desirability of those outcomes. 89 Thus, it should be clear that there are two key processes involved in articulating a plausible account of aggregation. First, we must be able to identify the elements or components that constitute the relevant parts of the whole in question. Second, we must be able to articulate a function that explains how to plausibly get from the goodness of the parts to the goodness of the whole. The fact that aggregation is composed of these two different processes should not come as a surprise, and all theorists who offer theories of aggregation at least implicitly recognize this. This notwithstanding, the failure to explicitly recognize that the process of identifying the relevant components is a process separable from—and indeed prior to—the process of articulating how to combine these parts, can lead to mistakes in theorizing. The second process—that of how to combine the relevant components—is of great interest, and there are various ways in which one could get from the “elements relevant to the goodness of lives or outcomes,” to use Temkin’s phrase, to the assessment of the whole. Among the possibilities, one might sum up components in an additive manner, one might average over the components, one might use weighted sums (as the prioritarian does), one might implement maximin, or perhaps one might implement some form of maximax (as the perfectionist might). Other options abound. These questions will be taken up in the majority of this dissertation, but the current chapter will largely leave this combination question aside, and instead focus primarily on the prior question of identifying the relevant components. As Temkin says, with respect to this first question, it is almost always claimed that the components relevant to the goodness of a life are the “different moments of a life.” 90 Thus, for a utilitarian, the elements relevant for the goodness of a life would be the “‘local’ utilities possessed 89 Temkin, p24. 90 Ibid. 87 by that individual over the course of her life, day by day, or perhaps moment by moment.” 91 Similarly, it is almost always claimed that the components relevant to the goodness of an outcome are the individual lives of sentient beings and their respective goodness or desirability. The goal of this chapter will be to examine the truth of the latter of these two claims—i.e. to examine whether the components relevant to the goodness of an outcome are in fact the individual lives of sentient beings and their respective goodness or desirability. Ultimately, I will argue that while the goodness of a life is a function of the moments of that life, the goodness of an outcome is a function not of the goodness of the lives in the outcome, but rather, of the goodness of the moments in that outcome. Of course, there will not necessarily be a divergence between the assessments of the goodness of outcomes made in line with my claim (that the components relevant to the goodness of an outcome are the moments) and those made in line with the conjunction of the two common claims just mentioned (that the components relevant to the goodness of an outcome are the lives and that the components relevant to the goodness of a life are the moments). My assessment of the outcome and the common assessment of the outcome would be identical given some approaches to combining the relevant components (such as total utilitarianism (TU)). With other approaches to combination (such as average utilitarianism (AU)), however, the assessments need not be identical. 92 Thus, it is this claim, that the goodness of an outcome is a function of the goodness of the moments in that outcome—and not of the goodness of the lives in the outcome—which (if true) will provide reasons to accept average per moment utilitarianism (AMU) and to reject average total lifetime value utilitarianism (ATLU). I will begin in Section 1.2 by laying out the assumptions underlying the discussion in this chapter. Then, in section 1.3, I will offer a positive argument for AMU and explain both why its 91 Ibid. 92 The assessments could be identical in some cases, but, in the vast majority of cases, they would not be identical. 88 conclusion follows from its premises and why its second premise is plausible. The bulk of the chapter will then be devoted to arguing for the plausibility of the argument’s first premise. Section 1.4 will provide a number of related arguments to this effect, and then section 1.5 will provide a wholly different argument for Premise 1 that is rooted in a view about the metaphysics of persons. Thus, the chapter hopes to provide reason to accept Premise 1, and thus reason to accept the argument for AMU. Various important objections to AMU are not addressed in this chapter, but, with the positive argument on the table after the completion of this chapter, these objections will be addressed in the other chapters in Part II. Perhaps the most serious of these objections is that AMU must be false since it cannot accommodate the intrapersonal maximizing intuition (IMI)—the claim that, all else equal, it is better agent-neutrally that a happy person live longer. This objection will be the main topic of Chapter 4. However, if one rejects AMU, one will need to reject a component of the argument offered in this chapter. Thus, assuming that its conclusion follows from its premises, if one rejects AMU, one might consider the argument to be either a reductio against Premise 1 or a reductio against Premise 2 (AU on the whole). As such, the stronger the argument in this chapter, the more difficult it will be to reject its conclusion. This chapter will strive to strengthen the argument itself and the following chapters will strive to reject other reasons for casting doubt on its conclusion. 1.2. Some Assumptions Before diving into this chapter’s arguments, it’s important to be clear about some assumptions I will be making (many of which were also flagged in Part I as assumptions that I will be making in the dissertation as a whole). First of all, in order to simplify things, the discussion will occur in a utilitarian 89 framework. 93 As such, I will be assuming that utilitarianism is true. This is being done solely to minimize the number of moving parts, and, as discussed in Part I, my conclusions will be relevant to the aspects of non-utilitarian theories that are concerned with welfare. Thus, to the extent that a theory places any importance on welfare, the current discussion will be relevant. Next, it will be assumed that it makes sense to speak of a “moment” of a life. The details of how precisely we are to understand this phrase, however, are unclear. For one, there are questions in the philosophy of time about whether there are analogues of atoms in the context of time, or whether the building blocks of time are best understood as duration-less units (i.e. duration-less moments). Similar—though slightly different—questions arise when we ask about the smallest temporal unit of conscious experience. This question—and its implications for Theory X—will be probed in Part III Chapter 3, but I will assume throughout this chapter that it makes sense to speak of “moments” of a life, and that the questions regarding the precise formulation of this term need not be addressed in order to address the main questions that confront us. Furthermore, it has been assumed in the literature that there is nothing problematic about referring to “moments of a life.” 94 Not only will I assume that there are moments of a life, but I will also make the assumption commonly made that it makes sense to discuss the welfare-level of a particular conscious moment, and that for any arbitrary moment, there is a level of wellbeing that exists. This assumption is happily made by both Broome and Temkin, among others. Broome writes: “[L]ocations of good are spread across three dimensions: states of nature, people and times. . . . [G]oods are strung out on a three dimensional grid of locations. Each location is a person, at a time, in a state of nature. When we compare alternatives, each alternative will distribute good across this grid.” 95 Similarly, as discussed above, Temkin frequently refers to the wellbeing at a particular time, at one point (for 93 See Part I Chapter 1 for an explanation of what I mean by “utilitarianism,” and for a description of various other assumptions that I will be making for the purposes of this dissertation. 94 Temkin, p.24; Broome, p.25. 95 Broome, pp.25, 27. 90 example) mentioning “‘local’ utilities possessed by [an] individual . . . moment by moment.” 96 Thus, I, too, will assume that we can make sense of a person’s wellbeing at a moment. Further, it will be assumed (with Temkin and Broome) that the goodness of a life is some function of the moments in that life and their respective goodness. Relatedly, it will be assumed that the goodness of any chunk of life, ranging in size from one moment to an entire lifetime, will be a function of the moments contained therein. Although these assumptions surely can be contested, they seem quite plausible and not only are they frequently employed in philosophy, but they are generally employed in public policy. For example, in addition to implicitly signing onto these assumptions, governmental decisions often explicitly employ the notion of “quality-adjusted life years.” Lastly, for the purposes of this chapter, I will bracket the question of which objects are bearers of intrinsic value. This is an important question, though, and it will be the focus of Part III Chapter 3. For our purposes here, however, I will loosely use the word “goodness” to refer to a property of all sorts of objects, including states of affairs, lives, and conscious moments. By doing so, I do not mean to assume that all of these objects are capable of being bearers of intrinsic value. Rather, I am suggesting that, at the very least, they can bear some relation to intrinsic value, and thus, I loosely refer to this property of theirs as goodness. A more precise account of intrinsic value will be worked out in Part III Chapter 3. With these assumptions in place—be they plausible or implausible—we can now proceed to see what follows in a framework where they are granted. 1.3. An argument for AMU Assuming AU, consider the following argument for AMU (as distinguished from ATLU): 96 Temkin, p.24. 91 Premise 1: For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than y. Premise 2: All else equal, it is not better or worse that there exist more people than fewer people. Lemma: All else equal, it is not better or worse that there exist more conscious moments than fewer conscious moments. Conclusion: AMU is true. In order to show that this argument’s conclusion is supported by its premises, and in order to make it more tangible, let’s consider a particular case. Consider two possible states of affairs in a far off land, perhaps in the distant future: In S1, one individual exists for 100 years, accruing a constant stream of happiness per moment, one unit per year, ultimately totaling to 100 units of happiness by the end of his lifetime. In S2, two individuals exist (be it simultaneously or sequentially), each living lives with an identical stream of happiness to that of the individual in S1, but in S2, each individual lives for only fifty years, each thus accruing one half the lifetime happiness accrued by the individual in S1. If Premise 1 is true, then S1 and S2 are equally good. (Recall that I explained, in Part I Chapter 2, why we should think that “neither better than nor worse than” is synonymous with “equally good as.”) The two states of affairs contain the same number of conscious moments, all else is equal, and thus it’s not better or worse that, in S1, the conscious moments all exist in one life whereas, in S2, the conscious moments exist in two different lives. If Premise 2 is true, then S2 would be equally as good as a state of affairs, S3, which is just like S2, except only one of the two individuals exists. If this is true, then S1 will be equally as good as S3, and the only difference between these two states of affairs is that the only individual who exists has a life in S1 that is twice as long as the life of the individual who exists in S3. Replicating this reasoning, we can generalize from this case, and thus arrive at the argument’s intermediate conclusion (Lemma)—that, all else equal, it is not better or worse that there exist more conscious moments than fewer conscious moments. As for the step from Lemma to the conclusion that AMU is true, this is in fact a step that 92 could be contested, but it is not a step that would be contested by a proponent of ATLU (or any other version of AU). ATLU will rely on the same reasoning when it comes to the averaging over the goodness of persons’ lives to arrive at the goodness of an outcome. As such, the step from Lemma to the conclusion will be largely set aside in this chapter. It is important to note, though, that this is not only being bracketed because the target of the above argument is a proponent of AU. It is also the case that there are positive reasons for believing this step. As described at length in Part I Chapter 2, as long as one rejects the viability of person-affecting intuitions, a view holding that, all else equal, more components is not better naturally leads to a view that determines relative goodness by averaging over the relevant components. Assuming that this argument’s conclusion is supported by its premises, what are we to make of the two premises? Let’s begin with Premise 2. Premise 2 is merely a restatement of population-size anti-more-is-better intuition (AMBP). In Part I Chapter 2, I argued for the plausibility of AMBP, and I also argued that the plausibility of this intuition (in conjunction with the difficulties afflicting person-affecting views) is one of the main motivating reasons for adopting average utilitarianism. Nevertheless, as I conceded in Part I Chapter 2, there will be people who are not drawn to AMBP. As I mentioned, however, even if one is not drawn to this intuition, there are other reasons why one might consider the plausibility of, or even endorse, AU: Other candidates for Theory X are afflicted by various difficulties—for example, TU is afflicted by the Repugnant Conclusion. Thus, for those readers who are not drawn to AMBP, but who are considering which version of AU might be the most plausible, these readers should treat Premise 2 not as representing a deeply felt intuition, but rather as a derived premise. They should consider Premise 2 to be, for the purposes of this chapter, a stipulation of AU—the larger family of AU sub-views. As such, we can assume that AU is true if our investigation here is into which version of AU is most plausible. Of course, if one is drawn to the intuition stated in Premise 2, one can 93 simply take the premise’s words at face value. Now let’s consider Premise 1. Premise 1 might not necessarily raise eyebrows (or at least not severely so) when it is realized that an implication of its truth would be the intermediate conclusion in the argument that S1 is equally as good as S2. What might raise more eyebrows, however, is the result of jointly holding Premise 1 and Premise 2—namely that, assuming what would be a continued constant flow of happiness, and all else equal, we should be indifferent about whether an individual has additional years added onto his life. This equal valuation of S1 and S3 might be harder for a reader to get on board with than the equal valuation of S1 and S2, because although both go against ATLU, the equal valuation of S1 and S3 seemingly involves an uncompensated affront to IMI. At least in the equal valuation of S1 and S2, the conscious moments not enjoyed by one person are enjoyed by an additional person. In the comparison between S1 and S3, however, there simply are fewer conscious moments that are enjoyed period. If, however, one accepts Premise 2, and if one is willing to accept the equal valuation of S1 and S2, one would have to accept the equal valuation of S1 and S3, at pains of inconsistency. Thus, if one objects to the conclusion and one accepts Premise 2, one must reject Premise 1, thus denying not only that S1 is equally as good as S3, but, importantly, also denying that S1 is equally as good as S2. Thus, what can be said for Premise 1? Shall we accept it? I must say that Premise 1 and the intermediate conclusion that S1 is equally as good as S2 strike me as quite plausible, and that the overall conclusion thus strikes me as plausible as well. Nevertheless, for those who don’t have this prima facie impression, which might be many readers, we must further explore the plausibility of Premise 1. 1.4. Reasons for accepting Premise 1 In order to argue for Premise 1, it will be useful to first introduce a few new concepts and 94 provide some new terminology. So far it has been said that the first of the two steps in aggregation consists in identifying the component parts of the whole that are relevant in assessing the goodness of the whole. It turns out, however, that this notion, “component parts of the whole that are relevant in assessing the goodness of the whole” is somewhat vague, and it includes at least two separate concepts that are important to distinguish from one another. On the one hand, it includes the notion that I will hereafter call “the unit relevant for ethical inquiry.” The unit relevant for ethical inquiry refers to the type of object, the goodness of the instantiations of which are the data inputs into the combination function that constitutes the second process of aggregation. In other words, the combination function takes as its inputs a string of numerical values, and each of these values represents, respectively, the goodness of one instantiation of the unit relevant for ethical inquiry. To make this more concrete: Take, for example, a theory whose combination function is an averaging function. The unit relevant for ethical inquiry is the type of object, the goodness of the instantiations of which is averaged over. This is straightforward: It is clear that any averaging view would have to average goodness over a relevant unit, and this is the unit that I call the unit relevant for ethical inquiry. Thus, for AMU, the unit relevant for ethical inquiry is the conscious moment. For ATLU, the unit relevant for ethical inquiry is an individual’s life. This notion, “the unit relevant for ethical inquiry,” is not, however, the only concept that falls under the umbrella of the broader notion “component parts of the whole that are relevant in assessing the goodness of the whole.” Let me explain. First consider AMU: If AMU were the only possible theory, “the unit relevant for ethical inquiry” would be the only concept that falls under the umbrella notion. This is because the unit relevant for ethical inquiry, according to this theory, is not itself a whole composed of smaller parts. This, however, is not the case for a theory such as 95 ATLU. According to ATLU, the unit relevant for ethical inquiry is an individual’s life, and a life, itself, is a whole composed of smaller parts—conscious moments. Thus, according to ATLU, there are two combining functions: There is a combining function that calculates the goodness of the outcome and takes as its data points the string of values representing the goodness of lives, and then there is a second combining function that calculates the goodness of an individual life and takes as its data points the string of values representing the goodness of conscious moments. In light of this, and in light of my definition of “the unit relevant for ethical inquiry,” conscious moments are not the unit relevant for ethical inquiry according to ATLU. This notwithstanding, according to ATLU, conscious moments still are component parts of the whole outcome that are relevant to assessing the goodness of the whole outcome, because the goodness of a life is a function of their goodness, and the goodness of an outcome is a function of the goodness of the lives in the outcome. Thus, while, on some theories, the unit relevant for ethical inquiry will be the only component part of the whole that is relevant in assessing the goodness of the whole, on other theories, in addition to there being a unit relevant for ethical inquiry that is a component part of the whole that is relevant in the assessing the goodness of the whole, there will also be other component parts of the whole that are relevant in assessing the goodness of the whole. I will now introduce a new concept that I will term a “building block.” As I will use the term, it will refer to a component part of a whole outcome that satisfies both of the following two criteria: (i) Be it the unit relevant for ethical inquiry or not, its goodness is relevant to the assessing of the goodness of the whole outcome in the sense described above (i.e. it is taken as data in a combining function that either produces the goodness assessment of the outcome, or it is taken as data in a combining function that produces the goodness assessment of a unit that is taken as data in a combining function that produces the goodness assessment of the outcome, or it is taken as a data point in an even more attenuated portion of a hierarchy of this sort), and (ii) It does not 96 constitute a whole, the goodness of which is determined by employing a combining function to combine the goodness of smaller components of this whole. As such, a “building block,” as defined here, constitutes the explanatory end of the line. A utilitarian of any sort will deem these building blocks to be conscious moments of happiness or welfare. Importantly, this will be the case both, on the one hand, for the AMU theorist, and, on the other hand, for the ATLU theorist—and for other theorists who think that the unit relevant for ethical inquiry is an individual’s life. Thus, it is possible for two utilitarian theories to agree about what constitutes the building blocks, yet disagree about what constitutes the unit relevant for ethical inquiry. This is the case for ATLU and AMU: They agree that conscious moments are the building blocks, yet AMU maintains that the unit relevant for ethical inquiry is the conscious moment, and ATLU holds that the unit relevant for ethical inquiry is the individual’s life. 97 With these concepts defined and explained, we can now proceed to assess the truth of Premise 1—the claim that “For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than y.” The purpose of this section (1.4) and the next section (1.5) will be to argue for the plausibility of this Premise 1, and it will aim to do so by arguing for the plausibility of the claim that the unit relevant for ethical inquiry is the conscious moment—and that it is not an individual’s life. If it can be shown that the unit relevant for ethical inquiry is in fact the conscious moment, the truth of Premise 1 will follow. (If the unit relevant for ethical inquiry were the individual’s life, Premise 1 could still be true, depending on which measure of combination were used, but it would not necessarily be true. If the unit relevant 97 It’s important to note that the combination phase of aggregation itself can be understood as confronting similar issues to the above issues confronting the process of identifying the relevant elements to combine. Even if one holds that the unit relevant for ethical inquiry is the conscious moment—thus holding that person data is not directly relevant for the assessment of outcomes—one might still confront questions such as (but not limited to) whether to aggregate with “pure” methods (such as summing or averaging) where the assessment of the outcome would always be affected by a change in the value of an element, or whether to aggregate with a method (such as maximin) where a change to an element’s value does not necessarily bring about a different assessment of the outcome. 97 for ethical inquiry is the conscious moment, then Premise 1 will always be true.) Thus, I now set my sights on arguing that the unit relevant for ethical inquiry is the conscious moment, and, to this end, the framework set up here and the terms introduced here will be helpful. 1.4.A. A first pass: Michael Huemer’s argument 1.4.A i. An overview of Huemer’s argument and an explanation of its relevance for me To start, let’s first consider a variant of an argument—the “Equivalence Argument”—that Michael Huemer provides for why we should accept the Repugnant Conclusion and, thus, for why we should accept TU. 98 In his argument, Huemer does not use my terminology (such as “the unit relevant for ethical inquiry”), but, in what follows, I will describe his thesis in these terms, and I will explain later in this section why this is an appropriate interpretation. In his argument, Huemer attempts to do two things. First, he argues that the conscious moment is the unit relevant for ethical inquiry. Second, he argues that this conclusion exclusively supports TU. I ultimately will argue that both of his attempts fail: His second claim fails because if the conscious moment is the unit relevant for ethical inquiry, while this does provide reason to reject ATLU, it provides equal support for TU and for AMU—as opposed to providing support exclusively for TU. Further, and more importantly for our purposes, his first claim fails, I argue, because it assumes its conclusion. Thus, neither of Huemer’s two claims should be accepted based on the arguments and discussions he provides. This notwithstanding, however, Huemer’s discussion is fruitful to explore both because of the usefulness of the framework of equivalences and dimensions that it sets up, and because taking notice of his mistakes can help us avoid committing the same or similar ones. 98 Huemer. 98 1.4.A ii. Huemer’s framework of “equivalences” and “dimensions” First, I provide some background that will both aid our understanding of Huemer’s “Equivalence Argument” and also further illuminate our topic. According to Huemer, we can consider different dimensions of welfare—such as intensity (per moment), duration, and widespreadness across recipients—and we can speak of dimensions being equivalent to each other just in case an increase (or decrease) in one is equally as good as a proportionate increase (or decrease) in another. For example, as Huemer says, “if the intensity of a pleasurable experience is . . . equivalent to the duration of the experience, then a doubling 99 of intensity is . . . [equally] as good as a doubling of duration.” 100 In his attempt to defend TU, Huemer ultimately wants to argue for two equivalences: The equivalence of duration and intensity of benefits and the equivalence of duration and widespreadness across recipients of benefits (and, thus, also the equivalence of intensity and widespreadness across recipients of benefits). To be clear, as the term is used here, doubling widespreadness across recipients means doubling the number of existing people and making the additional people’s lives be qualitatively equivalent to the original people’s lives. This is to be distinguished from a possible meaning of the term according to which doubling widespreadness would yield twice the number of people but keep the total benefits in the populations equal, thus yielding fewer benefits per person in the larger population. Different candidates for Theory X can, in many cases, be distinguished by their claims regarding the equivalence or non-equivalence of these dimensions, and, for TU, ATLU, and AMU, this is purely a function of which theories deem which dimensions morally relevant. (Perhaps for different hybrid theories, equivalences might not only be a function of which dimensions are deemed relevant, but also a function of what we might call exchange rates among the dimensions for each theory. In the case of these three theories, however, the relevance of the dimensions are 99 This requires measurability on a ratio scale. 100 Huemer, p.918. 99 all or nothing.) A proponent of TU, with Huemer, will find all three dimensions to be morally relevant, and thus he will deem all three to be equivalent to each other. A proponent of ATLU will deem intensity and duration to be relevant, but deny the relevance of widespreadness of recipients. As such, he will deem intensity to be equivalent to duration (both matter) but deny both the equivalence of duration with widespreadness of recipients and of intensity with the widespreadness of recipients. A proponent of AMU will deem intensity to be relevant, but deny the relevance of duration or widespreadness of recipients. As such, he will deem duration equivalent to widespreadness of recipients (i.e. neither one matters) but will deny the equivalence of both intensity with duration and of intensity with widespreadness of recipients. In light of this, the difference between AMU and ATLU (as we already knew) is that they disagree about whether duration is morally relevant. As such, one of the two equivalences about which they disagree is the potential equivalence of duration and widespreadness of recipients. AMU holds that they are equivalent but ATLU holds that they are not. TU holds that they are equivalent, though, and to this extent, the TU theorist is an ally of the AMU theorist against the ATLU theorist. As such, let’s consider Huemer’s argument in support of the equivalence of duration and widespreadness of recipients. (It is important to note, of course, that the other potential equivalence about which ATLU and AMU disagree is that of duration and intensity. And with respect to this dispute, TU is an ally not of AMU, but, rather, is an ally of ATLU. Thus, AMU disagrees with TU’s, and Huemer’s, claim that duration and intensity are equivalent. I will, however, return to the issues raised by this disagreement about potential equivalences below.) 1.4.A iii. Huemer’s Equivalence Argument Consider Huemer’s example, 101 which is similar to the example of S1 and S2 in section 101 Ibid. 100 1.3: Suppose that there are different individuals, Sue and Mary, who are numerically non-identical but who are qualitatively indistinguishable. In world G, Sue exists for ten minutes, experiencing a total of five units of happiness in the first five minutes due to eating a sandwich, and experiencing an additional five units of happiness in the next five minutes due to listening to music. Mary does not exist. In world H, Sue exists for five minutes, experiencing a total of five units of happiness due to eating a sandwich. Sue then goes out of existence and Mary comes into existence, experiencing a total of five units of happiness in her five minutes of existence, due to listening to music. Huemer concedes that some might argue that world G is better than world H because, all else equal, they might think it better for an already existing person to receive a benefit rather than to create a new person to receive that benefit, and, relatedly, one might argue that it might be worse for a person to die prematurely than for a potential person to not be created. This notwithstanding, Huemer argues that G and H are equally good. According to Huemer, “The intrinsically good-making feature[s] of the sandwich-eating [and music-listening] experience[s]—namely [their] pleasurableness—[are] equally present in either case.” 102 He continues: It seems that the two temporal halves of world G, [G1 and G2,] are equally good, since each consists in Sue’s enjoying an equally good benefit for five minutes. It is hard to think of a reason why one of these five-minute events should be better than the other. Similarly, it seems that the two temporal halves of world H, [H1 and H2,] are equally good. Sue and Mary are qualitatively indistinguishable people who enjoy equal benefits for equal lengths of time. So it is hard to see how one of their lives would be better than the other. Finally, it seems that the first half of world G and the first half of world H are equally good. Both consist in Sue’s enjoying the same benefit for the same length of time. If the intrinsic value of an event supervenes on its intrinsic, non-evaluative properties, then the first halves of G and H must have equal intrinsic value. 103 Thus, Huemer reasons that VG2 = VG1 = VH1 = VH2, 104 and thus that worlds G and H are equally good. Therefore, in light of the fact that there is nothing unique about the specifics of his 102 Huemer, p.920. 103 Ibid, p.921. 104 “V” means value, the second letter denotes the world, and the numeral denotes the temporal part. 101 example, he says that the duration of welfare for an individual is “evaluatively equivalent to widespreadness of welfare—the number of people enjoying the benefit.” 105 1.4.A iv. The problem with Huemer’s first claim Let’s assess this argument. First, it’s important to point out the fact that Huemer is assuming that events can be the bearers of intrinsic value. This is something that can be contested by his opponent. Nevertheless, to avoid this concern, and to be charitable where his argument requires charity, I suggest we interpret Huemer’s claims about the intrinsic value of an event not as positive claims about which objects are bearers of intrinsic value, but rather, in line with the disclaimer I myself make at the end of section 1.2. In other words, I suggest that where he refers to the intrinsic value of an event, we instead substitute in “goodness,” where “goodness” is being used loosely to refer to a property of all sorts of objects, including states of affairs, lives, and conscious moments. In so doing, I don’t mean to assume that all of these objects are capable of being bearers of intrinsic value. Rather, I’m suggesting that, at the very least, they can bear some relation to intrinsic value, and thus, we can loosely refer to this property of theirs as goodness. Even if we interpret Huemer’s argument charitably in the above manner, as I suggest we should, his argument still runs into problems. Clearly, a key qualification of Huemer’s argument is the phrase “If the intrinsic value of an event supervenes on its intrinsic, non-evaluative properties,” and it is precisely this phrase that is at the root of my first objection to his argument. If Huemer’s assumption is true, however, then his conclusion is supported, and, as he suggests, it is supported only if it is true. So far so good, but the problem is that it is precisely this assumption that his opponent—in this case, the ATLU theorist—will reject. This would be unproblematic if he 105 Huemer, p.921. 102 provided an argument for this crucial assumption, but he does not. 106 In light of our charitable interpretation of Huemer with respect to his first problematic assumption, however, it will make an assessment of this second problematic assumption clearer if we articulate the second concern in terms that I have suggested avoid the first concern. Thus, we can understand Huemer to be saying the following: “If the goodness of an event [i.e. how it contributes to the goodness of a state of affairs] is a function of its intrinsic properties [e.g. not properties regarding whether it exists in the same person as another event or whether it is the only event of a person’s life], then the duration of welfare for an individual is evaluatively equivalent to widespreadness of welfare—the number of people enjoying the benefit.” Or, to put this statement into other terminology of mine, we can understand Huemer to be saying the following: “If the unit relevant for ethical inquiry is the conscious moment, then ATLU must be false.” As should be clear with each of these new formulations as well, however, Huemer’s argument assumes its conclusion: The assumption that is required for the argument to follow is one that his opponent (the ATLU theorist) will not accept. Huemer assumes that the unit relevant for ethical inquiry is the conscious moment, and it is the truth of this claim that he is arguing for. As such, rather than providing a successful argument for the claim that the conscious moment is the unit relevant for ethical inquiry, Huemer’s argument instead simply makes it very clear what the relevant assumption is that needs to be defended if we are to show that the conscious moment is the unit relevant for ethical inquiry. Thus, rather than being the end of our defense of the claim that the conscious moment is the unit relevant for ethical inquiry, it in fact is just the beginning. As I have made clear, however, I do think that the goodness of an event is a 106 Lest there be confusion, it’s important to clarify the following. There is an extremely widespread view that moral properties supervene on “natural” properties. This, however, is not the assumption that Huemer is making here. He’s assuming something more specific. Huemer is assuming that moral properties of events supervene on the intrinsic properties of these events. Some natural properties are intrinsic properties, but not all natural properties are intrinsic properties. 103 function solely of its intrinsic features, and that these features are equally present in both G and H, and I agree with Huemer’s claim. My task, however, is to provide reasons to believe what in Huemer’s argument is simply an unsupported claim. 1.4.A.v. The problem with Huemer’s second claim Before continuing with this task, however, it’s important to raise my objection to Huemer’s second claim. After arguing for his first claim, that the conscious moment is the unit relevant for ethical inquiry, Huemer argues for a second claim—though it might be more appropriate to say that he asserts a second claim, since there is little argumentation to be found. Huemer’s second claim states that if the conscious moment is the unit relevant for ethical inquiry, then TU is true. Thus, the conjunction of these two claims constitutes Huemer’s argument for TU. Huemer is mistaken to think that the claim that the conscious moment is the unit relevant for ethical inquiry, if true, supports only TU and not AMU as well. While I would agree that his argument shows that worlds G and H are equally good in light of the fact that they are each composed of two components and all four components are equally good, he goes a step further. Huemer ultimately concludes that the value of world G is VG1 + VG2, and that the value of world H is VH1 + VH2. He arrives at this conclusion because of his belief in the position that, as it happens, is the common ground between TU and ATLU, but not AMU: the notion that duration is morally relevant. In short, Huemer finds it plausible that IMI is true: He believes that, all else equal, it is better agent-neutrally that a happy person live longer. Although Huemer is in the majority in believing that duration is morally relevant (and thus believing IMI), he does not provide an argument to this effect, and the truth of this claim cannot simply be assumed or taken as self-evident. Further, if my arguments in this chapter and the next 104 are successful, I will have shown that duration is not morally relevant. 107 Leaving aside this debate for a moment, however, one thing does seem clear: If the conscious moment is the unit relevant for ethical inquiry, this provides reason to reject ATLU, but it supports both AMU and TU—at least absent any further arguments regarding the moral relevance or irrelevance of duration. What precisely an AMU theorist’s account of intrinsic value will look like will be addressed in Part III Chapter 3, but for our current purposes, it is sufficient to say that both TU and AMU would be consistent with the claim that the conscious moment is the unit relevant for ethical inquiry. Further, this claim provides support for TU and AMU to the extent that there are competing theories (such as ATLU) that are inconsistent with this claim. Thus, in conclusion, neither of Huemer’s two claims should be accepted based on the arguments and discussions he provides. This notwithstanding, Huemer’s discussion is fruitful to explore both because of the usefulness of the framework of equivalences and dimensions that it sets up, and because taking notice of his mistakes can help us avoid committing the same or similar ones. 1.4.B. Bentham and Mill and the implications for AMU of TU’s partial, and contrasting, alliances with ATLU and AMU One of the (various) fruitful aspects of Huemer’s discussion of equivalences and dimensions is the way that it enables us to think about the partial alliances and non-alliances that TU shares with ATLU and with AMU. As discussed, TU is an ally of ATLU (against AMU) in one way and, in another way, it is an ally of AMU (against ATLU). On the one hand, TU and ATLU share the position, denied by AMU, that duration is equivalent to intensity. On the other hand, TU and AMU share the position, 107 I will return to the specific question of the plausibility of AMU’s claim that duration is not equivalent to intensity and why this is plausible despite being incompatible with the widely and deeply held, and seemingly plausible, intrapersonal maximizing intuition in Part II Chapter 2. 105 denied by ATLU, that duration is equivalent to widespreadness of welfare across recipients. Interestingly, though, despite TU’s non-wholehearted support of AMU, TU would give wholehearted support for Premise 1 in my argument for AMU: “For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than y.” This premise is essentially a statement of the equivalence of duration and widespreadness across recipients. TU theorists would, of course, deny Premise 2, which articulates AMBP, so they would not accept my argument’s conclusion. Nevertheless, AMU and ATLU theorists, alike, will accept Premise 2 (be it as an axiomatic or as a derived premise), and this is why we have been focusing on the plausibility of Premise 1 in our investigation into whether AMU or ATLU is the more plausible version of AU. AU has been taken as a premise of this whole discussion. The crucial point in arbitrating between AMU and ATLU is establishing whether Premise 1 is or is not plausible, and it might have seemed to many, at first glance, that Premise 1 should be denied and that the argument for AMU should be denied on these grounds. While the support from TU theorists in no way proves that Premise 1 is plausible—let alone that the whole argument that AMU is the best version of AU is plausible—its importance is not negligible. It suggests that the vast majority of utilitarians in history (who have been TU theorists) have found Premise 1 to be plausible. Bentham and Mill, as well as countless others, were TU theorists, and they all thought that duration was equivalent to widespreadness across recipients. 108 For them, it was equally good for happy conscious moments to be added to an existing person’s life as it would be for them to constitute a new person’s life. For them, it doesn’t matter, at least intrinsically, how happy moments are distributed across people. As such, it seems 108 Bentham; Mill. 106 as though, if anything, the burden of persuasion 109 should be placed squarely on the ATLU theorist and his attempt to explain his denial of Premise 1. 110 It’s important to note, however, that although TU clearly accepts Premise 1, it’s less obvious that TU accepts the claim that the unit relevant for ethical inquiry is the conscious moment and not the person. Huemer—who is a TU theorist—accepts this latter claim, and he assumes that others would as well. I, too, think that TU theorists do and should accept this claim. These points notwithstanding, and given the discussion so far, it seems that any of the following three positions might be available to the TU theorist: (1) Hold that the relevant unit is the conscious moment, (2) hold that the relevant unit is the person, or perhaps even (3) remain agnostic. The reason for this uncertainty is that TU’s summative aggregation method results in the same prescriptions regardless of which of these three options is the case. Regardless of this uncertainty, however, it still remains the case that the TU theorist accepts Premise 1. Further, the discussion below, in 1.4.C, regarding what the utilitarian should think can be seen as some evidence that TU theorists have held that the unit relevant for ethical inquiry is the conscious moment. Importantly, however, what is key for me is not merely to show that utilitarians have thought that Premise 1 is true, or that utilitarians have thought that the conscious moment is the unit relevant to ethical inquiry. Rather, I must show that this is what a utilitarian should think, and it is to this question that I now turn. Nevertheless, history does provide some support for my position. 109 In reality, any philosophical claim has a burden of proof. Only total agnosticism or suspension of judgment might have no such “burden.” Although I will use the phrase a fair amount in this chapter, I mean for it to be understood more colloquially rather than strictly. 110 Of course, in the very same way that I am enlisting TU’s powers against ATLU and in favor of AMU in the context of the question of whether duration is equivalent to widespreadness across recipients, so too an ATLU theorist will be able to enlist TU’s powers against AMU in chapter 4 and elsewhere where I argue, contra both ATLU and TU, that duration is not equivalent to intensity. 107 1.4.C. Why a utilitarian should think that the unit relevant for ethical inquiry is the conscious moment and, thus, why a utilitarian should accept Premise 1 1.4.C.i. Now, however, I turn to the heart of the matter: Why is it that a utilitarian should think that the conscious moment is the unit relevant to ethical inquiry, and thus why is it that a utilitarian should think that Premise 1 is true? The prima facie point that seemingly provides support for the claim that the unit relevant for ethical inquiry is the conscious moment is simply that the “conscious moment of happiness” seems to be a more primitive notion than “a life containing moments of happiness” or “a happy life.” 111 We can understand the former notion without the latter notions, but not vice versa. Similarly, it seems that the justification of goodness goes in one direction but not the other. It’s not the case that a moment of happiness is good because a happy life or a collection of happy moments is good, but rather, it is the reverse. These points are not dispositive, and an ATLU theorist could consistently maintain that the justification goes in the opposite direction, but this seems as though it would be a strained account. Furthermore, as discussed in the beginning of Section 1.4, it seems that an ATLU theorist would concede that conscious moments of happiness are the “building blocks.” Even the ATLU theorist maintains that the goodness of a life is a function of the conscious moments it contains. Thus, in what follows, I will not further consider this potential line of response for the ATLU theorist. In light of this, it seems that an ATLU theorist (or any other utilitarian theorist who is an advocate of an individual’s life as the unit relevant for ethical inquiry) might agree with my above claims about primitiveness and opt for a different avenue of disagreement. The move that the ATLU theorist will likely make will be to simply and baldly maintain the plausibility of the 111 Aristotle seemingly would disagree with this point and the following points. He, however, would likely be holding the minority viewpoint. See Aristotle, 1098a20. 108 following: While the goodness of a life is a function of the intrinsic features of its component conscious moments, and while the goodness of an outcome is a function of the intrinsic features of its component lives, the goodness of an outcome is not a function merely of the intrinsic features of its component conscious moments. The question confronting this position, however, is why we should introduce this intermediate unit and deem that it is the unit relevant for ethical inquiry. It seems as though the ATLU theorist owes us an answer to this question, but it is neither clear that the ATLU theorist has an answer, nor what such an answer might look like. However, an ATLU theorist could, of course, ask the analogous question of the AMU theorist: Why shouldn’t we introduce this intermediate unit and deem that it is the unit relevant for ethical inquiry. It seems to me, though, that, in this dialectic, the burden of persuasion 112 lies with the ATLU theorist. Despite the fact that I think both that the ATLU theorist has the burden of persuasion in this debate about the unit relevant for ethical inquiry, and that the ATLU theorist has not (and, indeed, cannot) satisfy this burden, it seems to be fairly clear that it is not contradictory to maintain that the unit relevant for ethical inquiry is the individual’s life. One could consistently maintain this position. Similarly, I imagine, an ATLU theorist would concede that one could consistently maintain that the unit relevant for ethical inquiry is the conscious moment. Thus, it seems that there might be no knockdown argument to be had for either camp, and that there might ultimately be an impasse between the warring intuitions. The fact that there might be no knockdown argument to be had and the fact that there might continue to be an impasse between these warring intuitions, however, does not mean that the two theories are, ipso facto, equally plausible, or that nothing further can be said. Thus, despite the fact that I do not purport to have a knockdown argument for the claim that the unit relevant for ethical inquiry is the conscious moment, in this section, I will provide considerations in favor of 112 As I have said, any philosophical claim has a burden of proof. Although I will use the phrase a fair amount in this chapter, I mean for it to be understood more colloquially rather than strictly. 109 accepting this claim. These considerations will be of two types: On the one hand, I will argue that the burden of persuasion, mentioned above, does in fact lie with the ATLU theorist. On the other hand, I will provide positive points in support of the claim that the conscious moment is the unit relevant for ethical inquiry. Underlying these two types of considerations offered here, however, will be a single theme that constitutes the backbone of my objection to ATLU: ATLU is a puzzling account that is unsteady and difficult to justify because of its hybrid nature. Section 1.4.C.ii explains the hybrid nature of ATLU, shows what’s puzzling about it, and shows how it constitutes a great departure from traditional utilitarian theories. Sections (1.4.C.iii) and (1.4.C.iv) further address different aspects of the hybridism claim. Section (1.4.C.v) then offers a possible explanation of ATLU intuitions. The arguments in this section will, perhaps, not persuade someone who is staunchly committed to ATLU, but they will, perhaps, persuade those who find ATLU appealing but who are not staunchly committed to it. Further, as for the reader who is staunchly committed to ATLU, perhaps these considerations will prompt him to further explain the appeal of ATLU and to further explain the rationale of maintaining its truth. 1.4.C.ii. It is a commonly noted feature of utilitarianism that, in many ways, it ignores personhood: As Rawls famously states, “Utilitarianism does not take seriously the distinction between persons,” 113 and, as Sen and Williams state, utilitarianism “merges the utility bits together as one total lump, losing in the process both the identity of the individuals as well as their separateness.” 114 This disregard for personhood is what underlies many, if not most, objections to utilitarianism, but it is also a feature of utilitarianism that is recognized and embraced by those 113 Rawls, p.24. 114 Sen and Williams, p.5. 110 who endorse utilitarianism. In short, it is well-known and non-controversial that utilitarianism disregards personhood. To further illustrate this feature of utilitarianism, it is helpful to consider the broader passages of Rawls and Sen and Williams from which the above quotations are taken. Sen and Williams write: A moral principle can be viewed as a requirement to exclude the use of certain types of information in making moral judgments, and utilitarianism imposes, by implication, severe information constraints. These constraints rule out a great deal of information from being taken into consideration in arriving at moral judgments. This amounts to taking, in the context of moral evaluation, a remarkably narrow view of being a person. Essentially, utilitarianism sees persons as locations of their respective utilities—as the sites at which such activities as desiring and having pleasure and pain take place. Once note has been taken of the person’s utility, utilitarianism has no further direct interest in any information about him. This view of man is a common feature of different variants of utilitarianism, including the indirect forms such as rule utilitarianism, disposition utilitarianism, etc., since no matter what is taken to be the relevant choice variable, the judgments of states of affairs, conduct, institutions, etc., are all ultimately based exclusively on the amount of utility and disutility generated. Persons do not count as individuals in this any more than individual petrol tanks do in the analysis of the national consumption of petroleum. . . . This reduces the collection of diverse information about the n persons in that state into n bits of utility . . . . [Utilitarianism] merges the utility bits together as one total lump, losing in the process both the identity of the individuals as well as their separateness. The distributional characteristics of the utility vector are also consequently lost. By now persons as persons have dropped fully out of the assessment of states of affairs. . . . [A] state of affairs is judged exclusively on the basis of utility information related to that state. 115 Rawls writes: [According to utilitarianism,] just as an individual balances present and future gains against present and future losses, so a society may balance satisfactions and dissatisfactions between different individuals. And so by these reflections one reaches the principle of utility in a natural way. . . . The principle of choice for an association of men is interpreted as an extension of the principle of choice for one man. Social justice is the principle of rational prudence applied to an aggregative conception of the welfare of the group. . . . [The utilitarian] view of social cooperation is the consequence of extending to society the principle of choice for one man, and then, to make this extension work, conflating all persons into one through the imaginative acts of the impartial sympathetic spectator. . . . There is no reason to suppose that the principle which should regulate an association of men is simply an extension of the principle of choice for one man. . . . Utilitarianism does not take seriously the distinction between persons. 116 Thus, as these passages make clear, utilitarianism disregards the distribution of utility 115 Sen and Williams, pp.4-5 (italics are theirs). 116 Rawls, pp.20-24 (emphasis mine). 111 across people. According to utilitarianism, it is not relevant which person obtains a certain quantity of utility, because in assessing the goodness of an outcome, we simply lump together all of the utility information and we ignore person-data. 117 Similarly, it is not relevant how concentrated in a small number of lives or how dispersed among a large number of lives a certain amount of utility is within a population of a given size. Premise 1 (and thus AMU) maintains that this principle of lumping and disregarding of person-data, which is the hallmark of utilitarianism, also applies when considering how many persons exist and enjoy for themselves a portion of a given quantity of utility that is to be had by a population. If, however, one rejects Premise 1 and one maintains that the unit relevant for ethical inquiry is the individual’s life (as would an ATLU theorist), then one would hold the following hybrid position: While we should disregard person-data for questions regarding fixed-sized population cases, person-data is crucially important for questions regarding non-fixed-sized population cases. To reiterate my comments above, this is not an inconsistent position. After all, the comments in the literature about utilitarianism disregarding persons is generally not directed explicitly at non-fixed-sized population cases. They are either directed at the more run of the mill fixed-size population cases, or, what is more likely, they are meant as blanket observations about utilitarianism. As the ATLU theorist would rightly point out, they certainly don’t seem to be comments that are meant to take a stand on the debate between AMU and ATLU about the unit relevant for ethical inquiry. These points notwithstanding, however, there is still something puzzling about the ATLU theorist’s position, because even if it is not inconsistent to disregard personhood in one context and not in the other, it does seem as though the question of whether or not to disregard personhood is not an entirely different one in the two contexts. It seems as though the two questions about 117 Recall that, by “person-data,” I mean facts about which conscious moments are parts of which persons. 112 whether to disregard personhood, while different due to the differing contexts, still share a common core and have much in common. It seems as though there should be a single principle explaining why person-data is relevant or why it is irrelevant, and it’s not clear what justifies treating person-data differently in the two contexts. We want an explanation for this dichotomy. One might ask an ATLU theorist: Why are you so relaxed about how welfare is distributed between people in fixed-size population cases, yet so uptight about how many people the welfare is distributed among in non-fixed-size population cases? It seems that ATLU owes us an explanation. Thus, although ATLU is not inconsistent with utilitarian principles, it does seem that, in failing to disregard person data in non-fixed-size population cases, ATLU is more remote from the utilitarian tradition than we might have thought. Furthermore, and more importantly, it’s not just the case that ATLU seems more remote from the utilitarian tradition. Rather, by seemingly departing from utilitarian tradition, it adopts a hybrid position about the importance of persons, and this hybrid position is both puzzling and greatly under-supported—if, indeed, it is supportable at all. Sections 1.4.C.iii and 1.4.C.iv provide additional illustrations of this same point. 1.4.C.iii. Another way of illustrating the unsteady hybridism that comes with holding that the unit relevant for ethical inquiry is the individual life (and thus that comes with holding ATLU) is as follows: The ATLU theorist has difficulty explaining what accounts for the discrepancy in value posited by his theory and that posited by theories that hold that the conscious moment is the unit relevant for ethical inquiry. The dialectic here is similar to the dialectic in discussions about utilitarianism and its treatment of equality and distributive justice. 113 Although equality and other distributive principles can certainly be of instrumental value to a utilitarian, these principles can, at best, have instrumental value. Thus, given a stipulated distribution of utilities among persons, a utilitarian does not have the tools to say, for example, that a more equal distribution of utility is better than a less equal distribution—leaving aside questions of how equality of utility in the present would have instrumental value for future utility levels. This is because utility is what matters, and utility values have been stipulated. Any assessment of betterness of one state of affairs would have to be due to there being greater utility in one of the “locations” of utility, and since the values in all “locations” have been stipulated, no such assessment is possible. A similar problem afflicts the ATLU theorist. To the extent that there is a divergence between the value assessment that an ATLU theorist makes and that which either the AMU or TU theorist makes, the question is where this additional value or disvalue accrues. Given that the ATLU theorist is a utilitarian, the value or disvalue must accrue somewhere, and it must be a function of conscious experience, but the AMU and TU theorists are already taking all conscious experience into account. Just as in the above case about equality, the state of affairs has already had its utility values fully stipulated. Thus, it seems that there would have to be some sort of consideration extrinsic to utilitarianism that is affecting the ATLU theorist’s intuitions and infiltrating his allegedly utilitarian account. Thus, just as utilitarianism does not and cannot care (intrinsically) about the distribution of utility between people, so too it seems that it does not and cannot care about how many people utility is distributed among. To the extent that this type of distribution seems to matter, it seems that an extra-utilitarian principle is doing the work. It might be thought, however, that this argument begs the question against the ATLU theorist. In a sense this is true. After all, the ATLU theorist maintains that the unit relevant for 114 ethical inquiry is the individual’s life, and thus, even though he thinks that the goodness of a life is a function of the goodness of its component conscious moments, he thinks that goodness of a conscious moment (i.e. its contribution to the goodness of an outcome) is not a function merely of the conscious moment’s intrinsic features—it can also be a function of extrinsic features such as whether or not the conscious moment is part of an individual’s life that also contains other conscious moments. Thus, the points in this section do beg the question against the ATLU theorist in the sense that the ATLU theorist is committed to denying an assumption underlying my objection—namely that the goodness of an outcome is a function of the intrinsic features of its component conscious moments. This notwithstanding, however, I think that this argument shows—from an angle somewhat different than employed elsewhere in this chapter—the commitments that an ATLU theorist must have. Specifically, it shows, from a slightly different angle, the hybrid nature of ATLU and a puzzling feature of the commitment to a hybrid view of this sort. The hybridism illustrated in the argument and analogy above, in this section (iii), can be described as follows: The conscious moments are the building blocks, yet their intrinsic features are not all that is relevant to the assessment of an outcome. In addition to the intrinsic features of these building blocks, we must also bring in considerations, akin in the analogy in the beginning of this section iii, to considerations of equality. There is something else that is relevant beyond the utility values that have already been stipulated. This is the hybridism at work. ATLU seems to be utilitarian in its determination of what constitutes the building blocks, yet it seems non-utilitarian in its formula for getting from the intrinsic features of the building blocks to a goodness assessment of the entire outcome. Thus, this provides an additional angle from which to view what I am arguing is the implausible and counterintuitive hybridism of ATLU. This puzzling hybridism provides reason to 115 find plausible the claim that the unit relevant for ethical inquiry is the conscious moment, and thus, in so doing, it also provides reason to find Premise 1 plausible 1.4.C.iv. Next, consider a slightly different angle from which to view the hybridism issue afflicting ATLU. In Part II of Reasons and Persons, 118 Parfit considers three different possible theories of rationality: the moral theory, the self-interest theory, and the present aim theory. According to the moral theory, one should do what is best without giving a privileged status for weighting to results for a particular person (e.g. oneself) and without giving a privileged status for weighting to results for a particular time (e.g. the present). Thus, according to the moral theory, reasons should be agent-neutral and temporally-neutral. According to the self-interest theory, however, reasons should be temporally-neutral but agent-relative. According to the present aim theory, reasons should be fully relative—both temporally-relative and agent-relative. In his discussion, Parfit points out that the self-interest theory resides in a particularly unsteady position, because it is subject to attacks from both sides. This is an especially unsteady position, because not only can it be attacked from both sides, but, in most cases, when it uses an argument against one side, a similarly structured argument can used against it from the other side. According to Parfit, the difficulty for the self-interest theory is that it is a hybrid theory (appealing to neutrality in one dimension but to relativity in another), whereas the present aim theory and the morality theory are both pure (in that they either appeal to neutrality in both dimensions or to relativity in both dimensions). 119 Despite this difficulty, Parfit doesn’t necessarily think that a hybrid theory of this sort is incoherent or that the challenges it faces are insurmountable. Nevertheless, it does seem as though a hybrid view is left with somewhat more of 118 Parfit, p.117. 119 Parfit, p.140. 116 a burden of explaining why two dimensions should not be treated analogously. Though Parfit’s discussion concerns theories of rationality, there is an analogy (albeit an imperfect one) to be found in the comparisons between TU, ATLU, and AMU. All else equal, TU holds both that more people is better and that more conscious moments per person is better, and, as such, is a pure theory. AMU is also a pure theory: more people is not better and more conscious moments per person is not better. ATLU, however, is a hybrid: more people is not better, but more conscious moments per person is better. In light of this, ATLU occupies a precarious middle ground and can come under attack from both directions. And also in light of this, it seems that ATLU might have somewhat of a greater burden to explain its position because of the fact that it is a hybrid position. As was the case with Parfit’s discussion, though, it may well be the case that the hybrid theory has enough to say in its defense, and we might think that ATLU can satisfy its burden. Nevertheless, it does seem important to note which theories are pure and which are hybrid. Interestingly, however, there is a way in which the three theories here are not analogous to Parfit’s three theories of rationality. Here, there is more than one way to describe the lines along which the labels “pure” and “hybrid” are applied. One could instead define the conceptual space by appealing to different terms used in the discussion of the equivalences of different dimensions in Section 1.4.A. As such, the three theories could be characterized in such a way that not ATLU, but either TU or AMU is considered to be a hybrid. Despite the fact that I think that the labels I applied were the most natural for thinking of the three theories as pure or hybrid, the fact that other classifications are possible potentially lessens somewhat the heightened burden incurred by ATLU. Further, as discussed, even if the heightened burden were not lessened, this consideration would still not be dispositive against ATLU. This notwithstanding, applying Parfit’s framework of pure and hybrid theories to our context might be a useful tool for our analysis of AMU, ATLU, and TU. 117 It’s important to note that the discussion in (1.4.C.iii), above, is slightly different in its scope than this discussion here in (1.4.C.iv). The considerations in (1.4.C.iii) were not only a problem for ATLU, but also for any form of aggregating that re-weights data of the “building blocks”—even if the same type of aggregation occurs intrapersonally and interpersonally. In other words, even if averaging were done intrapersonally and interpersonally, but if each life got the weight of one regardless of its length, this system of aggregation would fall prey to the same concerns. (Of course, if additive summing were used on both levels, the objection would not arise.) The consideration here in section (iv), however, applies more narrowly to ATLU—at least in the form that the consideration is currently stated. As such, it does not directly suggest that the unit relevant for ethical inquiry should be the conscious moment and not a person’s lifetime. This notwithstanding, however, it does do so indirectly: If one holds that a person’s lifetime is the unit relevant for ethical inquiry, and unless one adopts TU, there will be a potential asymmetry between one’s treatment of conscious moments that are treated if part of an existing person and if in additional person, thus resulting in a hybrid theory. Importantly, however, although the foregoing concerns do not apply if one adopts TU (since the different sub-versions of TU would yield the same assessments), it seems that even a TU theorist has reason to hold that the unit relevant for ethical inquiry is the conscious moment and not the person. While it’s true that just as a proponent of TU could maintain that it’s the goodness of moments and not the goodness of individuals’ lives that are the instantiations of the unit relevant for ethical inquiry, it still is the case that we need to look at the data about goodness of moments if we are to know the goodness of a life. On the other hand, if we know the goodness of the moments, there is no need—on our way to getting to an assessment of the outcome—to know about persons. Thus, even for TU, an asymmetry (albeit an asymmetry of a different sort) 118 remains, and there is reason to claim that the unit relevant for ethical inquiry is the conscious moment. 1.4.C.v. 1.4.C.v.a Up until now, I have been arguing that we should accept the claim that the unit relevant for ethical inquiry is the conscious moment, and thus, that we should accept Premise 1 (“For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than y.”). To supplement arguments for this position, however, it might be useful to have a plausible explanation of what might be leading those who reject Premise 1 to reject Premise 1. In what follows, I will refer to the intuition that Premise 1 is false as the “anti- Premise 1 intuition.” Importantly, there are two main sub-types of the anti-Premise 1 intuition: On the one hand, there is what I will call the “ATLU-friendly intuition.” According to this intuition, for all numbers x, all else equal, it is better that x number of conscious moments exist in a smaller rather than a larger number of lives. On the other hand, there is what I call the “anti-ATLU- friendly intuition.” According to this intuition, for all numbers x, all else equal, it is better that x number of conscious moments exist in a larger rather than a smaller number of lives. I think that, of these two sub-types of the anti-Premise 1 intuitions, the ATLU-friendly intuition is much more common, and below, in Section 1.4.C.v.b, I will offer what I think might be one of the most likely reasons for why people might hold ATLU-friendly intuitions. Interestingly, according to this hypothesis, the intuition that Premise 1 is false is not a function of reflecting directly on Premise 1 itself, but it is a function of (consciously or unconsciously) finding unappealing implications of the conjunction of Premise 1 and Premise 2 (AMBP), and then determining that rejecting Premise 1 is the least implausible solution. 119 After offering this hypothesis, I then return to focus, a bit more narrowly, on Premise 1 itself, and, in Section 1.4.C.v.c, I will offer an additional possible explanation of why people might hold the ATLU-friendly intuition. This will be meant to apply to a wide-range of cases. Then, in Section 1.4.C.v.d, I will offer additional possible explanations for anti-Premise 1 intuitions in two somewhat circumscribed cases. One such case might be governed by the ATLU- friendly intuition, and as such, the mechanism I propose here might, in these circumscribed cases, work in conjunction with the primary explanation that I will have proposed in Section 1.4.C.v.c. Second, also in Section 1.4.C.v.d, I consider a circumscribed case where one might be sympathetic to the anti-ATLU-friendly intuition. The mechanism I propose here, for this circumscribed case, cuts in the opposite direction of the primary explanation that I will have proposed in Section 1.4.C.v.c. The reader should keep in mind throughout this section that there might be various other possible explanations and sources of people’s anti-Premise 1 intuitions, and, thus, the comments here do not attempt to provide an exhaustive or complete account. To the extent that one’s anti- Premise 1 intuitions derive from any of the sources I describe, they should be discounted. However, if one believes that one’s anti-Premise 1 intuitions do not derive from these sources, then this section will not provide any additional support for Premise 1. But, even if someone who rejects Premise 1 does not believe his intuitions are explained by these considerations, no damage will have been done to my position. Rather, I will only have failed to provide additional support for my position. 1.4.C.v.b. Interestingly, I think that perhaps the main reason why people might reject Premise 1 is not directly a function of the plausibility or non-plausibility of Premise 1 itself. As discussed in 120 section 1.3, although Premise 1 itself might raise some eyebrows, the conjunction of Premise 1 and Premise 2 (AMBP) is particularly likely to raise eyebrows, because an implication of this conjunction is that the (seemingly plausible) IMI is false (i.e. that it is false that, all else equal, it is better agent-neutrally that a happy person live longer). I think that concerns about this implication might often be what ultimately lead people to reject Premise 1. There are two ways in which this reductio-type concern could result in one’s rejecting Premise 1. First, there might be individuals who don’t find Premise 1 by itself false, but who think that asserting its falsity is the most plausible response to the reductio-type concern. Second, there might be individuals who do find Premise 1 by itself false, but who unconsciously have developed this intuition in response to the reductio-type consideration lurking around the corner. It could well be that absent this reductio-type motivation for rejecting Premise 1, there would be nobody who would reject Premise 1. If so, then the hypotheses that follow in Sections 1.4.C.v.c and 1.4.C.v.d would not only be false, but they would be unnecessary, because Part II Chapter 2 will explain why rejecting Premise 1 is not the best solution when confronted with the fact that the conjunction of Premise 1 and Premise 2 is incompatible with IMI. Thus, if it were the case that nobody would reject Premise 1 if not for the reductio-type concerns, this would make my position easier to defend. Unfortunately, however, I think that there are people who have anti- Premise 1 intuitions who would still hold these intuitions even absent the reductio-type motivation for rejecting Premise 1. If I am right about this, then the following hypotheses about possible sources of anti-Premise 1 intuitions are not unnecessary (and they will not necessarily be false). In any case, to be safe, I will offer them. Thus, while the reductio-type motivation for rejecting Premise 1 will be addressed in Part II Chapter 2, for the time being I will focus on the plausibility of Premise 1 itself. Therefore, in what follows, I will limit the discussion to comparisons that isolate the question of whether 121 Premise 1 is true. In other words, the comparisons in what follows will be between cases that contain the same total number of moments of happiness, in which the moments of happiness are of the same intensity, and where the sole difference is in the number of persons who share these moments—and thus, in the length of these lives. Constructing the cases in this way will minimize the number of moving parts and focus the inquiry on Premise 1 itself. 1.4.C.v.c If the unit relevant for ethical inquiry is the conscious moment, all states of affairs fitting the mold of Premise 1 are equally good. While different views might be held in different cases, it seems that a somewhat prevalent intuition might be the intuition supporting ATLU—that the states of affairs in this mold will be better the fewer the number of people. Why might this be? I think that the intuition likely is “Happiness itself is not good, but happiness is good for people.” Of course, in any of the cases we are considering, the units of happiness will be experienced by people. However, I think that the ATLU-friendly intuition can be explained by good old person- affecting intuitions. Suppose one is choosing between doubling the life of person x (from 50 years to 100 years) or instead adding a person y who will live 50 years. It seems that x will be benefited by the former choice and no one will be harmed, whereas the second choice will harm x and benefit no one. As such, person-affecting intuitions suggest that the first option is the better one. As discussed in Part I Chapter 2, accounts can be given of how bringing a person into existence or failing to do so could benefit or harm him, respectively, but reasons were offered to doubt the plausibility of those accounts. Thus, to the extent that one is sympathetic to person-affecting intuitions, one will likely find some plausibility in the claim that the first choice is the better one. Instead of the above case, however, one might instead consider the case where neither x 122 nor y exists yet and where one is deciding between creating a world where x and y both exist for 50 years or a world where just x lives, but for 100 years. In this case it seems that the above person-affecting mechanism would not be applicable. It seems to me, though, that one’s intuitions in the first situation are likely to be infiltrating one’s assessment of the second situation, and thus that the person-affecting intuitions are still responsible for the assessment. Importantly, however, the shortcomings of person-affecting intuitions were discussed at length in Part I Chapter 2, and it was shown why they are unworkable in the context of population ethics and, thus, why we need an impersonal morality. More specifically, it was argued that person-affecting intuitions are heuristics for furthering impersonal moralities, and that they are highly useful in most cases we encounter, but that they are not designed for, and thus not helpful for, questions in population ethics. Thus, if person-affecting intuitions are in fact responsible for the ATLU-friendly assessment, these intuitions should be discounted. 1.4.C.v.d. While I think that the above explanation might be the driving factor that explains people’s anti-Premise 1 intuitions, there are two additional points worth making. First, I offer another explanation of ATLU-friendly intuitions that cover some types of cases. Second, I offer an example of a case where intuitions might go against Premise 1 but also go against ATLU, and I offer an explanation of why this might be the case. First, one might have ATLU-friendly intuitions in cases where the happiness units are spread out across a very large number of lives of minuscule duration. In addition to the person- affecting intuitions just discussed, it could also be the case that the intuition here that this would be bad could be explained by a different thought: One might think such a short existence would be too short to enjoy things that are sufficiently complex that they require more time. This concern is 123 simply denying that the happiness at the moment is as high as it has been stipulated that it is (or instead, denying that the moments in the longer life are of no greater intensity when it has been stipulated that they are not). Additionally there also might be non-utilitarian considerations seeping into the minuscule duration case. Second, one might have the anti-Premise 1 intuition that cuts in the direction opposite from ATLU: One might think that it would be better if one (or few) long life (lives) were distributed among more people, and this intuition might be strengthened if the long life or lives were particularly long. This intuition could perhaps be traced to an implicit belief that there would be diminishing returns for additional years for the person who lives the long life. Of course, this has been stipulated to not be the case, though, when the case was described in utility levels. Alternatively, and perhaps somewhat relatedly, intuitions based in non-utilitarian concerns such as equality or distributional justice could be seeping into this case as well—even though the individuals allegedly benefiting from this would not have existed in the state of affairs with the single long life, and thus equality and distributional justice would have had to seep into this case from a case in which they would be relevant, even before seeping into the allegedly utilitarian assessment. 1.4.C.v.e. As I mentioned, I do not mean to suggest that the hypotheses offered in this section constitute an exhaustive or complete account. There might be various other possible explanations and sources of people’s anti-Premise 1 intuitions. Here, I only offered a few. However, to the extent that one’s anti-Premise 1 intuitions derive from any of the sources I describe, they should be discounted. If, however, one’s anti-Premise 1 intuitions do not seem to derive from these sources, so be it, and this section will thus have not provided further support for my position. This 124 notwithstanding, even if nothing is gained in this section, no damage will have been done to my position. I will only have failed to provide additional support for my position. 1.4.D. Summary In Section 1.4, I have argued for the plausibility of Premise 1 of my four-step argument for AMU, and, in so doing, I have argued that the unit relevant for ethical inquiry is the conscious moment. After introducing the relevant terminology, I began by exploring Huemer’s argument. Although I explained why I Huemer’s argument fails, I think that it was fruitful to discuss for two reasons: First, the framework of equivalences and dimensions that Huemer sets up is useful for our inquiry, and second, taking notice of his mistakes can help us avoid committing the same or similar ones. After discussing Huemer, I then showed how the AMU theorist has an ally in the TU theorist in attempting to defend Premise 1. Both AMU and TU endorse Premise 1. Thus, although the support from TU theorists in no way proves that Premise 1 is plausible, it suggests that the vast majority of utilitarians in history (who have been TU theorists) have found Premise 1 to be plausible, and history does provide some support for my premise. However, what is key for me, of course, is not merely to show that utilitarians have thought that Premise 1 is true, but, rather, I must show that this is what a utilitarian should think. I then turned to the heart of the matter and provided reasons to think that Premise 1 is plausible. While I provided a number of different arguments to this effect, underlying many of the arguments was the single theme that constitutes the backbone of my objection to ATLU: ATLU is a puzzling account that is unsteady and difficult to justify because of its hybrid nature. After offering these arguments I then finished by considering some possible explanations of why people might hold ATLU-friendly intuitions, and I suggested that if these hypotheses correctly identify the source of one’s ATLU-friendly intuitions, then these intuitions should be discounted. 125 In sum, I hope to have provided reasons to accept the plausibility of Premise 1. The arguments I provide are not knockdown arguments, and an impasse of intuitions might continue to exist, but I think that the considerations here have force. At the absolute minimum, I hope to have shown that the ATLU theorist owes us an explanation of why we should find ATLU plausible despite its puzzling features. 1.5. The metaphysics of personal identity and how it supports Premise 1, and, more generally, AMU Recall that Premise 1 in my argument for AMU states: “For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than in y lives.” Although the ATLU theorist agrees with AMU that it is not better, all else equal, for there to be additional conscious moments in the form of additional lives, the ATLU theorist denies Premise 1, arguing that it is better that there exist additional happy conscious moments in the life of an already existing person. Thus, the ATLU theorist’s rejection of Premise 1 relies on his claim that, all else equal, although it is not better for there to be more good conscious moments if they exist in an additional person, it is better for there to be more good conscious moments if they exist in an already existing person. Although this dispute has already been approached from a couple of angles in this chapter, there is another approach that can be taken. In light of the two different views about the relevance of persons to the location of conscious moments, it will be helpful to address the prior question of what a person is, and then see if our metaphysical conclusions have implications for the debate between AMU and ATLU. This will enable us to approach the dispute in greater depth and from an additional angle. In this section, I will begin by defining a reductionist theory, and I will describe the two traditional accounts of personal identity—both of which are reductionist. I will then show both why these accounts are unsatisfactory and why a non-reductionist approach would not be fruitful. I 126 will then proceed to describe and endorse Parfit’s proposed solution to this problem: the empty question view of personal identity. I will then explain that, if true, the empty question view would provide strong reason to accept Premise 1, and further, having endorsed the empty question view, I will suggest that we have strong reason to accept Premise 1. In the second half of this section, however, I will consider a move that my opponent might make in response to these claims. An opponent of mine might concede that the empty question view is true and that this provides support for Premise 1, yet now—influenced by the issues raised in the discussion of the empty question view—offer a response to the four-step argument that has not yet been considered in this chapter. He might argue that, if we accept the empty question view, AMU will only follow if we can show that a slightly altered version of Premise 1 is true. This premise, which I will call Premise 1A, states: “For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x packets of Relation R continuity rather than in y packets of Relation R continuity.” In the second half of this section, I will explain what this premise means, I will consider whether or not we should accept it, and ultimately I will argue that we should accept it. Thus, having provided additional reasons for accepting Premise 1 and having provided reasons for accepting Premise 1A, I will conclude that our metaphysical conclusions will have implications for the debate between AMU and ATLU, and, in particular, our metaphysical conclusions will provide additional support for AMU. These claims notwithstanding, however, AMU certainly does not rely on the empty question view of persons or even on a reductionist metaphysics of persons. The considerations here merely provide additional reasons to believe AMU. Thus, neither the arguments of this chapter nor those of the dissertation as a whole will stand and fall with the success of the considerations in this section. As such, after this section—both in this chapter and in later 127 chapters—it will not be assumed that the reductionist view is true or that the empty question view is true. I should add one additional note: As will quickly be clear, this section as a whole draws heavily on Parfit’s discussions in Reasons and Persons, and for that I am deeply indebted to him. 1.5.A. Traditional approaches to personal identity and the problems that afflict them A reductionist theory is one according to which any true statements in the reduced vocabulary are true in virtue of facts that can be stated in the reducing vocabulary. The paradigmatic example of a reductionist theory is the reduction of heat to molecular motion: Any true statements about heat are true in virtue of facts than can be stated in terms of molecular motion. Similarly, ever since Locke, and perhaps much further back, philosophers have attempted to provide reductionist accounts of personal identity. The question asked is “What makes a person at time x and a person at time y one and the same person?” In response, philosophers have attempted to provide necessary and sufficient conditions for when personal identity obtains, and in so doing they attempt to reduce the term “personal identity” to statements in a different vocabulary. Though the arguments and innovative thought experiments have proliferated, the theories of personal identity still generally fall into two main camps: theories that offer a psychological criterion of personal identity and theories that offer a bodily (or brain-based) criterion of personal identity. There are different subtypes of the two accounts and there are also hybrid accounts. The debate continues, though, without there being a particular account that is obviously satisfactory. As Parfit argues, although bodily and psychological theories are in fact the best attempts to explain personal identity, both suffer from similar insurmountable difficulties. There are two general types of difficulties, each of which afflicts both the psychological theory and the bodily 128 theory. First of all, there are difficulties related to what Parfit calls spectra cases. 120 Consider a psychological version of the spectrum case, though the same comments can be made with respect to a bodily version and a combined version. Consider one’s current self and then consider a wide range of possible persons the identity to one’s current self of which is in question. On one end of the spectrum is a perfect psychological duplicate of one’s current self (with regard, for example, to memories, intentions, desires, beliefs, values etc.), and on the other end of the spectrum is a being that shares no psychological traits of one’s current self. On a typical psychological theory, if one’s current body is destroyed and the identical duplicate is created, it is clear that this person would be identical to the person whose body was destroyed. Similarly, if instead the body at the far end of the spectrum were created, it would be clear, on the psychological theory that this person is not the same person as he whose body is destroyed. The problem, however, is that one can consider a spectrum of possible replicas, each of which has one fewer psychological feature in common with the prior replica, and it seems that a psychological (or bodily) theorist who thinks there is a fact of the matter to personal identity would be forced to state a threshold of similarity separating the cases of replicas that are and are not the same person as he whose body was destroyed. However, it seems implausible that the addition of one particular memory or one particular brain cell could make the difference between Person X being and not being the same person as Person Y. As Parfit argues, if personal identity is a function of bodily or psychological connections or continuities, which come in degrees, it is implausible to think that personal identity is all-or-nothing and does not come in degrees. The second insurmountable difficulty afflicting theories of personal identity are so called fission cases. 121 It has long been known that this is a difficulty for psychological theories. If a 120 Parfit, p.231. 121 For illuminating and thought-provoking discussions of fission cases, see Jake Ross’s “Divided We Fall: Fission and the Failure of Self-Interest,” Philosophical Perspectives, 28, Ethics, 2014, 222-62 129 psychological theorist thinks that cases of teletransportation (where a person’s molecules are scanned, his body destroyed, and a new body created elsewhere with the blueprint of the original body) are cases that can preserve personal identity, the question arises what we should say if there is more than one qualitatively identical body resulting—be it either due to the machine not destroying the first body or due to the machine creating replicas in more than one location. There are various possible explanations a psychological theorist can give about what goes on in these cases, but none of them are particularly plausible. The most common response is to amend the psychological theory by adding a “no-branching” clause, making it the case that the replica is only identical to the first person if he is the only replica. As Bernard Williams, among others (including Parfit), have said, though, this is an implausible patch because it makes “something very important depend on seemingly insignificantly minute facts.” 122 Williams thought that this provided reason to prefer the bodily criterion, but as Parfit points out, research in the last fifty years on split brain patients 123 and on patients with one hemisphere of their brains destroyed, indicates that half a brain would be sufficient for a person to survive. In light of this, it seems theoretically possible that a person’s two hemispheres could be transplanted into different bodies, and thus the bodily criterion would be afflicted by the fission objection as well. Williams gestures at some reasons to deny that these examples have similar force, but his explanations are lacking. Before moving on, it will be helpful to elaborate on and clarify these two objections to the bodily and psychological criteria of personal identity. In fact, even Parfit’s discussion itself could use some clarification and more explicit discussion of the following. It’s important to note the role that is played in these arguments by a person’s feeling at t2—be it a true feeling, a mere illusion, or neither 124 —that he is the same person as a person (in the same body or in a different body) at t1. 122 Williams, p.209. 123 Parfit, p.245. 124 The “neither” option will be explained and discussed below when I introduce the empty question view. 130 The reason that the two objections mentioned have force is that when we think of the notion of a person being the same person at t1 and t2, we inextricably have this notion linked to some internal perception of sameness by the person. (Though it might seem to at first, this does not presuppose the psychological criterion. This perception of sameness by the person is consistent with the bodily criterion as well.) Thus, the reason, or at least a large reason, for why the spectra objection and duplication objection work is that it is difficult to imagine that this perception of sameness would exist in these cases. In the duplication case, it is difficult to imagine that both bodies at t2 have a perception of sameness as each other. In the spectrum case, it is difficult to imagine that the perception of sameness, which appears to be an all-or nothing perception, would be or could be affected by a single added memory. Thus, since we doubt that the perception of sameness would obtain in these examples in the way that the theories of personal identity respectively say that they would, we don’t think that the theories provide a plausible account of the perception of sameness. And then, since our notion of personal identity is inextricably linked to the perception of sameness, we think that these objections show that the theories in question do not provide a plausible account of personal identity. There are many more nuances in this subject than included in this brief summary, but, as indicated here, there are strong reasons to think that we have not yet found a plausible theory of personal identity. Confronted with these difficulties in articulating a reductionist account of personal identity, one might consider a non-reductionist account. What would a non-reductionist account look like? As Parfit describes the position, non-reductionists believe that we are “separately existing entities” 125 —that personal identity consists in something beyond our physical or psychological continuity. The best-known version of this view is the Cartesian Ego. While perhaps 125 Parfit, p.245. 131 not incoherent, views of this sort are mysterious at best. What then are we to do? As Parfit says, non-reductionism is a non-starter, and we have yet to find a plausible reductionist theory. According to Parfit, however, the fact that we do not have a satisfactory reductionist account of personal identity is not due to a failure of the relevant theorizing, but, rather, it is due to the fact that we have misunderstood the nature of persons at a more fundamental level. I will now describe Parfit’s “empty question” 126 view of personal identity. 1.5.B. The “empty question” view of personal identity According to Parfit, the error in the personal identity literature up to this point has been to hold what he calls the “Simple View” of personal identity. 127 According to the Simple View, facts about personhood and personal identity have a special nature: “They are further facts, independent of certain more specific facts; and in every case they must either hold completely, or completely fail to hold,” and further, they hold in a one-to-one manner. 128 Parfit thinks we should instead adopt a view according to which personhood consists only in specific facts (e.g., bodily or psychological connections), and thus is both a matter of degree and not necessarily a one-to-one relation. According to Parfit, if we adopt such a view, we will no longer be afflicted by the two insurmountable problems discussed above—spectra cases and duplication cases. Importantly, for Parfit, the question of whether person X, at t1, is the same person as person Y, at t2, is an empty question—there is no fact of the matter. This can be best understood by way of analogy to entities that it is uncontroversial to understand through an empty-question 126 Parfit himself actually calls his view “the reductionist view.” However, it seems to me, that even the views that he distinguishes his view from are reductionist in nature. As such, I have taken the liberty to give his view a name that seems to be a more appropriate one. 127 Parfit, p.227. 128 Ibid. 132 lens, such as clubs or nations (or, even less controversially because it does not involve people, a boat). Suppose, with Parfit, that there is a club or a nation that undergoes various changes—be it in membership or procedures or strategies and actions that it carries out. If, at t2, we ask whether the club (or nation) is the same club (or nation) as at t1, Parfit says: The claim ‘This is the same club’ would be neither true nor false. Though there is no answer to our question, there may be nothing that we do not know. . . . If we know all the facts about how people held meetings . . . This is why we would not be puzzled when we cannot answer the question, ‘Is this the very same club?’ We would not be puzzled because, even without answering this question, we can know everything about what happened. If this is true of some question, I call this question empty. . . . Different answers to our question are merely different descriptions of this fact or outcome. . . . There are not here two different possibilities, one of which must be true. 129 Thus, Parfit thinks that this is the case for personal identity. If we know all the specific facts of a situation, this is all there is to know. Different answers to the question of personal identity, “Will it be me?,” are merely different descriptions of a single outcome. Despite the fact that Parfit thinks that the identity of a person is less metaphysically “deep” 130 than we thought when we adhered to the Simple View, Parfit does still think that connections between the past and the future can potentially matter. According to Parfit, just as we thought that surviving as the same person in the future was important, and all-or-nothing, potentially of similar importance is whether there are selves in the future that bear “Relation R” to our current selves. Relation R, for Parfit, is just like personal identity, except for two differences: First, Relation R need not be a one-to-one relationship, and thus we can bear Relation R to multiple future selves. Second, Relation R comes in degrees and is not all-or-nothing. On Parfit’s account, the relevant specific facts that make up Relation R are certain psychological connections, such as relations of intention, memory, and character, and these connections can take the form of connectedness (the holding of particular direct connections) and continuity (the holding of overlapping chains of connectedness—chains that I will hereafter call 129 Parfit, p.213. 130 Ibid, p.138. 133 “Relation R continuity packets”). They are specific insofar as there is clearly a fact of the matter as to whether or not they obtain. Again, it’s important to add a clarification to this account that would have been useful in Parfit’s text itself. It’s important to be clear that these specific facts can refer to various types of connection, many of which might be very narrow as a particular memory might be, but also some of which might be broader. For example, if a person at t2 has the feeling that he is the same person as a person at t1 (be it a true feeling, a mere illusion, or neither), this is a specific fact. Thus, even this type of a fact should be understood to count as a connection. It’s just that this feeling of sameness would only constitute one specific fact—among many. Importantly, however, unlike the answer to the question of whether a person at t2 and a person at t1 are one and the same person, there is a fact of the matter about whether a person at t2 has the feeling that he is the same person as a person at t1. Thus, the existence of this feeling would constitute a specific fact in Parfit’s framework. It is important to note, however, that a proponent of the empty question view need not endorse a psychological version of Relation R. One might espouse a version of Relation R grounded in bodily connections or perhaps a hybrid of psychological and bodily connections. Though this is a very interesting question (which in many ways is just as provocative and perplexing as the debate between proponents of different criteria of personal identity on the Simple View), what is relevant for my purposes is not an assessment of “what matters for survival,” but rather, the key notion of the empty question view of personhood. As such, I will bracket the question of how best to fill out a plausible description of Relation R. I will return in more detail to Relation R, however, below, when considering the implications of Parfit’s empty question view. Parfit concedes that people might, at least at first, find the theoretical shift from identity to 134 Relation R to be a counterintuitive one. After all, one seemingly would want a future person to not just be R-related to the person that one currently is, but rather, for the future person to be the same person. According to Parfit, however, we should not view science-fiction-type cases where one is “merely” R-related to a future person as being any step down from what we experience in daily life, and this is because even everyday cases do not involve personal identity, but rather they merely involve Relation R. As such, since we are content with the connections between our current and future selves in everyday life, we already find Relation R to be satisfactory, and we should not dream of anything further—such as identity. Parfit illustrates this point nicely: “When I fear that, in teletransportation, I shall not get to Mars, my fear is that the abnormal cause may fail to produce this further fact. As I have argued, there is no such fact. What I fear will not happen never happens. I want the person on Mars to be me in a specially intimate way in which no future person will ever be me. My continued existence never involves this deep further fact.” 131 132 Parfit acknowledges that his empty question theory is, prima facie, implausible in some ways, but he argues that it is considerably less implausible than the other available options— namely either, on the one hand, espousing a non-reductionist view, or, on the other hand, rehabilitating the psychological or bodily theory of personal identity in one of various ways so that it avoids the two objections discussed above, in section 1.5.A. I wholeheartedly agree with this conclusion. 1.5.C. A possible reply and the need to show that Premise 1A is plausible Recall that the goal of this section has been to provide support for Premise 1: “For all 131 Parfit, p.279. 132 Though I think that Parfit’s points regarding teletransportation here are correct, I do think that there is one confounding factor: Ultimately I think that bodily continuity plays a necessary role (perhaps in addition to there being a role for psychological continuity) in a defensible account of Relation R, and, as such, the teletransportation case might be particularly scary because it won’t even satisfy the relevant account of Relation R. 135 numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x lives rather than y.” If the empty question view is correct, this provides strong support for Premise 1. If there is no fact of the matter about (i.e. if it is an empty question) whether a person at time x and a person at time y are one and the same person, then there is no fact of the matter about how many different lives exist. If there is no fact of the matter about how many different lives exist, then Premise 1 is trivially true. It certainly cannot be better or worse that z number of conscious moments exist in x lives rather than y lives if there is no fact of the matter about how many lives there are. This is not the case due merely to an inability to obtain data about how many lives there are, but, even more seriously, it is the case because no such facts exist. Furthermore, I have argued not only that we would have strong support for Premise 1 if the empty question view were plausible, but also that the empty question view is plausible, and thus that we do have strong support for Premise 1. Supporting Premise 1, however, while an important goal, is, of course, merely an intermediate goal. The broader goal is to support my four-step argument for AMU. In light of this, an opponent of mine might concede that the empty question view is true and that this provides support for Premise 1, yet now—influenced by the issues raised in the discussion of the empty question view—offer a response to the four-step argument that has not yet been considered in this chapter. My opponent might argue that, even if Premise 1 and Premise 2 are (trivially) true, the truth of the lemma does not follow, and thus the conclusion does not follow. Instead, my opponent might say, if the empty question view is true, Premise 1 and Premise 2 need to be replaced, respectively, by Premise 1A and Premise 2A (see below), and only if the truth of these two premises can be established, will the truth of the lemma and of the conclusion follow. Thus, consider the following version of the four-step argument: Premise 1A: For all numbers x, y, and z, all else equal, it is not better or worse that z number of conscious moments exist in x packets of Relation R continuity 136 rather than in y packets of Relation R continuity. Premise 2A: All else equal, it is not better or worse that there exist more packets of Relation R continuity than fewer packets of Relation R continuity. Lemma: All else equal, it is not better or worse that there exist more conscious moments than fewer conscious moments. Conclusion: AMU is true. As for the shift from Premise 2 to Premise 2A, arguing for the truth of this premise should not occupy our time here, and this is the case for two reasons. First of all, the rationales underlying Premise 2 and 2A are the same, so it seems that if one believes Premise 2, one will also believe Premise 2A. Further, as discussed in section 1.3, Premise 2 (and similarly Premise 2A) is an expression of AMBP, and this is a view that was supported in Part I Chapter 2. As such, the truth of Premise 2 was stipulated for the purposes of this chapter, and the same should go for Premise 2A. While 2A can be contested, here is not the place to do so. Supporting Premise 1A, however, not only is crucial for the success of my argument, but it is crucial to support it here. The task of assessing whether Premise 1A is true will require us to consider the different normative implications that the empty question view might have—implications for rationality and for ethics. Sections 1.5.D and 1.5.E will do so, and they will provide reasons to believe that Premise 1A is true. 1.5.D Possible views about the implications of the empty question view Adopting the empty question view of persons will have normative implications. Parfit offers two different views that one might take, and ultimately he argues that both provide support for utilitarianism. I will return, in section 1.5.F to his claims about downstream implications of the two views, and the implications for utilitarianism in particular, but what is most relevant for our purposes at this point is his discussion of the two views themselves. There are different views that one might hold about the implications of the empty question view, and Parfit introduces two possible views that he calls the “moderate view” and the “extreme 137 view.” These are frequently mistakenly taken to be diverging descriptive views about the empty question view itself, and this is exacerbated by the fact that Parfit refers to them as different empty question views, but the moderate and the extreme views in fact are about rationality and ethics, and thus involve normative claims. According to Parfit, whether one adopts the moderate or extreme view with respect to the empty question view’s implications for rationality is conceptually and logically distinct from whether one adopts the moderate or extreme view with respect to the empty question view’s implications for ethics. As such, a proponent of the empty question view could hold one view in one context and the other view in the other. This notwithstanding, there are similarities between the moderate view in the two contexts, and also, similarities between the extreme view in the two contexts. Parfit writes: “Some writers claim that, if the [empty question] view is true, we have no reason to be concerned about our own futures. I call this the Extreme Claim.” 133 According to proponents of the extreme view, this is the case even though, as a matter of fact, special concern for one’s own future exists and has been selected for by evolution. Parfit writes: “Such concern would remain, as a natural fact, even if we decided that it was not justified. By thinking hard about the arguments we might be able to briefly stun this natural concern. But it would soon revive . . . Since there is this explanation, we would all have this attitude even if it was not justified; so the fact that we have this attitude cannot be a reason for thinking it justified. Whether it is justified is an open question, waiting to be answered.” 134 On the moderate view, however, unlike on the extreme view, the claim is that Relation R gives us a reason for special concern, and that “extremists are wrong to assume that only the deep further fact gives us a reason for special concern.” 135 Parfit himself does not choose between the 133 Parfit, p.307. 134 Ibid, p.307. 135 Ibid, p.311. 138 moderate and extreme claims in the context of rationality. According to him, both positions are defensible and it could be that we will be unable to show that one or the other should be rejected. He writes: “There is a great difference between the Extreme and Moderate claims. But I have not yet found an argument that refutes either.” 136 After discussing the moderate and extreme views in the context of rationality, Parfit proceeds to consider their counterparts in the context of ethics. He addresses the implications of the empty question view of persons for desert, commitments, and, much more at length, distributive justice. As Parfit says, according to the extreme view in the context of ethics, we might believe that only the further fact of personal identity (the existence of which the empty question view denies) “enables some later benefit to compensate some earlier burden.” Thus, on this view, compensation “does not merely presuppose personal identity . . . it presupposes the deep further fact.” 137 Thus, since this further fact does not exist, no compensation is possible at all. On this view, the only “relevant units are the experiences that people have at any particular time,” 138 and thus there is no sense in which maximization across these atomic units within persons is any different morally from maximizing across these atomic units across different persons. Thus, on the extreme view, we should extend the principle of maximization that we use in intrapersonal choice to the interpersonal context. In other words, these considerations suggest that the relevant unit for ethical inquiry is the conscious moment. According to this view, Premise 1A is true. Packets of Relation R continuity are not the unit relevant for ethical inquiry, but conscious moments are the relevant units. Thus, on this view, it does not matter intrinsically for ethical purposes how many packets of Relation R continuity a certain number of conscious moment are distributed among. 136 Ibid, p.312. 137 Ibid, p.342. 138 Ibid, p.344. 139 According to Parfit, the moderate view in the context of ethics states that Relation R does provide grounds for compensating a person at t2 who is R-related to a person at t1. Relation R not only gives us reason for special concern for ourselves, but it also is an ethically relevant notion. This notwithstanding, Parfit thinks that the grounds that Relation R provides for compensating a person at t2 who is R-related to a person at t1 are weaker than the grounds that would be provided if the person at t1 were the same person as the person at t2. This is because Relation R comes in degrees, and most cases of Relation R’s holding will be significantly less than a 100% connection. According to Parfit, the grounds for distribution will be weaker than they would be on the Simple View, because, if personal identity is less “deep,” then there is less to the fact that a person who bears a burden to day is the same person who is compensated tomorrow, and thus compensation and equality seem less important (though perhaps not unimportant). Thus, on this view, Premise 1A would be false. It’s not clear precisely what the relevance would be of the number of packets of Relation R continuity, but there could be some relevance, and thus Premise 1A would be false. As was the case in the context of rationality, Parfit thinks that the moderate view and the extreme views are each defensible and that neither one can be refuted. As such, he does not choose between the two views. The question now confronting us is whether we should accept Parfit’s claims and, if so, whether we should adopt the extreme view or the moderate view (or, as Parfit does, remain agnostic about their relative plausibility). To this end, it is useful to consider some objections. 1.5.E. Objections and an additional possible view Parfit’s claims about the implications of the empty question view have come under attack from various angles, and the critiques fall into two different categories: Those that allege that either the extreme view or the moderate view (or both) lack internal coherence, and those that 140 allege that the views, while perhaps coherent, lack plausibility for other reasons. As one example of an objection regarding internal coherence, consider the following objection to the extreme view: It has been argued that it might deal adequately with some issues, but that it fails because it “overlooks the fact that any moral theory requires some coherent notion of a minimally unified moral agent capable of putting the theory into effect, acting in accordance with it, and so on.” 139 Parfit, however, has responded in what I believe to be a successful manner to this concern, raised by Scheffler, 140 and I don’t think we need to focus here on objections of this sort. Similarly, objections have been raised against the internal coherence of the moderate view: Some authors have voiced concerns about the “vague, metaphorical language about the depth of certain facts and the weight of moral principles” that Parfit’s moderate view employs. 141 These concerns in this context are not to be ignored, but Parfit has plausible responses and explanations. More importantly, however, we should shift our attention to the second type of objection to Parfit’s views, because they cut more deeply and reach the heart of the issue that interests us. This type of an objection is the type that contests the general plausibility of Parfit’s two views. As they tend to be offered, however, they are slightly misplaced, and the confusion here seems to be related to the confusion, discussed earlier, regarding the fact that the extreme and moderate claims are normative claims and not descriptive ones. The objections in this camp are generally couched as objections to the internal features of or the implications of the moderate view, but what these objections amount to are not objections to implications of the moderate view (which is already a view about implications of the empty question view), but rather, arguments for rejecting the moderate view and espousing a third view. Thus what the authors below should do is state their claims as explanations of why they favor a third view about the implications of the 139 Scheffler, p.230. 140 Parfit. 141 Daniels, p.269. 141 empty question view of personal identity—a view that we might call the “no difference” view. A view of this sort is offered by various authors, including, among others, Bart Schultz and Susan Wolf. 142 According to Schultz and Wolf, even if there is a connection between metaphysical depth and moral weightiness, the moderate view doesn’t give us any reason to believe that the relevant unit is the conscious moment and not the person. 143 As they argue, even if the empty question view about persons is correct, the fact of the matter is that we still have Relation R and Relation R seems to be sufficient to ground our ethical intuitions. As I stated, their claim should not be understood as contesting the implications of the moderate view, but rather, it should be understood as offering a third view, the “no difference” view. As mentioned above, if the extreme view is correct, Premise 1A is true. If, however, the no-difference view is correct, then Premise 1A is false. If the moderate view is correct, Premise 1A will also be false, though it’s unclear precisely how important the distribution of conscious moments into packets of Relation R continuity will be on the moderate view as compared to on the no difference view. With these three views about the ethical implications of the empty question view now on the table, which one should we endorse? For the reasons below, I think that the extreme view is the most plausible and that the moderate view and the no difference should be rejected. In arguing for this conclusion, I will not separately address the moderate view and the no difference view, but rather I will argue in one fell swoop that Relation R should not matter ethically at all. In my argument I will explicitly address only the no-difference view, but the arguments apply mutatis mutandis to the moderate view. Before continuing, it’s important to mention that I think that the extreme view in the context of ethics is correct even if we endorse the moderate view or the no difference view in the context of rationality. We must remember that claims about rationality are claims about agent-relative goodness. In the context of ethics, 142 Schultz; Wolf. 143 Schultz; Wolf. 142 however, we are seeking an agent-neutral theory, and there can be a gap between the two. 1.5.F. Why Relation R should not matter ethically According to the no-difference view, the unit relevant for ethical inquiry is the packet of Relation R continuous conscious moments, as described in section 1.5.B. To the extent that this is the case, the no-difference view avoids the difficulties of the empty question view of personal identity, because there is a fact of the matter about whether Relation R holds, and thus it is possible to know how many packets exist. So far so good. The difficulty with the no-difference view, however, is that it seemingly will not converge with people’s intuitions, and this divergence is most apparent when we consider duplication cases and spectra cases. For the purposes of this discussion, I will discuss a psychological version of Relation R, but the points I make apply to bodily versions of Relation R as well. First consider spectra cases. Suppose that one memory of person x at t1 is inputted into person y at t2 and that the body of person x is destroyed just after t1. According to the no- difference view, Relation R obtains between these two persons, despite the minimal amount of connectedness, and thus x and y would be part of the same packet or Relation R continuity. It is difficult to believe that this packet should be the unit relevant for ethical inquiry. The original intuition was that a single person’s lifetime should be the unit relevant for ethical inquiry, and the mere sharing between two persons of a single memory does not seem to be a compelling reason for considering two persons to be portions of the same unit relevant for ethical inquiry. Further, if we approach things in a fine-grained enough manner, it seems that it is a common natural occurrence that many people will have an identical memory of, say, seeing the President take the oath of office on television. One shared memory does not seem to be sufficient for rendering persons part of the same continuity packet that is a unit for the purposes of ethical inquiry. 143 In response to this concern, one might offer the following explanation of why the above is implausible, and offer the following tweak to Relation R: While Relation R is described in terms of continuity and connectedness, it’s important to not forget that Relation R itself is a mental notion, and although the mental notion is being reduced to connections, it is important to not forget the mental notion itself. Instead of calling it Relation R, it might be easier to grasp if we call it the illusion-of-being-the-same-person 144 relation—or Relation I. Now I will explain both why Relation I will maintain Relation R’s ability to avoid the empty question concerns, and why it avoids the one-memory problem mentioned in the previous paragraph. The first way in which Relation R avoided empty question concerns was that it was not exclusively a one-to-one relation. Relation I maintains this feature, as more than one person (e.g. x and y) at t2 could potentially have the illusion of being the same person as person z at t1. The problem with Relation R, however, is twofold: First, the mental state of feeling that one persists through time does not appear to be felt in degrees, but rather it is all-or-nothing. Second, it doesn’t seem as though this illusion would be instantiated by one connection, as described in the case above. Thus, it seems that Relation I is an improvement if it requires of connectedness that there be enough connections for this illusion to exist. Thus, Relation I will involve a threshold notion of required connections. At first glance, it might seem that, with this change, Relation I loses the second feature of Relation R that avoided the empty question concern associated with spectra cases. This, however, is not so. While there might be no fact of the matter if two persons are one and the same person, there is a fact of the matter of whether there exists an illusion of being the same person. It may be mysterious what type and quantity of connections bring about this illusion, but this does not matter for our purposes, because what matters is that 144 By referring to this as an “illusion,” I do not mean to limit the understanding of this to psychological accounts of Relation R. The term “illusion” is meant to be understood broadly and encapsulate any account of how this illusion is experienced or what the illusion is. 144 there is a fact of the matter of whether the illusion exists. 145 Thus, there is also a fact of the matter about how many continuity packets of this sort exist. Thus, in this sense, Relation I is able to maintain the two important features of Relation R, yet it is able to give us the seemingly correct result in the problematic one-memory case described above. Thus, with the move from Relation R to Relation I, it seems that spectra cases might not pose a problem for the plausibility of using packets of continuity of this type of relation as the unit relevant for ethical inquiry. As such, it might seem for a split second as though Premise 1A could be false. Premise 1A, however, is not false, and this is because neither Relation R nor Relation I is able to give a plausible account of duplication cases. Suppose that, at t2, there exist two persons that each are duplicates of a person that existed at t1. In light of this, each person at t2 will be connected to the person at t1, and thus the two persons at t2 and the person at t1 will all be part of one packet of Relation R (or Relation I) continuity, and thus they will be one unit for the purposes of ethical inquiry. (Of course, this is because a Relation R (or Relation I) continuity packet would have to be an equivalence relation, which would mean that the two persons at t2 would be in the same packet.) While not incoherent, this will seem implausible to many who see that the two persons at t2 appear to go about life separately 146 and can even interact with one another if in the same room. Furthermore, this concern about duplication is just as serious for Relation I as it is for a view that does not maintain a threshold, such as Relation R. There seems to be no reason to doubt that both duplicates would have the illusion of being the same person as the person at t1. (One could doubt that the illusion existed in both persons at t2, but this view would either be non- reductive or it would be strongly opposed to actual scientific findings, including those regarding 145 Further, it seems that this would be the case even if it the facts were not verifiable by others. 146 Strictly, there will be two different ways of defining "packets". On one way (developed by David Lewis), there can be two packets occupying the same body and having exactly the same mental states at the same time—because of some later event of fission. On the other way, fission results in what look like two people constituting the same packet (as I describe). 145 split-brain patients.) In light of these considerations, it is difficult to believe that this packet should be the unit relevant for ethical inquiry. To further illustrate the implausibility of the claim that the Relation R (or I) continuity packet is the unit relevant for ethical inquiry that is revealed by the duplication example, consider a more extreme version of the duplication example that is described by Parfit in a different context. 147 One can imagine a population of creatures (perhaps humans) that continually undergoes repeated cycles of fusion and fission, with the result being a complex tube-shaped lattice-like map of connectedness. To better grasp this possibility, see Parfit’s drawing in his text, 148 but the potential result is that the whole population—which could be enormous—is all part of one single Relation R (I) continuity packet. This possibility should evoke the same response as the original duplication case above—i.e. the Relation R (or I) continuity packet cannot be the unit relevant for ethical inquiry. In response to this point, someone who rejects AMU and my claim that unit relevant for ethical inquiry is the conscious moment (and thus who, perhaps, wants to defend ATLU or something akin to ATLU, and who wants to reject Premise 1A, and who wants to reject potential premises akin to 1A) might reply as follows: What really seems to matter for ethical inquiry might not be Relation R continuity packets, but rather the number of conscious experiences at a given time, as long as the appropriate connections in the lattice obtain. If we used this number as a denominator for our utility calculation of the lattice-type case, we could get a grouping of sorts and thus derive a result that would bear some resemblance to the assessment made by an ATLU- theorist. While this is an interesting response, it fails. The question is why the dimensions of time and space should be treated differently for the purposes of this calculation, and it seems that any attempt to explain this distinction would fail. If we are to consider spatially separate, but 147 Parfit, p.303. 148 Ibid, p.303. 146 concurrent, conscious experiences to be separate for the purpose of an ATLU-like calculation, it is unclear why we should not do the same for temporally separate conscious experiences. Another point might be made by the theorist who is struggling mightily to reject the claim that the empty question view supports the claim that the unit relevant for ethical inquiry is the conscious moment. Perhaps, he might say, longevity just is a good thing. Perhaps, it is better for R (or I) Relations to hold, and perhaps it is better for conscious moments to be linked by Relation R (or I). First of all, this point certainly seems to be true instrumentally. Though this would be an empirical question, it seems quite likely that conscious moments are able to be of much greater welfare in a world where there are R (or I) relations. This would be reflected by AMU’s assessments as well, though. As for the non-instrumental version of this point, though, it seems that it would be a consistent position to maintain that longevity is something that is of value in the world. Taken one way, however, it seems that it would be a non-welfare-based value, and thus, while it certainly could come into the mix in a person’s assessment of value, non-welfare-based goods have been bracketed in this dissertation—just as questions about, say, the value of equality and distributive justice have been bracketed and can be considered independently once the welfare-based assessment is complete. Taken another way, however, this point might be meant to be a claim sitting within the bubble of welfare-based goods. If this is the case, then the proponent of this position must precisify his position and explain explicitly how he responds to the various objections raised throughout section 1.5, and, more specifically, those raised here in section 1.5.F. Without such clarification and elaboration, it’s not clear exactly what that position is, or how that position takes itself to be responding to the objections raised here. Thus, it’s not possible, as is, to begin to assess its plausibility, and for the reasons given in this section, it seems doubtful that a 147 plausible account of this sort can be given. In sum, it does seem as though Relations R or I provide an apparatus enabling there to be a fact of the matter regarding the number of packets. At first, this seems promising. Upon closer inspection, however, it seems that it is precisely the same non-typical and non-everyday cases that led to the empty question view that now are the downfall of Relations R and I. It is in these cases that we do not feel that the packets described by Relations R or I should constitute the unit relevant for ethical inquiry. As such, it seems that we must accept Premise 1A, and, similarly, we must reject views that hold that Relations R or I are ethically relevant—in other words we must reject both the moderate view and the no difference view. This should be sufficient, but a few additional considerations are worth mentioning. If the relations between conscious moments are not metaphysically important, why would it be ethically important how Relations R and I do or do not connect different conscious moments? Even proponents of the moderate view and the no difference view—views that Schultz, Wolf, and others stipulate to for the purpose of their arguments—hold that reality could be described equally well without referring to the existence of persons or packets of R-related or I-related continuity. In light of this, it seems implausible that the person or the packet could be the unit relevant for ethical inquiry when persons and packets are not even necessary components of the description of a state of affairs. One could of course maintain that while we need not refer to persons (or packets) for the person (or packets) to be the unit relevant for ethical inquiry, it is not clear what—within the reductionist framework—could ground this claim. Here is a further point that’s important to remember: When a person at t1 has, for example, special concern for a person at t2 to whom he is R-related, this concern is a property of the person at t1, and this would be the case even if it turned out the person at t2 did not exist. (The converse would be the case for Relation I.) As is the case with other aspects of R- and I-relations, the relation in question is manifested and experienced at a 148 single time. Thus, it seems that while Relations R and I do suggest that persons are important, it suggests that they are important merely instrumentally. Though there isn’t direct evidence of the following, one possibility here is that the above authors, in arguing for the no difference view, (and even Parfit, in arguing for the moderate view) are wearing their empty-question-view hats, but when it comes to determining the implications of the empty question view, their non-empty- question-view intuitions might unwittingly be seeping back in. It’s important to elaborate on the point just made regarding the claim that, on the extreme view, Relations R and I do render personhood important instrumentally, even if not intrinsically. In no way does Parfit’s empty question account (if true), or similarly, AMU, ignore personhood or packethood altogether, and this is a point that will be returned to at various points in this dissertation. Rather, just as a utilitarian might follow rules of thumb and appeal to desert and fairness instrumentally, so too, a proponent of the empty question view can and should take personhood or packethood into account for instrumental purposes. As Parfit says, even if we are intellectually persuaded by an empty question view of personal identity, we will have a difficult time believing it at a lower, more mundane level—just as we might intellectually believe that we lack free will (in whichever sense we deem relevant), yet not be able to maintain this belief at a lower, more mundane level. This is a fact that any account must take as given. As such, in practice, various prescriptions that Parfit (or AMU) would endorse do not stray as far from common sense morality as Schultz, Wolf, and others might think. To the extent that this is the case, and if this is recognized and appreciated, perhaps the extreme view will come to be seen as less extreme. 1.5.G. Similarities and differences between my conclusions and those of Parfit The goal of section 1.5 was to provide additional reasons for accepting Premise 1 in my 149 four-step argument for AMU. To this end, I first argued that the reductionist view of personal identity and, further, the empty question view of personal identity, are plausible. As I argued, if they are true, this provides strong support for Premise 1. I next argued that if my opponent were to accept these points, he might raise a new objection to my four-step argument for AMU, pointing out that I also must show that Premise 1A is true if AMU is to follow. I then considered Premise 1A, and I provided reasons for accepting it as well. In short, Premise 1 and Premise 1A are both true because neither lives of persons nor packets of Relation R (or Relation I) continuity is the unit relevant for ethical inquiry. The conscious moment is the unit relevant for ethical inquiry. It’s worth briefly mentioning some of the similarities and differences between the conclusions Parfit draws and those that I draw when it comes to the implications that the metaphysics of persons has for ethics. One difference, as discussed earlier, is that Parfit does not reject the moderate view in the context of ethics, and he remains agnostic between it and the extreme view. I, however, endorse the extreme view. (In the following comments, I will refer solely to Parfit’s claims regarding the extreme view, but he holds that the moderate view provides support for the same claims—though merely somewhat less support.) Parfit and I share the view that the empty question view supports the claim that the conscious moment is the unit relevant for ethical inquiry, and thus he and I would agree that Premise 1 is true and he (on his extreme view) and I would agree that Premise 1A is true. As for the broader implications of this agreed upon intermediate conclusion, Parfit and I focus on different implications, and our implications are partially in tension. Starting from our intermediate conclusion, Parfit focuses his efforts on arguing that, if true, it lends support to utilitarianism (the broad theory) because it defuses Rawls’s egalitarian objection to utilitarianism. According to this objection, utilitarianism makes a mistake in not giving equality or other distributional concerns intrinsic value and by simply maximizing benefits 150 across people, and this mistake is due to a failure to take seriously the separateness of persons. 149 Armed with our intermediate conclusion, as Parfit says, there are at least two ways in which one can consistently proceed. One can either apply egalitarian distributional principles to all conscious moments or one can not apply these principles and instead apply utilitarian maximizing principles. Parfit considers both possibilities in depth, but he concludes that most people (including Rawls) who favor the use of principles of equality would think that equality loses its importance when it’s applied to the units of conscious moments. Thus, Parfit concludes that the empty question view of persons provides support for utilitarianism, because it completely defuses Rawls’s objection. 150 I agree with Parfit on this score. Interestingly, however, Parfit assumes, without any discussion at all (thus, eerily similarly to Huemer), that the version of utilitarianism the empty question view supports is TU. He assumes that maximizing benefits summatively is what is best in the intrapersonal context (in fact, a Rawlsian would sign onto this as well), and thus, in light of his conclusion that conscious moments—and not the individual’s life—are the relevant unit, he assumes that the empty question view supports a summative maximizing of benefits in society—i.e. TU. Once we conclude that reductionism provides support for the claim that the conscious moment is the unit relevant for ethical inquiry, I think that we reach another question, which is akin to the question Parfit confronts of whether distributive principles should be applied to the unit of the conscious moment or not at all. We are confronted with the question of how to aggregate the goodness of conscious moments. Importantly, I don’t think that the empty question view, if 149 Rawls, p.24. 150 This is the case, he says, if one subscribes to the extreme view in the context of ethics. If, however, the moderate view is true, Rawls’s objection isn’t fully defused, because Relation R is still ethically relevant. But the moderate view still does help to defuse Rawls’s objection, even if it does so less so than the extreme view. As for how much even a full defusing of Rawls’s objection would support utilitarianism, Parfit writes: “The argument that I have given cannot show that we must accept the utilitarian view. It can only yield conclusions about relative plausibility. When we cease to believe that persons are separately existing entities, the utilitarian view becomes more plausible. Is the gain plausibility great or small, my argument leaves this question open.” Parfit, p.342. 151 true, speaks to this question at all. It only speaks to the question of the relevant unit to which principles of aggregation should be applied, and thus, at this point, it remains an open question which principle of aggregation is appropriate. I, of course, will be arguing that the appropriate principle of aggregation is AU—that the anti-more-is-better intuition should apply to these units (conscious moments). In this respect, Parfit and I are in disagreement. Nevertheless, at this point, this is immaterial. What is relevant here, in this chapter, is that if the empty question view of personal identity is correct, this appears to provide support for the fact that the unit relevant for ethical inquiry is the conscious moment— and not the lives of persons and not packets of R-related (or I-related) continuity. Thus, I hope to have provided reasons to accept both Premise 1 and Premise 1A, thus providing (additional) strong support for AMU and TU, while providing strong (additional) reasons to reject ATLU. 1.6. Summary Part II Chapter 1 has attempted to provide a positive argument for why AMU is the most plausible version of AU. After the introduction in section 1.1 and the explanation of my assumptions in section 1.2, section 1.3 offered a four-step argument for AMU that the chapter attempted to support. Section 1.3 explained the plausibility of Premise 2 and why the conclusion follows from the premises. Section 1.4 then provided reasons to accept Premise 1, and section 1.5 offered additional reasons—based on the metaphysics of persons—to accept Premise 1. Section 1.5 also considered an additional response that an opponent might give to my argument, and it explained why we should accept Premise 1A—and thus why we should reject my opponent’s response. Much remains to be done, however. Various highly important objections to AMU were not addressed in this chapter, but, with the positive argument now on the table, these objections will 152 be addressed in the following chapters. Perhaps the most serious of these objections is that AMU must be false since it cannot accommodate IMI. This objection will be the main topic of Part II Chapter 2. However, if one rejects AMU, one will need to reject a component of the argument offered in this chapter. Thus, assuming that its conclusion follows from its premises, if one rejects AMU, one might consider the argument to be either a reductio against Premise 1 or a reductio against Premise 2 (AU on the whole). As such, the stronger the argument in this chapter, the more difficult it will be to reject its conclusion. This chapter—after offering the argument—attempted to strengthen the argument itself. The following chapters will strive to reject other reasons for casting doubt on its conclusion. ===================================================================== Part II Chapter 2: Considering counterarguments: AMU and the Intrapersonal Maximizing Intuition (IMI) Part II Chapter 1 argued that the different location for additional conscious moments is not morally relevant, and thus that the conscious moment itself is the unit relevant for ethical inquiry—if true, thus meaning that ATLU is not correct. Thus, after Part II Chapter 1, it appears that AMBP (the population-wide anti-more-is-better intuition) and IMI (the intrapersonal maximizing intuition, which states that, all else equal, it is better agent-neutrally that a happy person live longer) are incompatible. We cannot simultaneously hold both of them. It thus seems that we must adopt the same method of aggregation both intrapersonally and interpersonally—be the method that of AU or TU (or of some unidentified other option). 151 It is the need to make this decision that leads to the fact—alluded to in the text of Part II Chapter 1—that the four-step 151 Throughout the dissertation, I will have referred to the TU and AU methods of aggregation for the goodness of an outcome or state of affairs. Here and in what follows, however, I introduce the notions of “AU-type aggregation” and “TU-type aggregation” for assessing the goodness of a lifetime. AU-type aggregation in the intrapersonal context assesses the goodness of a life by averaging the goodness per moment of all the moments of the lifetime. TU-type aggregation in the intrapersonal context assesses the goodness of a life by additively summing the goodness per moment of all the moments of the lifetime. 153 argument of Part II Chapter 1 can be seen as a reductio against Premise 2 (i.e. AMBP, and, more generally AU), if one accepts Premise 1 and one accepts that the conclusion follows from the premises. I, however, do not think that AMBP (and, more generally, AU) is false, and I think that the argument is a good one. Thus, I think it is IMI that must be rejected. IMI, however, seems to be prima facie plausible to many, and this might be a tough pill to swallow. This notwithstanding, I think it must be done. The goal of this chapter is to explain why, contra prima facie intuitions, it is plausible to reject IMI. One important note: As mentioned earlier, in what follows I will not be assuming the acceptance of or the denial of the empty question view or of the reductionist view of persons. I will remain agnostic about whether or not each of them is true. Section 2.1 explains the attraction of IMI and it offers a four-step argument that I think is a formalized version of why people find IMI plausible. I then explain why this argument fails: Section 2.2 argues that Premise 2 is false. Section 2.3 argues that Premise 3 is false. Section 2.4 then considers other routes for getting to the conclusion of the argument. Section 2.5 considers further implications. One prefatory note is in order here: As mentioned at various points in the dissertation, I will, throughout, be using the term “happiness,” but it is meant to be an expansive concept that catches just about all notions of well-being 152 within its net. If one is not a hedonist, or if one is not any form of a utilitarian, one should feel free to understand the term in a broader sense that includes all understandings of well-being, and which enables one to understand the discussions here through the lens of one’s own framework. 152 Except, of course, those notions of well-being that do not allow us to assign levels of well-being to moments. 154 Section 2.1: The Reason for the Prima Facie Appeal of IMI: “The Common Argument for IMI” 2.1.A. Introduction One implication of AMU is that it is not better, all else equal, if more conscious moments exist. This implication holds in each of two different contexts: First, it is not better, all else equal, if the potential additional conscious moments occur in new people, and second, it is not better, all else equal, if the potential additional conscious moments occur as additional moments experienced by a person who already exists. This latter claim conflicts with the intrapersonal maximizing intuition (IMI), which states that, all else equal, it is better agent-neutrally that a happy person live longer. Therefore, if IMI is true, then AMU is false. And thus, as an AMU theorist, I must cast doubt on the truth of IMI. Unfortunately, IMI is generally thought to be a plausible principle, so my work is cut out for me. To further describe IMI, and to make the intuition in favor of IMI more concrete and more tangible to the reader, I offer the following example. Consider two worlds, w1 and w2. W1 and w2 both contain many people and are identical in all respects except for one: In w2, a particular individual, Sam, lives a life that is ten years longer than the life he lives in w1. Sam’s life is a happy one (and he accrues a constant non-fluctuating stream of happiness throughout each version of his life), but his average per moment happiness is below the average per moment happiness of the world as a whole. According to AMU, w2, the world in which he lives a longer life, is worse than w1. According to common sense intuition, however, w2 is better than w1, and, at the very least, w2 certainly is not worse than w1. After all, people will think: Sam is better off in w2, and everybody else is equally well off in the two worlds, thus how could it possibly be the case that w2 is not better than w1 or that w2 is actually worse than w1? This common sense intuition that I have drawn out is the intuition in favor of IMI. It is this intuition that I will attempt to reject. In order to do so, I will carefully explore what leads people to 155 accept IMI. The greater clarity brought about by this exploration, I argue, will enable us to see why we should not accept IMI. 2.1.B. “The Common Argument for IMI” I believe that the prima facie appeal of IMI is due to reasoning made formal by the following argument, which I’ll call “The Common Argument for IMI”: Premise 1: Intrapersonal maximizing fact (IMF): Perhaps with some rare exceptions, it is a psychological fact about human beings that what we want at any particular time is to maximize the amount of happiness we will experience, going forward, as measured by TU- type aggregation. Premise 2: If IMF is true, then the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime. (Said differently, state of affairs A is better than state of affairs B for a person if the person experiences more happiness in A than in B, as measured by TU-type aggregation, over the course of his lifetime.) Premise 3: The Pareto principle: State of affairs x is better than state of affairs y, agent-neutrally, if there is at least one individual who is better off in x than in y, and there is no individual who is better off in y than in x. Conclusion: IMI is true. The general appeal of this line of reasoning was drawn out in the example above (regarding w1 and w2), but the question now is what we should think of this specific argument. I think that the argument has prima facie plausibility, and that’s because each component of the argument has at least prima facie plausibility: (i) The conclusion appears to follow from the premises, (ii) Premise 1 seems plausible, (iii) Premise 2 seems plausible, and (iv) Premise 3 seems plausible. In what follows, I argue that we should accept (i) and (ii), and, though I do not think we should accept (iii) or (iv), I will describe their prima facie plausibility. (i) Whether the conclusion follows from the premises Although it could be disputed that the conclusion follows from the premises, it seems fairly clear that the conclusion follows from the premises, and I think that we should accept that it does. We can describe the form of the argument as the following: (1) A, (2) If A then B, (3) If B then C, 156 (4) Therefore, C. In an argument of this form, the conclusion follows from the premises, and I think it’s fair to say that our argument takes this form. (ii) Whether Premise 1 is plausible One prefatory comment is in order here: In assessing the plausibility of Premise 1, in particular, the reader should keep in mind the comment I made in the introduction to Part II Chapter 2 about the term “happiness” and how the reader should construe it in as broad a sense as he likes so that the comments here apply to his theory of wellbeing. I think that Premise 1 is plausible, and that it is plausible regardless of what one’s views might be about persons or about ethics. With some rare exceptions, it seems to be a psychological fact about human beings that what we want at any particular time is to maximize the amount of happiness we will experience, going forward, as measured by TU-type aggregation. Generally speaking, and all else equal, we want two things when it comes to our future: First, we want the conscious moments that we will experience to be as happy as possible. Of two possible conscious moments of differing happiness levels, we want the one that we experience to be the one with the greater happiness level. Second, if given the choice between two futures, identical except that one includes additional conscious moments at the end that the other future does not, we want to experience additional conscious moments if they are moments of positive happiness, and we do not want to experience them if they are moments of negative happiness. 153 As such, what human beings want at any particular time is to maximize happiness that will be experienced in the future, and as measured by TU-type aggregation. As I mentioned, exceptions to this descriptive theory might exist, and there are various types of exceptions that one might cite. For one, one might think that there are concerns rooted in 153 Of course, this characterization seems to imply a non-arbitrary zero point for happiness. I will argue in Part IV for the claim that a non-arbitrary zero point for happiness does not exist. 157 things other than one’s own welfare—such as the welfare of others or non-experiential goods. The current chapter brackets these considerations. It might be argued, however, that exceptions exist even within the framework of a person’s wellbeing. To give three examples here: Some might think that there are exceptions related to the phenomenon of akrasia; some might think that happiness, if there is enough of it over time, can have diminishing value, and thus that at some point it might not be better to experience additional happy moments—or even that it might be better not to experience additional happy moments; some might think that there is some quantity or quality of negative happiness that might be so bad that even if happiness that is experienced afterward would yield, on aggregate a positive sum of happiness for a particular future, as measured by TU-type aggregation, it might be preferable to not add this mixed bag to one’s future. Other alleged exceptions exist as well, and the exceptions alluded to here and others will be addressed in greater depth in later chapters. Some will also be addressed in section 2.4. However, it is far from clear that these prima facie exceptions are in fact exceptions. Furthermore, even if these prima facie exceptions do turn out to be exceptions, I think that they are much more the exception than the rule, and Premise 1 allows that there might be exceptions. Premise 1 simply states that it is a psychological fact about humans that, leaving aside potential rare exceptions, what we want at any particular time is to maximize the amount of happiness we will experience, going forward, as measured by TU-type aggregation. Interestingly, I think that Premise 1 can be and will be accepted regardless of one’s ethical view. For example, I think it can be and will be accepted by proponents of TU, proponents of ATLU, and proponents of AMU. Accepting this does not create any tension for TU or ATLU, because both of these theories are consistent with IMI. Accepting this does appear to create tension for the AMU theorist, though, because AMU is inconsistent with IMI, and this is the first of three premises, from which, if they are all accepted, IMI follows. This is precisely why we are 158 discussing this topic. This notwithstanding, I think that even the AMU theorist will (and should) accept Premise 1. Also interestingly, I think that substance of Premise 1 can and will be accepted regardless of one’s view of persons. For example, to hark back to the empty question view of persons discussed in Part II Chapter 1, I think that not only will a proponent of the common sense view of persons accept Premise 1, but so too will a proponent of the empty question view—a theorist who believes that it is an empty question whether any two conscious experiences are part of (or experienced by) the same person. An empty question theorist might quibble with the usage of the word “person” in Premise 1, but he will agree with Premise 1’s substance, and would fully agree to Premise 1 if it were slightly rephrased. Such a theorist will concede that we (or, rather, moments of consciousness) have the intuition that we will persist into the future as persons in a non-empty-question sense. Further, just like non-empty-question theorists, empty question theorists see persons as aiming to persist and maximize the happiness, as measured by TU-type aggregation, in what they believe to be their future—be it by having each conscious moment packed with as much happiness as possible, or by adding additional conscious moments of happiness. As was the case for the non-empty-question theorist (though more vividly here for the empty question theorist), this latter option amounts to a person trying to include as many conscious moments of happiness within his net of personhood as possible—somewhat like a country trying to increase its population not only by increasing birth rates, but also by expanding its borders by swallowing the land and population of other neighboring areas. In sum, I think that we should accept Premise 1 of the argument for IMI. (iii) Whether Premise 2 is plausible Recall that Premise 2 states: “If IMF is true, then the metric for assessing the goodness of a 159 life is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.” In section 2.2, I will argue that we should doubt the truth of Premise 2, but, notwithstanding this claim I will make, I think that Premise 2 is prima facie plausible and that most people do in fact find it plausible. Furthermore, I think it plays a key role in explaining the common sense belief that IMI is true. Thus, why is it that one might find Premise 2 plausible? I think that the common belief that Premise 2 is plausible is actually a function of people thinking that two sub-premises (let’s call them 2A and 2B) are plausible and that Premise 2 follows from the conjunction of 2A and 2B. I’ll call this “The Three-Step Argument for Premise 2.” 2A states: “If what people want at any point in time is to maximize their future happiness as measured by TU-type aggregation, it is also the case that what people want at any point in time is to maximize the happiness that they accrue during their lifetime.” This seems prima facie plausible enough. After all, one cannot change the past, so it seems that maximizing one’s future happiness as measured by TU-type aggregation will always maximize one’s lifetime happiness as measured by TU-type aggregation. 2B states: “Assuming perfect information, and aside from rare exceptions such as akrasia, if a person wants something, then we can assume that the state of affairs where he gets such a thing is a better for him than that in which he does not.” This, too, seems prima facie plausible enough. If what a person wants is to maximize the happiness he accrues over his lifetime as measured by TU-type aggregation, absent certain exceptions, it seems that we can assume that what’s best for this person is the life that maximizes the happiness that it accrues, on the whole, as measured by TU-type aggregation. Further, it seems fairly clear that Premise 2 follows from the conjunction of 2A and 2B. In light of this, and in light of the prima facie plausibility of 2A and 2B, it seems that there is prima facie reason to accept Premise 2. I will ultimately cast doubt on Premise 2, but prima facie, it certainly seems that the above considerations are plausible, they appear to exhibit common sense, 160 and they would generally be signed onto by most people. (iv) Whether Premise 3 is plausible Recall that Premise 3 offers a statement of the Pareto principle: “State of affairs x is better than state of affairs y, agent-neutrally, if there is at least one individual who is better off in x than in y, and there is no individual who is better off in y than in x.” This is in fact a statement of the Pareto principle, and the Pareto principle is a widely accepted and seemingly non-controversial principle. Thus, this seems like it should be benign enough. Further, the substance of the principle is seemingly quite straightforward. After all, if a state of affairs consists only in the lives of persons, and if two possible states of affairs (x and y) are identical in all respects except for that fact that Sam’s possible life in y is better than in x (i.e. all people other than Sam have possible lives in x that are equally as good as their (respective) possible lives in y), how could it possibly be that state of affairs y is not better than state of affairs x? We seemingly appropriately employ this principle both explicitly and implicitly all the time, and there appears to be no reason to doubt its plausibility in this context. I will ultimately reject Premise 3, but prima facie, it certainly seems that it is plausible, it appears to exhibit common sense, and it would generally be signed onto by most people. 2.1.C. Outlook In sum, I think that the prima facie plausibility of IMI is due to reasoning made formal by the Common Argument for IMI. Further, I have indicated that it is at least prima facie plausible that each of the three premises is true and that the conclusion follows from the three premises— since we can describe the form of the argument as “(1) A, (2) If A then B, (3) If B then C, (4) Therefore, C.” In the foregoing, I also argued that not only is it prima facie plausible that the 161 conclusion follows from the premises, but that it in fact does. I also argued that not only is Premise 1 prima facie plausible, but that it in fact is true. This notwithstanding, I think that the argument fails, and this is because there are strong reasons to doubt both Premise 2 and Premise 3. Section 2.2 explains why we should doubt the truth of Premise 2 and section 2.3 explains why we should reject Premise 3. As such, I argue that the failure of the argument is over-determined, since the falsity of either Premise 2 or Premise 3, on its own, would be sufficient to make the argument fail. Importantly, however, my goal is not merely to reject this argument, but, rather, to provide reasons to reject IMI. As such, section 2.4 considers a surrogate for Premise 2. As stated in the above formal argument structure, the purpose of Premise 2 was to get us from A to B. Having argued in section 2.2 that Premise 2 (“If A then B”) is doubtful, section 2.4 considers other arguments for B itself—i.e. arguments for the claim that “the metric for assessing the goodness of a life is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime,” even if this claim does not follow from IMF. Section 2.2. Casting Doubt on Premise 2 of the Common Argument for IMI 2.2.A. Introduction The goal of this section is to cast doubt on the truth of Premise 2, which, you’ll recall, states “If IMF is true, then the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.” It’s important to remind the reader here that the goal of this section is simply to reject this conditional statement in Premise 2, and it is not to reject the conditional’s consequent. In other words, this section will not attempt to show that it is false that the metric for assessing the goodness of a life is the amount of happiness the person experiences, as measured by TU-type aggregation. Given that my broader goal is to cast doubt on IMI, it will in fact ultimately be 162 important for me to cast doubt on the claim that the goodness of a life is the amount of happiness the person experiences, as measured by TU-type aggregation (and I will do so both in sections 2.4 and in Part II Chapter 3). Nevertheless, the task in this section is more modest. I aim only to show that Premise 2 should be rejected, and thus, there is a fatal gap in The Common Argument for IMI. As stated in section 2.1, it seems that the prima facie plausibility of Premise 2 can be traced to the prima facie plausibility of 2A and 2B, and Premise 2 follows from the conjunction of these two statements. Recall that 2A states: “If what people want at any point in time is to maximize their future happiness as measured by TU-type aggregation, it is also the case that what people want at any point in time is to maximize the happiness that they accrue during their lifetime.” 2B states: “Assuming perfect information, and aside from rare exceptions such as akrasia, if a person wants something, then we can assume that the state of affairs where he gets such a thing is a better state of affairs for him than that in which he does not.” I will begin by contesting the truth of 2A. To do so, I will rely on intuitions elicited by an example that Parfit provides in Part II of Reasons and Persons. One further prefatory comment is in order—and one which is structurally similar to a prior prefatory comment. As I stated, while the goal of the chapter is to cast doubt on IMI, since The Common Argument for IMI is just one way of supporting IMI (albeit the way I think almost everyone would find most plausible), rejecting The Common Argument for IMI is not the last task. Section 2.4 considers other routes that might be taken to IMI. The rejection of Premise 2 by means of rejecting either 2A, 2B (or both) is of an analogous structure: In section 2.2, I cast doubt on Premise 2 by arguing that The Three-Step Argument for Premise 2 fails. I think that The Three- Step Argument for Premise 2 is the route of defending Premise 2 that almost everyone would find most plausible, so denying this argument would cast serious doubt on the truth of Premise 2. This notwithstanding, a rejection of Premise 2 would require additional points suggesting that no other 163 route to Premise 2 is plausible. This task will not occur here, but, rather, it will be deferred to section 2.4. The reason for this is that casting doubt on routes to Premise 2 other than The Three- Step Argument for Premise 2 is a very similar task to casting doubt on routes to IMI that don’t employ Premise 2. Since I maintain that IMF is true, satisfying my burden in both cases will require the same task: showing that the metric for assessing the goodness for a person is not necessarily the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime. Thus, this task will be saved for section 2.4. These loose ends notwithstanding, I contend that the vast majority of people who find IMI plausible do so because of The Common Argument for IMI and The Three-Step Argument for Premise 2, and thus those loose ends actually are merely loose ends. Section 2.2.B. Casting Doubt on Premise 2A 2.2.B.i. My argument against Premise 2A Consider an example that Parfit provides and terms “My Past or Future Operations.” 154 In relevant respects, the example goes as follows: Parfit asks us to consider an amnesiac individual who wakes up in a hospital and who is told by a nurse that she knows the plight of two patients, but that she does not know which of the two patients he is. Either he is the patient who just underwent excruciating pain for ten hours and will experience no more going forward, or he is the patient who underwent no pain but will soon undergo excruciating pain for three hours—with the pain’s per-moment intensity being equal in the two cases. Who should this patient prefer to be? The example itself couches the question in terms of what should the patient prefer, and not in terms of what would the patient prefer. Both of these are interesting questions, and also very interesting is the relationship between these two questions. For now, I will hold off on the should 154 This example and its variants will prove to be key examples throughout this dissertation. The intuitions they elicit will be drawn upon to support a wide range of arguments that I will make. 164 question and the question about the relationship between the should and the would questions, and I will return to these two topics in Chapter 5. At this point, I will focus solely on the would question: What would the patient prefer? Or, rather, and more precisely, I will focus on the related question: What would you prefer if you were in the patient’s shoes? This latter version of the would question is more tractable than the former version, and furthermore, the answer to it would likely be imported as the answer to the former version anyway. Thus, what would you prefer if you were in the patient’s shoes? Though people’s intuitions do differ both on this question and on the question of what an intuition in this case might show, Parfit takes it that most people, himself included, would prefer to be the individual who has already undergone the pain. I share Parfit’s intuition in this case. Assuming that this assessment is correct, what does this show? First of all, this provides additional evidence in support of the truth of Premise 1 in The Common Argument for IMI—i.e. IMF (“Perhaps with some rare exceptions, it is a psychological fact about human beings that what we want at any particular time is to maximize the amount of happiness we will experience, going forward, as measured by TU-type aggregation.”). I take it that the preference to have already undergone surgery is an instance of IMF. This, however, is not all that the example shows. It also shows that 2A is false—it is not the case that “If what people want at any point in time is to maximize their future happiness as measured by TU-type aggregation, it is also the case that what people want at any point in time is to maximize the happiness that they accrue during their lifetime.” In other words, the example provides a case where a potential desire to maximize one’s happiness temporally neutrally and a potential desire to maximize one’s future happiness would not be coextensive in their applications to facts about the world. Furthermore, the example shows that in cases where these two potential desires come apart, the desire that people actually have is the desire to maximize one’s future happiness and people would not have a desire to maximize 165 their happiness temporally neutrally. In light of this, this suggests that we never have a (intrinsic) desire to bring about the best possible life for oneself—maximizing one’s happiness temporally neutrally. Even in the cases where our potential desire to maximize happiness temporally neutrally is coextensive with the potential desire to maximize future happiness, it is the desire to maximize future happiness that is operative. The frequent coextensiveness of the two potential desires is what makes it appear as though we have a desire to maximize happiness temporally neutrally. Thus, even if 2B is true, we will not get the conclusion of The Three-Step Argument for Premise 2—namely, that Premise 2 is true. We do not have a desire about our lives on the whole, and thus we don’t have the grounds for determining a metric for a goodness assessment of a life. It could be that the goodness of a life is measured by maximizing happiness temporally neutrally as measured by TU-type aggregation, but it could also be that the goodness of a life is measured by maximizing happiness temporally neutrally as measured by AU-type aggregation, or it could even be that some other function is the appropriate measure. Thus, we do not have the grounds to assert “If IMF is true, then the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.” 2.2.B.ii. Considering Objections to My argument against Premise 2A At this point, various objections can be raised in response to my claims about the implications of the Parfit’s example. I will raise and respond to three. Objection #1: According to Objection #1, we are misguided if we think an implication of the hospital example is that one does not have a desire to maximize happiness for one’s life. The conclusion that we lack this desire was based on the assumption that potential desires to maximize happiness 166 for one’s life and happiness for one’s future can come apart, and that in such cases we have a desire to maximize one’s happiness for one’s future. However, we have no evidence that these potential desires can in fact come apart. In any actual case, the two potential desires remain coextensive. In any actual case of choice, they are coextensive and cannot be separated. Thus, the intuitions in Parfit’s case do not generalize to cases of choice. Further, even if hospital example cases do not generalize, we are mistaken in our interpretation even of these circumscribed examples. Although we refer to the patient as expressing a desire or a preference, instead it really is something more like a hope, or some other mental state. Thus, while we might have something like a hope that our pain is in the past, we still have the desire to maximize happiness in one’s life. Response to Objection #1: First of all, I agree that it’s possible that the hospital example cases might not generalize to cases in the decision making context (absent the decision maker being a proponent of evidential decision theory, a possibility that I will not pursue here). This, however, does not cast doubt on the importance of the example and the implications we can draw from it. It takes precisely this kind of an example to separate these two otherwise coextensive potential desires, and it thus takes this type of an example to pull them apart and explore which desire is actually the desire that we have. This feature of the example is what enables it to do its work, and there is nothing suspect about this. In this sense, it is just like all good hypotheticals in philosophy that separate coextensive cases and make us think about which of two or more factors is actually the relevant one. Examples of this sort abound, but, to name one group of such examples, the various hypotheticals in the personal identity literature that separate the otherwise coextensive features of bodily and psychological continuity fit this mold. As for the second part of the objection that suggests that IMF is actually more of something like a hope: The concerns about the example do have some merit, and perhaps a more 167 clear and defined description of the nature of these preferences or hopes is important, but the basic thrust of the mental state is clear—a person is indifferent to the existence of pain in his past, and is deeply concerned about the existence of pain in his future. It seems that someone attempting to discount the intuitions in the hospital case must provide an explanation of our intuitions—and not merely gesture at some oddities of the mental states in question—if he is to persuade us to reject our intuitions in this type of a case. Furthermore, it seems that calling the IMF phenomenon a hope, a desire, a preference, an impractical preference, or whatever else will not change anything. Even if we call it a hope, the question then is just why our hopes are tied to our future happiness and not to the happiness of our whole lives, temporally neutrally. Further, the objection also posits the claim that even in the hospital case, we have a perhaps-hidden desire to maximize happiness temporally neutrally. I see no evidence for this. Why suggest that the absence of a feeling of a desire to maximize happiness temporally neutrally is better evidence of such a desire existing dormant than of the claim that no such desire exists? Lastly, at the absolute best, the objector gets the result that the two potential desires are coextensive. If this is the case, we still do not have reason to believe that the desire that we actually have is the temporally neutral desire, as opposed to the forward-looking one. At best, we would be in equipoise. Objection #2: This objection is somewhat related to Objection #1. As was the case with that objection, the claim here is that we do have a desire to maximize happiness temporally neutrally and that we do not have a desire to maximize merely future happiness. The rationale is somewhat different, though. According to the current objection, what we perceive as IMF is actually a heuristic. It is a heuristic aimed at maximizing future happiness in order to maximize happiness temporally neutrally for the person. Furthermore, as a practical matter, we cannot affect the past, so this is a 168 very useful heuristic—and in fact a perfect one. After all, heuristics are meant to affect actions, and while choices are actions, hopes are not. Thus, not only is IMF a heuristic designed for the purpose of maximizing happiness for a person temporally neutrally, but its prescriptions are also coextensive with the prescriptions that would be produced if one sought to maximize one’s accrued happiness temporally neutrally. In conclusion, according to this objection, what we actually desire at any particular time is to maximize happiness temporally neutrally. The perception of IMF that the hospital patient exhibits is merely the dread of impending pain. Having a heuristic of this sort focused exclusively on the future will be much more efficient than a heuristic that wastes any of its resources trying to avoid pain anywhere other than the future. Response to Objection #2: This point is not an unreasonable one, and it is one that is difficult to falsify. However, the following considerations explain why it seems contrived and implausible to maintain that IMF is a mere heuristic aimed at maximizing accrued happiness for a person temporally neutrally. In the hospital example, we’re provided with a case where we can think clearly about what we prefer, and even if we actively keep in mind the heuristic theory just offered, it’s very clear to us what we prefer and what is most important to us—happiness going forward. Generally, even for the most useful heuristics, if we are presented with the heuristic theory and presented with a case which clearly is a case where a heuristic does not prescribe the result that it is designed to promote, we will see this and express a preference that is not in line with the heuristic, but in line with what the heuristic is designed to promote—what we truly care about. We can conceive of a heuristic that is so ingrained that we are unable to avoid viewing its prescriptions as valuable in and of themselves, but in a case of this sort, it seems that the proper thing to say would be that what once had merely instrumental value now in fact has intrinsic 169 value. Following what before was a heuristic is now no longer a case of following a heuristic at all. I have doubts about whether this historical heuristic theory is true for two reasons: First, I think that if it were correct, when confronted with the hospital case, we would likely feel not only a desire from what was once a mere heuristic, but we would also feel a desire in the direction of what was valuable in and of itself. I don’t think we feel the pull in both directions. Secondly, for independent reasons, I doubt whether we would have ever had a desire to maximize one’s life’s happiness temporally neutrally. Regardless of these doubts, however, even if we are to maintain the heuristic theory, it seems that IMF has clearly reached the threshold where it constitutes a desire in and of itself. Further, there is no evidence in favor of there being a desire to maximize happiness temporally neutrally. The hospital example provides all the evidence we need in the opposite direction. Thus, it does not seem right to suggest that what we really do seek and want is to maximize the total amount of accrued happiness in our lifetimes. Objection #3 Somewhat related to the other objections, but still importantly different is the following objection: One might object that, while we do have the intuition that we would prefer to have already undergone the surgery, the best explanation of this is that the fact that the case would never arise in the context of practical choice makes our intuitions confused and trustworthy—the example is too far removed from reality. Response to Objection #3 While perhaps it is true that the hospital example can’t be generalized to cases of practical choice, this does not mean that it is a situation at all removed from reality. Generally the cases that philosophers say are too divorced from reality for our intuitions to be reliable are cases of science 170 fiction, some fancy technology we have not experienced, or some seemingly non-existent entity (such as the famous “utility monster”). Here in the hospital example, however, there seems to be nothing difficult to imagine about the case. It seems like a fairly commonsensical situation that the patient is in, and it wouldn’t be overly surprising if there have in fact been a handful of patients in history who have been confronted with the exact situation described in Parfit’s example. In sum, it seems that our intuitions regarding the hospital example should not be discounted as unreliable. Section 2.2.B.3. Outlook Thus, it seems that Premise 2A is false. I return below, in Section 2.4, to address whether we do indeed need Premise 2A after all. I conclude there that substitutes for Premise 2A fail as well. Section 2.2.C. Casting Doubt on Premise 2B Leaving aside the plausibility of 2A for the time being, should we accept 2B? Recall that 2B states “Assuming perfect information, and aside from rare exceptions such as akrasia, if a person wants something, then we can assume that the state of affairs where he gets such a thing is a better one for him than that in which he does not.” Thinking about 2B, it does appear that there is a gap between the principle’s antecedent and its consequent. After all, if there were no such gap, there would be no principle needed to close a gap. While on first thought it might seem that this gap-closing principle is not justified, it’s not clear what else could get us to a claim about goodness other than a claim of some sort involving a desire. Thus, perhaps if we are to make claims about the goodness of something for a person, this must be rooted in desire at least somewhat. I think that this reasoning in support of 2B would be plausible in most respects, but not in all. Importantly, however, I think that the context of this section’s discussion is a context in 171 which it should be rejected. It is generally (and, if assuming perfect information, always) good for a desire to be satisfied, because satisfying a desire at t1 will bring about more happiness at t1 than the frustration of the desire at t1. There are two reasons relevant to the current discussion for why this principle fails in this context, and thus the falsity of 2B is over-determined. First, one feature of our current context is that comparisons are being made between possible remainders of one’s life (or total lifetimes) where, for example, one possible remainder involves x number of years at constant happiness level y, and where a second possible remainder involves x + 1 number of years at constant happiness level y. However, when it comes to the time at which we know whether the desire has been satisfied, there is no possible life that is better than another possible life—because at this point there only is one possible life. The other one has ended. Thus, as stated, the benefit of desires being satisfied occurs at the time they are satisfied (if at all), but in these cases there is no person who is better off in the potential life where the desire is satisfied than in the one in which it is not. Second, and more comprehensively, the context in which we’re talking is desires for happiness. This type of a desire is a higher-order desire and is thus different from most first-order desires. First-order desires, when satisfied by some object or state of affair’s presence, bring about happiness. The desire for happiness, when satisfied, however, does not bring about (additional) happiness. Thus, the standard reason for why the satisfaction of a desire is a good for a person does not apply here. In other words, what is good for a person is happiness itself. But once we are out of the domain of desires and back in to the domain of happiness, the question is simply what the appropriate measure is for aggregating the happiness—be it TU-type aggregation, AU-type aggregation, or some other method. While one might have thought that a stated preference for a remainder of one’s life that maximizes the remainder of the life’s happiness according to one of 172 these measures would indicate that that possible remainder would in fact be the better remainder for the person, for the reasons I mention, we are not entitled to come to this conclusion. It seems that the question of what the appropriate measure is for the goodness assessment of a life or a remainder of one’s life is a question that is distinct from the question of which possible life or remainder of life one desires, and this is the case even if the person with the desire neither is akratic nor lacks perfect information. Not only are the two questions distinct, but it’s not clear that there is any relationship between them whatsoever. Thus, in sum, I argue that 2B is false because it need not be the case (even leaving aside cases such as akrasia and imperfect information) that just because a person has a desire for something, it follows that the state of affairs where he gets such a thing is a better one for him than that in which he does not. In some contexts this will be true. In our context however, it need not be. It could be the case that the connection holds true in our context and it could be the case that it fails to hold true in our context. At this point, we do not have the grounds to make either claim. While it would not be inconsistent to say Person X has a desire to live an additional happy year and it is good for him if he does, it also would also not be inconsistent to say Person X has a desire to live an additional happy year and it is not good for him if he does, or further, it also would not be inconsistent to say Person X has a desire to live an additional happy year and it would be bad for him if he does. Thus, just as I will be arguing below that Premise 3 fails because there is a gap in The Common Argument for IMI’s reasoning from claims about agent-relative goodness to agent- neutral goodness, the falsity of 2B shows that Premise 2 fails because there is a gap between claims about desires—which are mere psychological facts about persons at particular times—and goodness assessments (even if the goodness assessments are merely agent-relative). These desires, such as IMF, are mere psychological facts about persons, and they are consistent with various 173 types of claims about goodness assessments—and thus information about these psychological facts is orthogonal to coming up with a goodness assessment of a person’s life or a remainder of a person’s life. Of course this gap—said otherwise, the falsity of 2B—was not the only reason for rejecting Premise 2. The first half of this section argued that we should also reject 2A, and either the rejection of 2A or the rejection of 2B would be sufficient to reject The Three-Step Argument for Premise 2. As discussed, however, though I think that The Three-Step Argument for Premise 2 is far and away the most common route to Premise 2, there might be other ways of arguing for Premise 2’s plausibility. I will consider this possibility in section 2.4. Section 2.3. Casting Doubt on Premise 3 2.3.A. Recall that Premise 3 of The Common Argument for IMI offers a statement of the Pareto principle: “State of affairs x is better than state of affairs y, agent-neutrally, if there is at least one individual who is better off in x than in y, and there is no individual who is better off in y than in x.” As I stated in section 2.1, the Pareto principle is a widely accepted and seemingly non- controversial principle, and thus, prima facie, Premise 3 seems plausible. After all, if a state of affairs consists only in the lives of persons, and if two possible states of affairs (x and y) are identical in all respects except for that Sam’s possible life in y is better than in x (i.e. all people other than Sam have possible lives in x that are equally as good as their (respective) possible lives in y), how could it possibly be that state of affairs y is not better than state of affairs x? Despite the apparent plausibility of Premise 3, this section will argue that it is false. If Premise 3 is false, then The Common Argument for IMI will fail—and it will fail even if Premise 2 were true. In light of this, and in order to focus the analysis on Premise 3, this section will 174 assume (counterfactually) that Premise 2 is correct and (non-counterfactually) that Premise 1 is correct. Thus for the purposes of this section, we will assume that the metric for the goodness for a person is the amount of happiness that the person accrues, as measured by TU-type aggregation, over the course of his lifetime. I will go through various examples below, but here, in a nutshell, is my take on Premise 3. The problem with Premise 3 is not the fact that it is a statement of the Pareto principle, but, rather, the problem is that it actually is a statement of one particular instantiation of the Pareto principle that imports a view about the unit relevant for ethical inquiry that I find implausible. Recall that, as an AMU theorist, I maintain that the unit relevant for ethical inquiry is the conscious moment (as opposed to the person), and I argued for this at length in Part II Chapter 1. The instantiation of the Pareto principle in Premise 3, however, is one that takes the unit relevant for ethical inquiry to be the person. While this person-based instantiation in Premise 3 is certainly the most common instantiation of the Pareto principle—and, in fact, it is mistakenly assumed to be the only version of the Pareto principle—there are other possible instantiations, and they can be created using a template that remains agnostic about what constitutes the unit relevant for ethical inquiry. This generic version of the Pareto principle states: “State of affairs x is better than state of affairs y if there is at least one [insert relevant unit] that is better in x than in y, and there is no [insert (same) relevant unit] that is better in y than in x.” Thus, an AMU theorist would wholeheartedly accept this generic version of the Pareto principle. Furthermore, filling in the brackets, the AMU theorist would get the following as the relevant Pareto principle: “State of affairs x is better than state of affairs y if there is at least one conscious moment that is better in x than in y, and there is no conscious moment that is better in y than in x.” Although an AMU theorist would reject the instantiation of the Pareto principle in Premise 3, he would happily sign onto and defend the version of the Pareto principle I just stated. Further, this is the case regardless of whether the 175 AMU theorist in question is a proponent of the AU-type or TU-type view about the value of lives—both of which are consistent with the AMU theorist’s account. Having made these clarifications about what aspect of Premise 3 is problematic, I will now show exactly how Premise 3 fails. As I mentioned, for our present purposes, I am counterfactually granting the rest of the IMI theorist’s argument in The Common Argument for IMI, and this includes the claim that the metric for the goodness of a life is the amount of happiness it accrues, as measured by TU-type aggregation, over the course of its lifetime. In what follows, however, I will not restrict my analysis of the Pareto principle to cases where we adopt the TU-type measure. The analysis applies just as well to the AU-type measure. In light of the fact that we have yet to determine which (if either, or both) is the correct measure, and in light of the fact that both views are consistent with AMU, I will keep both on the table. Further, even though the IMI theorist only needs Premise 3 to be true for cases of TU-type aggregation, Premise 3 (as is it as been written) is not restricted in this way, so it can’t hurt to probe both. Though the main purpose of this section is to undermine Premise 3 in the argument for IMI, it’s important to note that the fact that Premise 3 is false shows that not only is the TU-type view consistent with AMU, but so too is an AU-type view. If not for the falsity of Premise 3, however, both of these views would be inconsistent with AMU. 2.3.B. First let’s consider Premise 3 as it is applied to the AMU theorist who espouses an AU- type view about the value of lives. It could be the case that everyone’s (other than Sam’s) lifetime average-per-moment happiness values are the same in w1 and w2 but that Sam’s lifetime average- per-moment happiness is greater in w2 than in w1 due to ten additional great years that have a higher average per moment happiness than his prior years, and which thus raise the average-per- 176 moment happiness of his life. It still could be, though, if Sam’s additional years are worse by average-per-moment-happiness standards than the average-per-moment happiness of the world, that he would have a better life, everyone else would have an equally good life, and the state of affairs, as a whole, could be worse. Further, the state of affairs, as a whole, could be worse in w2 than in w1, even if Sam has a better life in w2 and if everyone else in the population also had slightly better lives according to average-per-moment happiness, in w2. This would be the case if the improvement of the state of affairs as a whole, caused by the improvements to the lives of others did not outweigh the detriment to the state of affairs as a whole, caused by the improvement of Sam’s life. This result is an instance of the Batting Average Paradox, according to which player x might have a higher batting average (i.e. percentage of attempts in which one successfully gets a hit) than player y in both years 1 and 2, yet have a lower overall average for those two years than player y. This can occur if the number of at-bats (i.e. attempts) are not equal for the two players or for the two years. One player might have his average in year where he performs better count more heavily than his average in the year in which he performs worse toward his overall two-year batting average, whereas the other player might have the opposite be the case, due to the number of at bats each has in each year respectively. Similarly, in the most recently described case of w1 and w2: Those other than Sam have better lives in w2 than in w1, Sam himself has a better life in w2 than in w1, but the state of affairs on the whole (which is comprised solely of Sam’s life and the lives of those other than Sam) is better in w1 than w2. This is because the w1 and w2 diverge in the percentage of the respective world’s total conscious moments that are Sam’s conscious moments (and similarly the two worlds diverge in the percentage of the total conscious moments that are conscious moments of those other than Sam). As such, the world in which Sam’s moments constitute a greater percentage of the society’s moments will have Sam’s average-per-moment 177 happiness weigh more heavily into the society’s overall average. Similarly, as mentioned already, Premise 3 can straightforwardly be applied to the AMU theorist who espouses a TU-type view about the goodness of a life. Consider the case of Sam, just given, where Sam’s life has less average-per-moment happiness than the average-per-moment happiness for the world on the whole. On the TU-type view, if Sam’s life is extended by ten years and these ten years have not even a higher average-per-moment happiness, but merely the same average-per-moment happiness as his prior years, this will increase the value of his life. This, however, will lower the world’s overall average-per-moment happiness, even if the lives of everyone other than Sam are unaffected. Similarly, it could be that the lives of everyone other than Sam are made slightly better (either by extending their happy lives or by improving the average- per-moment happiness of the lives of their original length), and Sam’s life is extended, but that the detriment to the world as a whole brought about by the improvement of Sam’s life outweighs the improvement to the world brought about by the improvement of the lives of those other than Sam, thus meaning that everyone’s life is better in w2, but that, on the whole, according to AMU, w2 is worse than w1. 2.3.C. Do these examples show that Premise 3 is false? Yes, as long as the instantiation of the Pareto principle is person-based, then the AMU theorist should reject the principle. The broader point here is that for an instantiation of the Pareto principle to be plausible, it must employ the unit relevant for ethical inquiry that one employs in one’s ethical theory. Only when the same unit relevant for ethical inquiry is used in one’s ethical theory and in the Pareto principle will one get plausible results. Thus, if one had a coherent account of why the person were the unit relevant for ethical inquiry, then Premise 3 would be plausible. However, as explained in Part II Chapter 1, the 178 person is not a plausible unit relevant for ethical inquiry. Instead, the unit relevant for ethical inquiry is the conscious moment. Though this disconnect is the broader explanation, here is a slightly more specific account. Premise 3 fails because it requires a companion ethical framework where the unit relevant for ethical inquiry is the person. In other words, it requires an ethical theory where each person counts equally, no matter what the length of his life is. This Pareto principle would be consistent with aggregating over multiple people by averaging or also by adding, but only as long as all persons are counted equally. The reason that this Pareto principle doesn’t work in the AU-type and TU- type Sam cases, above, is that w1 includes a different weighting of persons’ values than does w2. This tinkering of weights enables a version of the Batting Average Paradox to take hold in both the AU-type and TU-type cases. Importantly, the reason that the different weightings of persons occurs is that in the AMU framework, the unit relevant for ethical inquiry is not the person, but the conscious moment, and thus persons only get equal weights in factoring into the state-of-affairs- wide average if they contain the same number of conscious moments. Thus, other candidates for Theory X that take the conscious moment as the unit relevant for ethical inquiry could similarly potentially come into conflict with the version of the Pareto principle in Premise 3 and thus be forced to reject it. Despite this explanation, one might still have the thought that it is not merely the generic Pareto principle template that is plausible, and not merely the “conscious moment” version of the principle that is plausible, but it is also (or instead) the original person-based formulation of the Pareto principle (stated in Premise 3) that is plausible and that we should accommodate. If this is one’s view, what this means is that one thinks that the unit relevant for ethical inquiry is the person. Thus, the debate isn’t really about the Pareto principle at its core—one’s choice of Pareto principle is an implication of one’s choice of the unit relevant for ethical inquiry. As for the 179 question of what is the unit relevant for ethical inquiry, I can merely point you to Part II Chapter 1 and let that make my case. Despite the fact that that is where the real action is, there are two further points an AMU theorist can make here in the attempt to provide additional support for his position about the Pareto principle in particular: First: Recall that in Part I Chapter 2, person-affecting moralities were briefly considered as possible candidates for Theory X. These were views according to which a state of affairs could only be worse than another state of affairs if it were worse for a person. Despite the fact that they accorded with our intuitions in many cases, I argued that these views proved to be non-viable candidates because of their inability to satisfactorily solve the non-identity problem. Thus, despite the fact that many of our intuitions are person-affecting in form, and despite the fact that a viable Theory X will have to recognize (instrumentally) their importance to our thinking, person- affecting moralities were rejected as candidates for Theory X. Now, to return to our discussion of the Pareto principle: I think that people’s attraction to the person-based Pareto principle might, in part, be related to people’s attraction to person-affecting moralities in general. This, I think, is a point that is different from (and secondary to) the (primary) point in the previous paragraph about how people generally think that the unit relevant for ethical inquiry is the person. Though separate points, I think that they both work to strengthen people’s support for the person-based Pareto principle. In light of this hypothesis about person-affecting intuitions, the point here is as follows: In Part I Chapter 2 we realized the need to reject person-affecting moralities despite the fact that we have, in many contexts, strong inclinations to think in person-affecting terms. Here, in the context of the Pareto principle, a related issue arises, and I think that if we remember our willingness, in Part I Chapter 2, to relinquish person-affecting intuitions, we will see that a similar decision here is easier to make, and less unpalatable than it might otherwise have seemed. Second: The AMU theorist should point out that there might be a second only-somewhat- 180 related consideration that might be infiltrating one’s judgment in favor of the person-based Pareto principle in the TU-type aggregation context (and in cases of the AU-type context where additional moments are good for the life of the person in question, but bad agent-neutrally). This is as follows: How, one might say, can it be bad to add extra good (i.e. happy) conscious moments to a person’s life, and thus to the world? After all, they are good conscious moments, and nothing else is affected! Though related to the main consideration in favor of the person-based Pareto principle, this consideration is different, because it accepts the AMU theorist’s claim that the unit relevant for ethical inquiry is the conscious moment. It just asks how the mere addition of a single good moment could possibly be a bad thing. While this is certainly a fair (and common) objection, it must be realized that this is not an objection that is unique to AMU. Rather, as will be discussed in Part III Chapter 1, this objection takes the form of a more general objection to AU theories on the whole, including ATLU. The version of the objection aimed at ATLU asks: How can a world, w2, be worse than w1 if the only difference is that in w2 an additional happy person exists (albeit a person who is less happy than the average person)? This objection is addressed (successfully, I believe) in Part III Chapter 1, and I think that the treatment there will go a long way toward diffusing the objection here in the specific context of AMU. Further, to the extent that this consideration was in fact infiltrating one’s judgment in favor of the person-based Pareto principle, one’s support for the person-based Pareto principle should be proportionately weakened—both by the disentanglement of the two considerations, and, by the fact that, even if it were relevant, the second consideration can be met with a good response. 181 Section 2.4. Considering Additional Arguments that Could Be Made (and Additional Routes that Could Be taken) by the Proponent of IMI 2.4.A. Overview At this point, I have shown that the Common Argument for IMI fails. This is because Premise 2 is false and Premise 3 is false, the falsity of either of which would be sufficient to make the argument fail. Further, I have argued that Premise 2 is false because both Premise 2A is false and Premise 2B is false, the falsity of either of which would be sufficient to make the Three Step Argument for Premise 2 fail, thus casting strong doubt on the plausibility of Premise 2. Thus, there are various independently sufficient reasons for us to reject the Common Argument for IMI. As I’ve said, however, the goal in this chapter is not simply to cast doubt on the plausibility of the Common Argument for IMI or to cast doubt on the plausibility of the Three Step Argument for Premise 2. Rather, the goal is to cast doubt on the plausibility of IMI itself. After all, it is IMI that is inconsistent with AMU, and thus a proponent of AMU must cast doubt on the plausibility of IMI. While this is the overarching goal here, it nevertheless was of no small importance to cast doubt on the Common Argument for IMI, because I think that the prima facie plausibility of this argument is what explains and causes people’s intuitions in favor of IMI. Similarly, it was of no small importance to cast doubt on the Three Step Argument for Premise 2, because I think that the prima facie plausibility of this argument is what explains and causes people’s intuitions in favor of Premise 2. Thus, having cast doubt on these arguments, I think that I have already sufficiently cast doubt on the plausibility of IMI. Despite this, however, since the plausibility of IMI itself ultimately is what I seek to cast doubt on, I consider, in this Section, some additional points that could be made in favor of IMI. First, in Section 2.4.A, I consider the question of whether Premise 2A indeed is needed to make the Three Step Argument for Premise 2 work. I explore this route, but I conclude that Premise 2A is indeed needed to make the Three Step Argument for Premise 2 work. 182 Next, in section 2.4.B, I briefly address some additional loose ends—i.e., other possible points that could be made in favor of IMI. Specifically, I address two possible points, both of which are of a similar structure. First, I consider the possibility that we can get the output (i.e. conclusion) of the Three-Step Argument for Premise 2—namely Premise 2 (““If IMF is true, then the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.”)—by a route other than the Three-Step Argument for Premise 2. In other words, I consider whether we can foreclose the possibility that there is another way of establishing the plausibility of Premise 2. Second, and analogously, I then consider the possibility that we can get the output of Premise 2— namely, the consequent of its “if then statement”: “[T]he metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime”—by a route other than the plausibility of Premise 2 as a whole (i.e., without using the route of establishing the “if-then” statement). In other words, I consider whether we can foreclose the possibility that there is another way of establishing the plausibility of the claim that “the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.” If I’m able to cast doubt on the plausibility (and, rather, it would be a success to even just cast some doubt on there being obvious plausibility) of there being other routes to the position that TU-type aggregation is the correct aggregation mechanism for assessing goodness for a person, this would provide strong additional support for AMU, because it would remove just about any possible motivation for espousing IMI (i.e., that all else equal, it is better agent-neutrally that a happy person live longer). It is hard to imagine a person who espouses IMI but who does not espouse TU-type aggregation as the metric for assessing goodness for a person. Of course, even if one has reason to doubt the plausibility of TU-type aggregation as the metric for assessing 183 goodness for a person this does not, by definition, mean that one similarly has reason to doubt IMI, because one could have other reasons for espousing IMI, but the possibility of this is so remote that it does not even seem to be worth considering. Thus, to be clear, I do not need to disprove the possibility that TU-type aggregation is a plausible account of aggregation in the context of determining the goodness of a life. Rather, it will be a substantial accomplishment to show that AU-type aggregation and TU-type aggregation are equally plausible accounts in this context. After all, the prima facie plausibility of IMI is driven by the assumption that TU-type aggregation is far and away the most plausible account of aggregation in this context. Showing that TU-type aggregation is no more plausible in this context than is AU-type aggregation would go a long way toward removing the prima facie plausibility of IMI. I now proceed to consider the loose ends that I have just described. 2.4.B. An Alternate Argument that Could Be Given by a Proponent of IMI in Response to My Rejection of Premise 2A 2.4.B.1. In light of the apparent falsehood of 2A, the question might arise whether we do in fact need 2A in The Common Argument for IMI. As the argument is currently constructed, the answer is yes, but it is, however, worth considering a slight variant of The Common Argument for IMI that would still have IMI as its conclusion and which would not include 2A. Thus, consider the following argument: Premise 1: Intrapersonal maximizing fact (IMF): Perhaps with some rare exceptions, it is a psychological fact about human beings that what we want at any particular time is to maximize the amount of happiness we will experience, going forward, as measured by TU- type aggregation. Premise 2*: If IMF is true, then the metric for assessing goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, in his future. (Said differently, state of affairs A is better than state of affairs B for a person if the person 184 experiences more happiness in A than in B, as measured by TU-type aggregation, in his future.) 155 Premise 3: The Pareto principle: State of affairs x is better than state of affairs y, agent-neutrally, if there is at least one individual who is better off in x than in y, and there is no individual who is better off in y than in x. Conclusion: IMI is true. This argument is the same as the original argument except for Premise 2* is constructed in such a manner that it does not require 2A to be true. The question, though, is whether Premise 2* in this argument would be true if we stipulate that 2B is true. First of all, to be clear, according to Premise 2* in this new argument, the metric for assessing goodness for a person is one that only takes into account the remainder of one’s life. As the hospital example showed us, though, there can be a divergence between the state of affairs that constitutes the best remainder of one’s life and that which constitutes the best life as a whole, and this is the case regardless of which aggregation method one uses for one’s goodness assessment for a full life. Even if it employed TU-type aggregation (thus matching the measure for the goodness assessment of the remainder of one’s life), the results could diverge. IMF is completely indifferent to one’s past pleasures and pains—except insofar as one’s past will affect one’s current and future experiences through memory, dispositions, or other mechanisms. For IMF, a person’s whole set of experiences can matter instrumentally, but only one’s current and future experiences can matter intrinsically. As such, Premise 2* of this new argument’s goodness assessment shares this feature of IMF. The question, however, is whether Premise 2* is true, and there is an immediate problem with it. The problem is that it says that goodness for a person is a function of a person’s future happiness, but while there is only one person, each time-slice of the person has a different future, and thus there would be no single value associated with “the amount of happiness in a person’s future” in any particular state of affairs. There are two related difficulties that this entails. First of 155 I will show below, in a few pages, why this incorporation here of the notion of “future” is problematic. 185 all, the infinite number of different “futures” will lead to diverging values for goodness of a state of affairs, depending on the assessment point. In other words, not only will the value of every future likely diverge from the goodness of a life as a whole, but they will also all likely diverge from each other. Thus, the absolute value of a goodness assessment will be undetermined, and worse, undeterminable, and worse yet, inconsistent. Second, as the hospital example illustrates, allowing for different points of assessment can potentially lead to conflicts even about the relative value for a person of different states of affairs. In the hospital example, the conflict was called a conflict between a temporally-neutral and a forward-looking approach, but it can also be articulated as a conflict between forward-looking approaches at different times. For example, in the hospital example, the patient prefers state of affairs B to state of affairs A at the time the nurse conveys the information to him (t2), but if the patient were told about the two states of affairs a day earlier (t1)—before either patient had undergone surgery—then the patient would prefer state of affairs A to state of affairs B. 156 This in and of itself, however, poses no problem. While the instances of IMF at t1 and t2 conflict with each other, these are simply desires, and while two desires may conflict, it is not a logical impossibility for one to have a desire at t1 that conflicts with a desire one has at t2. So far so good. The problem, however, is what occurs when 2B comes into the picture. Recall that 2B states “Assuming perfect information, and aside from rare exceptions such as akrasia, if a person wants something, then we can assume that the state of affairs where he gets such a thing is a better one for him than that in which he does not.” If this is the case, then the two instances of IMF, which were merely conflicting, are now transformed into two logically contradictory goodness assessments: “State of affairs A is better for the patient than state of affairs B” and “State of affairs B is better for the patient than state of affairs A.” It cannot be the case, of course, that both of these 156 There has been some interesting discussion of this preference-reversal phenomenon lately—especially by Tom Dougherty. 186 statements are true. (In light of the fact that this contradiction occurs when 2B comes into the picture, one might think that this constitutes grounds to doubt 2B. For reasons discussed below, however, I don’t think that that is the right approach—as long as 2B is used in the appropriate context. I did, however, provide independent reasons to doubt 2B in Section 2.2.C.) Is there any way to resuscitate the new argument’s Premise 2*? One might attempt to dissolve this difficulty by importing into the goodness assessments the time-indexing that was attached to the desires themselves. Though prima facie appealing, this move will not work. A goodness assessment is not something that admits of time-indexing. If it is the case at t1 that a state of affairs is good for person x, it cannot be that, at t2, that very same state of affairs is bad for person x. The statement “State of affairs A is good for person x (as compared to state of affairs B)” is either a timeless truth or a timeless falsehood, but its truth value cannot change depending on when the statement is uttered. Perhaps there is a different account that can make sense of Premise 2* (“If IMF is true, then the metric for assessing goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, in his future”). Perhaps we could understand the claim not as being about goodness for a person per se, but rather, as shorthand for comparisons of goodness of possible remainders of a life. In other words, perhaps the statement that state of affairs A is better than state of affairs B for person x at t1 really is shorthand for saying that the remainder of x’s life in A is better than the remainder of x’s life in B. If so, we avoid the previous problems, because there is no need for time indexing a goodness claim. While there is a time component in the statement about the remainder of one’s life, this is to pick out what remainder is being referred to, and it is not making a time-indexed claim about goodness for a person. Notwithstanding the fact that this account avoids the difficulties that plagued the prior 187 explanations, the account is not satisfactory. The “goodness for a person” aspect of Premise 2* (as well as this aspect of the original Premise 2) is a piece that is needed if the conclusion of the argument is to follow from the premises. This is because a key input of Premise 3 (the Pareto principle) is a fact about goodness for a person. Thus, the account in this paragraph avoids difficulty, but only by steering clear of what is needed. I now give one final potential account that might make Premise 2* plausible. Suppose that there is a sense in which the appropriate understanding of the term “person” only includes the present and future time-slices of what we might otherwise call the “person.” Exactly what might be meant by this will be addressed further in Part II Chapter 3. If the notion of “person” were a forward-looking account, though, then the goodness assessment—the “goodness for a person”— need not be time-indexed, because the time index would come from the notion of personhood itself. This would avoid the time-indexing of a goodness assessment problem in the same way as the account in the above paragraph that spoke of the goodness for a remainder of a life, but it seemingly would avoid the pitfall of that account—i.e. not referring to goodness for a person. Whether this forward-looking account of personhood is coherent and plausible will be considered in Part II Chapter 3 (after it is described in greater detail), but there are doubts that one could have. Even leaving this aside, though, there is a different reason to doubt the success of this attempt to resuscitate Premise 2*. In order for the conclusion of the four-step argument to follow, both the Pareto principle (Premise 3) and one’s ethical theory that is employed in the conclusion about IMI will need to be described in forward-looking terms. In other words, for someone who employs a forward-looking notion of personhood, one’s ethical theory can only be concerned with the present and future. However, it seems implausible that agent-neutral goodness would not be temporally neutral. I will defend this claim at length in Part III Chapter 3, but to give one illustration of why I think it seems implausible, consider the following: Consider a version of the 188 hospital case where there is only one patient—you—and, in fact, no one else exists in the world. You read a note that says that you either experienced ten hours of pain and have none more in store or you experienced no pain and have three hours of pain in store. It seems that it would be a better state of affairs, agent-neutrally, to have not yet had surgery. If this claim of mine is correct, then the problem with the current account to resuscitate Premise 2 is not that it fails to make Premise 2 true—it does make Premise 2 true. It’s just that in order to do so, it changes assumptions about agent-neutral goodness that are employed in Premise 3 and in the conclusion, and in what I take to be implausible ways. The thoughts expressed here are merely preliminary, however, and this account will be considered in further depth in Part II Chapter 3. 2.4.B.2. To summarize the findings of this sub-section: In light of the fact that I argued that 2A is false, I suggested we consider whether 2A actually is necessary for The Common Argument for IMI (or a sufficiently similar argument) to succeed. After noting that 2A is needed for the argument to succeed, I considered a similar argument that did not require 2A. The changes required in order for the argument to no longer need 2A, however, were such that the new argument fails—or so I argue. Thus, 2A is an important piece of The Common Argument for IMI, and the following offers a more general point about why “the desire to maximize one’s happiness in the future” cannot play the role that “the desire to maximize happiness for one’s life as a whole” plays in serving as a spring board—when combined with 2B—to get to a metric for goodness for a person. While it is true that both are desires, and they are similar in both this respect and some more specific respects, there is an important difference between the two potential desires. First, 189 however, two similarities: Suppose that at any point in time, a person can have a desire regarding his full life and the desire regarding the remainder of his life. In one sense, both of these types of desires have new tokens at each new instant of consciousness, because there is a new conscious moment of experiencing the desires of each type. This is a similarity. Further, in a sense, both types of desires—at all times at which they are instantiated—take as their object a single thing: in one case, one’s full life, and in the other case the remainder of one’s life. These similarities notwithstanding, there is an important difference: Unlike the desire to maximize one’s lifetime happiness temporally neutrally, one’s desire to maximize one’s future happiness—i.e. the happiness of the remainder of one’s life—does not always refer to the same thing. What constitutes the remainder of one’s life is a function of the time-indexing of the statement. What constitutes the remainder of one’s life at t2 will be different from what constitutes the remainder of one’s life at t1—namely, the latter will not include the portion of one’s life between t1 and t2, whereas the former will include this portion. In light of the points made in this section, it seems that serious doubt has been cast on (the original) Premise 2, and, more generally, on The Common Argument for IMI. 2.4.C. Two Additional “Loose End” Arguments The two additional loose ends I stated that I would consider are: (1) whether we can get the output of the Three-Step Argument for Premise 2—namely Premise 2 (““If IMF is true, then the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.”)—by a route other than the Three-Step Argument for Premise 2. In other words, I consider whether we can cast doubt on the possibility that there is another way of establishing the plausibility of Premise 2; and (2) whether we can get the output of Premise 2—namely, the consequent of its “if then statement”: “[T]he metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime”—by a route other than the plausibility of Premise 2 as a whole (i.e., without using the route of establishing the “if-then” statement. In other words, I consider whether we can cast doubt on the possibility that 190 there is another way of establishing the plausibility of the claim that “the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime.” It turns out that doubt can be cast on the success of both of these types of alternative routes if I can cast doubt on the claim that the metric for assessing the goodness for a person is the amount of happiness the person experiences, as measured by TU-type aggregation, over the course of his lifetime. I do intend to address head-on this question about what is the most plausible metric for assessing goodness for a person, but I will largely defer this discussion to Part II Chapter 3, which will deal with this question in great depth. My conclusions there, however, will be that both TU- type aggregation and AU-type aggregation are plausible accounts of goodness for a person, and that there is no knock-down argument available for either one. I will then proceed to argue, however, that there are considerably stronger reasons to espouse AU-type aggregation, rather than TU-type aggregation, as the aggregation mechanism used to assess goodness for a person. Even if I weren’t to make the argument that AU-type aggregation is a more plausible account than TU-type aggregation as an account of goodness for a person, however, the mere fact that I argue that both accounts are plausible views—and that neither is susceptible to a knock- down argument against it—is sufficient to accomplish my goal here in Part II Chapter 2. This conclusion sufficiently casts doubt on the thesis that IMI is correct, because it shows that it is not necessarily the case that TU-type aggregation is the correct metric for assessing goodness for a person. Further, to the extent that I argue that AU-type aggregation is the more plausible metric for assessing goodness for a person, this further strengthens my argument against the plausibility of IMI. Although I will wait until Part II Chapter 3 to provide the arguments for these conclusions in depth, I will now provide a brief preliminary argument to cast doubt on TU-type aggregation as 191 the aggregation method for determining the goodness of a life. (This argument will be one among many made in Part II Chapter 3.) Consider the following two thought experiments: First consider the first thought experiment, raised by Temkin, and called “The Single-Life Repugnant Conclusion”: 157 If TU-type aggregation is the correct measure for assessing the goodness of a life, then the following is an implication of this view: For any life of however many moments, each of fantastically high happiness level, there is a better possible life that exists that is composed only of moments (as long as there are enough of them) of a positive value yet such an infinitesimally small positive value that a person is almost indifferent to being alive, and is just the slightest notch above the happiness level at which one would prefer to not be alive. This is an implication of TU-type aggregation, because if there are enough moments of the positive yet tiny value, they will outweigh the number of units of happiness in the shorter life (despite all the moments in the shorter life being of a fantastically high happiness level. This conclusion seems to be repugnant, though, and thus the fact that this is an implication of TU-type aggregation provides reason to think that TU-type aggregation cannot be correct. It elicits the intuitions that there is at least some importance of and relevance of the happiness level that one experiences at the different moments of one’s life and that everything does not boil down to the total number of units of happiness experienced. While this thought experiment was couched in fairly abstract terms and also in terms of a person whose moments of experience are at such a low level that they are just a notch above the level at which the person would prefer not to be alive, I can offer an additional thought experiment that elicits the same intuitions, but which is slightly more applied and which does not necessarily employ a person being quite so close to the level where he would prefer to not be alive. This 157 Temkin, Rethinking the Good, p. 119. 192 thought experiment, attributable to Roger Crisp, 158 and entitled Haydn and the Oyster, asks: Would one prefer to have the life of the composer Haydn, who lives a good life and dies at the age of 75, or would one prefer to live a two-thousand-year life as an oyster? This thought experiment, like the Single-Life Repugnant Conclusion, is meant to elicit the intuitions that TU-type aggregation cannot be correct, and that there is at least some importance of and relevance of the happiness level that one experiences at different moments of one’s life and that everything does not boil down to the total number of units of happiness experienced. Thus, although I will wait until Part II Chapter 3 to fully explore the question of the relative plausibility of AU-type aggregation and TU-type aggregation in the context of determining the goodness of a life, the foregoing thought experiments are a preview of my arguments in Part II Chapter 3. My point is that although people seem to dogmatically assume that TU-type aggregation is the correct aggregation method in the context of determining the goodness of a person’s life, this is not obviously the correct view—and the two thought experiments I just provided do a good job of illustrating that. I will argue in Part II Chapter 3 that AU-type aggregation and TU-type aggregation are both plausible accounts in this context, and that it is an open question which account is more plausible. I will ultimately conclude, though, that there are stronger reasons in favor of AU-type aggregation than TU-type aggregation, and thus AU-type aggregation is the more plausible account. 2.4.D. Outlook To conclude the section here, it’s worth emphasizing that, even if the various arguments provided here in Section 2.4 were to fail (and, also, even if the arguments provided in Section 2.2 were to fail), it still is the case that I have successfully cast doubt on IMI. After all, in addition to 158 Roger Crisp, Mill on Utilitarianism, London and New York, Routledge, 1997, p.23. 193 making the above arguments, I have also argued that Premise 3 of the Common Argument for IMI is false: The person-based Pareto principle is false. This point, on its own, is sufficient to cast doubt on IMI. Of course, the success of my various other arguments, however, do provide important further reasons to doubt the plausibility of IMI. Thus, on the whole, this chapter has provided strong reasons to doubt the plausibility of IMI. Since IMI is inconsistent with AMU, and since IMI has prima facie plausibility, casting doubt on the plausibility of IMI provides substantial support for the plausibility of AMU. Section 2.5: Further considerations 2.5.A. Implications of the possible cases of divergence between goodness for a person and agent-neutral goodness At this point it is clear that the AMU theorist claims that the person-based Pareto principle is false. Further, the AMU theorist need not deny that something can be good or bad for a person. In light of this, the AMU theorist denies precisely what the person-based Pareto principle accepts: that there is a direct link between what is good for a person and the notion of agent-neutral goodness. Importantly, though, it’s not that there is no link. What, precisely, the link is, though, will depend on various factors, including but not limited to: what one’s measure is for the good for a person (e.g. an AU-type or TU-type view), how many other individuals beyond the individual in question exist (if any), how the happiness of one person’s life compares to the happiness of the lives of others. These indirect links notwithstanding, though, there is certainly a less direct link, for AMU, between good for a person and agent-neutral goodness than there is for views that take the unit relevant for ethical inquiry to be the person—and not the conscious moment. What, then, are the implications for human reasoning of this less direct link? Is it normatively important that something is good for persons? Does an agent have normative reasons to do what’s good for himself? 194 In short, I don’t think that adopting AMU changes the landscape here much, if at all. To the extent that one had moral reasons to do what complied with or furthered one’s moral theory, this is still the case. To the extent that AMU prescribes different content than did one’s prior moral theory, one will have moral reasons to do something different. Is it normatively important morally that something is good for a person? It both was relevant on a person-based morality and also is relevant on a conscious-moment-based morality, but with different degrees of directness. On a person-based morality, that something is good for a particular person would be relevant, but only one piece of the picture. For example, other persons would be relevant as well. On a conscious- moment-based morality, as described in the previous paragraph, there still is a link between good for a person and agent-neutral goodness, so good for a person would still be normatively important morally, but more and perhaps more varied information would be required before having the pieces to carry out satisfactory moral deliberation. Knowing that doing b would be good for person x can count in favor of doing b for both a person-based theory and a conscious-moment- based theory, though it would do so in different ways. Thus, while moral reasons might change somewhat if one adopts AMU, it seems that not much will change. As for one’s normative reasons to act in one’s own self-interest, this, too, does not seem to change much, if at all. The change that was allegedly brought about was a less direct link between what’s good for a person and agent-neutral goodness. As such, normative reasons about what is good for oneself remain unchanged. Of course, I have yet to argue for which view of goodness for a person is the correct one (be it the AU-type view, the TU-type view, or some third option), and I have not articulated whether one has normative reasons to do what’s best for oneself or to follow IMF—if they do in fact diverge. Lastly, I have not taken a stand about whether we even have self- regarding normative reasons at all. It could just be that it’s not the case that we ought to care about ourselves (or a subset thereof) and it’s just that nature constrains us and it is a psychological fact 195 that we do in fact care about ourselves (or a subset thereof). None of this uncertainty, however, has been brought about by the moral theory AMU. All of these different approaches to self-regarding reasons are consistent with person-based moralities as well. Similarly, it doesn’t seem as though there need be any change, after adopting AMU, to the relationship between potential normative reasons regarding oneself and potential moral reasons. Perhaps Sidgwick’s claims regarding the “dualism of practical reason” are correct, or perhaps other claims regarding the relationship between self-interested reasons and moral reasons are correct. I will remain agnostic on these debates. An adoption of AMU does not bring about any new considerations to these debates. All that AMU does do is show that what’s good for a person does not plug as directly into the balance for agent-neutral goodness as it would have if one were working within a moral theory that treated the person as the unit relevant for ethical inquiry. 2.5.B. Further points regarding the relationship between ATLU and AMU Despite the arguments in this chapter, one might still think that there is a tension between AMU and IMF. One might think that, regardless of how the details work out regarding what is best for a person on the whole, the truth can’t be too far from what IMF seemingly prescribes, and the gap between AMU and IMF is just too great to think that they are compatible. This, one might think, is particularly the case in light of the fact that ATLU—which shares much with AMU— diverges precisely in regard to their treatment of IMF. Furthermore, ATLU, unlike AMU, appears to have no tensions or difficulties in accommodating the commonsense view about IMF. Thus, one might think, perhaps one should reconsider ATLU. I disagree, both because I think the arguments in Part II Chapter 1 succeed, and because I think the concerns raised in this chapter have been quelled. Nevertheless, some additional points can be made to hopefully appease the ATLU 196 sympathist. In what follows, I will address some of the divergences between AMU and ATLU. I will do this first in general terms, and next, I will specifically address AMU’s treatment of death. I will explore how death might be bad and how it might not be bad impersonally—i.e. the question here is not whether death is bad for a person, but rather, if and how death might be bad ethically, for society on the whole. In what follows, I will bracket the question of what is good for a person and I will simply assume, for the purposes of what follows, what the ATLU sympathist is inclined to believe—that the only relevant fact about self-interest is either IMF or something that is similar in relevant respects. The commitment to something along these lines (and the difficulty in reconciling this with AMU), I take it, is what is keeping the ATLU sympathist from getting on board with AMU. 2.5.B.i. General Considerations Regarding the Relationship Between ATLU and AMU One thing the AMU theorist can say is as follows: As has been stipulated, even on AMU, it is a psychological fact about persons that we strive to maximize our own accrued total happiness over the course of our lifetimes. No matter whether we intellectually ascribe to a reductionist or non-reductionist view of persons, at a lower level we view ourselves in a non-reductionist light. Similarly, and likely because of evolutionary reasons, we view others through a similar lens—in most cases our moral thought takes the person as the unit relevant for ethical inquiry. As such, I hypothesize that our everyday moral sense is much more closely related to ATLU than AMU. (Importantly, however, in Part III, I will consider the various ways in which AU, on the whole, seemingly diverges from our intuitions. I will handle those concerns at that time.) As such, in light of the fact that we generally view what is best for ourselves, as persons, and also for others, as persons, as maximizing our total happiness accrued during our lifetimes, it is not a stretch to think that this might naturally affect what we think is best tout court. In fact, in light of how deeply our 197 lives are steeped in thought, talk, and action about persons, it would be surprising for AMU to correspond to our pre-theoretical moral intuitions. The second point that the AMU theorist can and should make is related to the first point: As a matter of fact, there won’t be all that many cases that we confront where ATLU and AMU diverge in practice. This is due largely to two facts: First, we have the desire to maximize our own total accrued happiness; second, folk morality will largely correspond to ATLU (or, at least in the sense relevant here: that, all else equal, a longer life is good impersonally as well as for the person in question himself). Thus, taking these two facts as given, it will usually—though certainly not always—be the case that states of affairs that maximize ATLU will also maximize AMU. In fact, a proponent of AMU could and might use ATLU as a rule of thumb to help maximize AMU. The hypothesis here can be compared structurally to the case of utilitarians taking as given the fact that people might have a taste for fairness and or equality, and thus, utilitarians might end up placing high instrumental value on fairness and or equality. In some cases, there could even be such high instrumental value associated with these notions that the prescriptions of utilitarianism might not noticeably diverge from the prescriptions given by fairness theorists or equality theorists. Thus, further, in light of the fact that ATLU’s prescriptions often will converge with those of AMU, this might further entrench people’s belief in ATLU and lessen the likelihood that people reflectively adopt AMU, even if AMU is correct. Thus, while the AMU theorist’s first point aimed to explain the role of our pre-theoretical intuitions, his second point aims to show how our post-theoretical intuitions can potentially explain our sympathy for an ATLU-type theory—and this is particularly the case in light of the facts about our psychology addressed in the first point. 2.5.B.ii. The badness of death (including murder), ethically speaking ATLU as a theory has a fairly commonsensical treatment of death: If one’s life was going 198 to continue as a happy one on net, then death is bad (from a societal moral standpoint) because while the denominator considered in the societal average stays the same (i.e. one life), the total amount of happiness accrued will be lower than it would have been if not for the death. An analogous analysis can be carried out if, but for the death, the individual’s remaining life would have been an unhappy one on net. This, however, is not the case for AMU. Although death may well be bad for a person, it is not the case that this would therefore mean that it is bad impersonally. The fact of the matter is that on AMU, while the death of a person (including a death brought about by murder) who would otherwise live additional happy years can be a bad thing, it could also be a good thing. It depends on how the AMU of the additional years compares to the AMU of society not including these additional conscious moments. This is something that many readers might find counterintuitive, and thus a proponent of AMU needs to have an account of why our intuitions about death—and especially about murder—are the way they are. According to AMU, persons aren’t the units relevant for ethical inquiry, and as such, an action that is good for all persons affected is not necessarily good impersonally. This notwithstanding, the units relevant for ethical inquiry—conscious moments—occur in persons, and persons have a taste for person-affecting intuitions. IMF is a psychological fact about us, and these mental states also infiltrate our views about the lives of those other than ourselves. Thus, the units that are relevant for ethical inquiry are greatly affected by what they perceive to be happening to persons—be it themselves or others. Thus, even on AMU, life and death and persons are highly important. Our perceptions regarding the goings ons of persons make up the vast majority of our perceptions and our pleasures and our pains. This is the case even though persons are only of value instrumentally, or derivatively. Thus, despite the fact that something like a murder might not have intrinsic disutility, in no way am I saying that a murder would not be very 199 bad. In the vast majority of cases it would have disastrous results for utility. But, in many cases (all of those in which the additional years that would have been lived would have brought the societal AMU down) the disutility would not be for intrinsic reasons. Thus, in no way am I saying we should ignore our person-affecting intuitions or our IMF- based intuitions. Very much the contrary. In practical cases, we need to keep these tastes in mind. Furthermore, these person-affecting and IMF-based intuitions will have particularly strong force in the short term, in decisions that affect people that we know or otherwise are relatively easy to identify with and whose shoes we can fairly easily place ourselves in. As such, these tastes will be more instrumentally relevant for actions whose effects reside primarily in the present than for those whose effects are more distant. These claims can be corroborated somewhat by comparing our views of killing animals to our views about killing other human beings. As described above, even if AMU is true, it is quite likely that people not only will continue to believe something more like ATLU. Furthermore, this might be precisely what an AMU theorist would prescribe for the commonsense morality of the populace to be. Not only might ATLU work fairly well as a rule of thumb of action in order to bring about the good results on AMU terms, but it might even be the case that not using these rules of thumb would yield worse results on AMU terms than would be yielded by using them, and if so, then AMU is what Parfit would call an indirectly self-defeating theory. If true, this is not necessarily a strike against AMU, and Parfit certainly wouldn’t think it is. The question of whether AMU is indirectly self- defeating, and if so, what the implications of this might be, will be addressed in more detail in the dissertation’s conclusion. Also discussed there will be how these issues regarding AMU relate to Sidgwick’s claims about utilitarianism and whether a morality should be “esoteric.” 200 2.6. Summary Part II Chapter 1 argued that the different location for additional conscious moments is not morally relevant, and thus that the conscious moment itself is the unit relevant for ethical inquiry—if true, thus meaning that ATLU is not correct. Thus, after Part II Chapter 1, it appeared that AMBP (the population-wide anti-more-is-better intuition) and IMI (the intrapersonal maximizing intuition, which states that, all else equal, it is better agent-neutrally that a happy person live longer) are incompatible—we cannot simultaneously hold both of them. It thus seemed that we must adopt the same method of aggregation both intrapersonally and interpersonally—be the method that of AU or TU (or of some unidentified other option). It is the need to make this decision that leads to the fact—alluded to in the text of Part II Chapter 1—that the four-step argument of Part II Chapter 1 could be seen as a reductio against Premise 2 (i.e. AMBP, and, more generally AU), if one accepts Premise 1 and one accepts that the conclusion follows from the premises. I, however, do not think that AMBP (and, more generally, AU) is false, and I think that the argument is a good one. Thus, I think it is IMI that must be rejected. IMI, however, seems to be prima facie plausible to many, and this might be a tough pill to swallow. This notwithstanding, I think it must be done. The goal of this past chapter (Part II Chapter 2) has been to explain why, contra prima facie intuitions, it is plausible to reject IMI. Section 2.1 explained the attraction of IMI and offered a four-step argument that I think is a formalized version of why people find IMI plausible. I then explained why this argument fails: Section 2.2 argued that Premise 2 is false. Section 2.3 argued that Premise 3 is false. Section 2.4 then considered other routes for getting to the conclusion of the argument. Section 2.5 considered further implications. At this point, I thus hope to have cast doubt on the plausibility of IMI, and casting doubt on the plausibility on IMI in turn provides strong support for AMU (because AMU and IMI are 201 incompatible and IMI has prima facie plausibility, thus rendering IMI’s seeming plausibility an obstacle in the attempt to defend AMU). Although a brief preview of the following argument was provided in Section 2.4.C, Part II Chapter III will argue that TU-type aggregation, despite prima facie intuitions one might have to the contrary, is not necessarily the correct aggregation mechanism for agent-relative goodness; not only is AU-type aggregation also a plausible account of aggregation for agent-relative goodness, but it is actually somewhat more plausible of an account of agent-relative goodness than is TU-type aggregation. These conclusions of Part III Chapter 3 will provide further support for the conclusions of Part II Chapter 2—that IMI should be rejected. Thus, the conclusions of Part II Chapter 3 will provide further support for AMU. =================================================================== Part II Chapter 3: The debate about the badness of death for the person who dies, and the underlying accounts of what is good for a person 3.0. Introduction What I include here, under the heading of “Part II Chapter 3,” is a summary of Part II Chapter 3, and not its full text. The full 120-page chapter is available upon request. Despite not including the full text of the chapter here, however, including a summary of what occurs in the chapter is indeed important. This is for a few reasons, the two main reasons of which are the following: Firstly, and most importantly, this chapter makes points that are highly relevant to the project of Part II—both in terms of the exploration of topics that are central to the project of Part II, and, importantly, also in terms of providing strong support for the thesis of Part II, thus substantially contributing to the success of my arguments in Part II. Secondly, this chapter also sets the foundation for a number of concepts, points, and ideas that end up playing a large role and end up getting relied on significantly in arguments and 202 discussions in later parts of the dissertation, especially in Part III. Absent the foundation set here, some of these later discussions and arguments would lack an adequate foundation. For these reasons, it is useful to provide at least a sketch of what is accomplished in the full text of Part II Chapter 3. Having provided this explanation of why a summary of this chapter is being included, I now proceed to sketch out what occurs in the chapter. * * * * Up until now, the focus of the dissertation has been on ethics, and it has been argued that AMU is a plausible theory. This chapter, however, will temporarily move away from ethics and address the topic of goodness for a person. In Part II Chapter 2, the question of what is best for a person arose, and I addressed at least to some degree what I deemed to be the two components to an account of goodness for a person: the aggregation mechanism component and the temporal component. As to the aggregation mechanism component, I argued that both AU-type aggregation and TU-type aggregation to assess goodness for a person are at least prima facie coherent, but I then largely bracketed the question of which view is more plausible (though, in Section 2.4.C, I did provide a brief preview my arguments of this chapter for the claims that (1) both AU-type and TU-type aggregation, even after full reflection, are tenable accounts, and that (2) all things considered, there are, however, stronger reasons to be in favor of AU-type aggregation rather than TU-type aggregation in this context). As to the temporal component, I also briefly touched on the question of which of the possible temporal accounts of personhood is most plausible, but I then ultimately bracketed this question as well. I return to these two questions now. Thus, putting these two sub-questions together (and, more generally), this chapter will address the question of how to assess goodness for a person. Unsurprisingly, the question of what is good for a person is inextricably tied to the question of whether death is bad for the person who dies. Thus, this chapter specifically explores 203 the badness-of-death-for-the-person-who-dies debate, and it does so before getting to the general question of what the most plausible theory is for assessing goodness for a person. Before proceeding, it’s important to reiterate to the reader that, even though this chapter is devoted to exploring the notion of goodness for a person, this chapter will indeed have key implications for AMU and its plausibility—and thus key implications for the plausibility of the thesis of Part II (that AMU is the most plausible version of AU) and for the plausibility of the whole dissertation (that AMU is the most plausible candidate for Theory X). 159 If I’m able to cast doubt on the seemingly-universally-believed claim that TU-type aggregation is the correct aggregation mechanism in the context of determining the goodness of a person’s life, this will strongly support my argument in Part II Chapter 2 that IMI is false, and this, in turn, will strongly support the plausibility of AMU. Further, I believe that in this chapter I do indeed successfully cast doubt on the claim that TU-type aggregation is the correct aggregation mechanism in the context of determining the goodness of a person’s life. Although people seem to dogmatically assume that TU-type aggregation is the correct aggregation method in the context of determining the goodness of a person’s life, I argue in this chapter that this is not obviously the correct view— and the two thought experiments I introduced in Part II Chapter 2 (the Single-Life Repugnant Conclusion, and Haydn and the Oyster), and which I address in greater depth here in this chapter, do a good job of illustrating that. I will argue in Part II Chapter 3 that AU-type aggregation and TU-type aggregation are both plausible accounts in this context, and that it is an open question which account is more plausible. I will ultimately conclude, though, that there are stronger reasons in favor of AU-type aggregation than TU-type aggregation, and thus AU-type aggregation is the 159 In addition to having key implications for the plausibility of AMU, this chapter also has merit because it (1) addresses a question that is interesting in and of itself, and (2) provides a foundation for discussions and arguments in later parts of the dissertation. 204 more plausible account. 160 * * * * Having made these introductory points, Part II Chapter 3 proceeds as follows. In Section 3.1, I address the long-standing debate about whether death is bad for the person who dies. I first canvas the different positions, and I next attempt to reconstruct the respective positions, which, at least in the case of Epicurus’s position, is far from clear—and thus the task of reconstruction is not a simple one. In carrying out this task, I explore different possible accounts of goodness for a person that the two camps might subscribe to—even though the camps do not speak in terms of a general theory of goodness for a person. Finally, I attribute a particular theory of goodness for a person to each camp, and this thus enables me to bring clarity to what the badness-of-death debate is actually about—and this is a clarity that, surprisingly, does not exist in the literature as of the time of my writing this. In Section 3.2, I turn to the various possible accounts of goodness for a person. I identify and articulate what the various possible accounts in the conceptual landscape are, and I assess their initial plausibility. I then more rigorously assess the plausibility of the possible accounts, and, in so doing, I separately address the two components of these accounts—the aggregation mechanism component and the temporal component—and I assess the plausibility of the various possibilities for each component. Then, in Section 3.3, I address some further implications of my conclusions regarding the most plausible account of goodness for a person. Among these implications is the finding that the framework I introduce regarding what I call the “first-person account of personhood,” the “second-person account of personhood,” and the “third-person account of 160 Despite the key differences between the questions of what is good agent-neutrally and what is good agent- relatively, it should not come as a surprise that there are ways in which the questions are related. First of all, this was discussed at length in Part II Chapter 2. Second of all, among the various ways in which the two domains are related, AMU has various implications, one of which is the perhaps-counterintuitive position that death—and even murder—is not intrinsically bad (agent-neutrally). (Of course, as discussed, however, in light of psychological facts about us, the instrumental disvalue associated with death and, especially, murder is typically enormous.) 205 personhood” provides interesting and important new insights that help shed light on and explain what is going on in instances of “weak will.” Lastly, in Section 3.4, I discuss the various implications that the conclusions in this chapter about goodness for a person have for the plausibility of AMU. (In addition to the main implication discussed earlier in this introduction— i.e. that showing that TU-type aggregation is not the only plausible view for aggregating the goodness of a life (and that indeed is somewhat less plausible than is AU-type aggregation)—, there are other implications that the discussions in this chapter have for AMU.) Thus, I conclude the chapter by explaining the various ways in which the findings in this chapter relate to my arguments for AMU. 3.1. Is death bad for the person who dies? Epicurus 161 says that it is not, whereas Nagel, 162 McMahan, 163 Feldman, 164 and Kagan, 165 among others, say that death can be bad for the person who dies. In this chapter, I begin by arguing that it is not immediately clear exactly what is being debated in this debate. I then attempt to clarify the positions in question—most importantly the sparsely described view of Epicurus—and, especially with respect to Epicurus, I will offer what seem to be the most coherent and plausible versions of the accounts in question. This clarification of the authors’ positions will be valuable in and of itself. Having done this, I will then provide insight into where precisely the disagreements in the debate lie. This chapter begins by describing Epicurus’s arguments for the claim that death is not bad for the person who dies. It then proceed to describe the arguments offered by McMahan, Nagel, 161 Epicurus. 162 Nagel. 163 McMahan. 164 Feldman. 165 Kagan. 206 Kagan, and Feldman for why Epicurus was wrong. According to them, their deprivation account provided the tools to see that death can in fact be bad for the person who dies. As I discuss, there seems to be a general consensus that the deprivation account is correct, and that it successfully responded to Epicurus’s arguments. Despite this consensus, however, I argue that the deprivation account does not successfully dispatch Epicurus’s position. While I argue that it is perhaps the case that the deprivation theorists successfully responded to the few arguments that Epicurus offered, I argue that there are further points that could be raised in defense of the Epicurean position, and I argue that rejecting Epicurus’s position solely on the basis of the arguments of the deprivation theorists would be premature. In order to see why this would be premature, and to better understand the differences between the warring camps, I argue that it is necessary to widen our scope to focus on what the different camps’ positions would be on a broader question. The debate between the two camps has been cast in terms of whether death is bad for the person who dies, but this, of course, is a specific instance of a more general question—the question of what is good (or bad) for a person. By exploring what the camps’ answers could or should be to this broader question, I argue, we would be better equipped to see where precisely the disagreements between the camps are located, and we would thus potentially be better equipped to provide a rebuttal to the deprivation theorists on Epicurus’s behalf. Despite the fact that most of the deprivation theorists do not offer a direct answer to the more general question about what constitutes goodness for a person, I argue that we could reconstruct an answer on their behalf, and that their account is fairly straightforward: According to them, goodness for a person is measured by aggregating a person’s happiness according to TU- type aggregation. Although this is clear, what is less clear is what temporal component this camp thinks is appropriate. Nagel, McMahan, and Kagan do not address the issue (even indirectly) at all, 207 and while Feldman argues that the analysis should be about a person’s complete lifetime, the explanation is brief and doesn’t consider the reasons for adopting a current / forward-looking account, and it’s not clear that he would maintain (or that he would switch) his position if he were confronted with the various considerations in this chapter—including Parfit’s hospital example. Furthermore, all four of these authors—even Feldman—seem to assume that temporally neutral and current / forward-looking versions of TU-type aggregation would be functionally equivalent and thus that it is not a useful line of inquiry when attempting to articulate an account of goodness for a person. What Epicurus’s answer to the more general question of what constitutes goodness for a person would be, I argue, is far from straightforward. While we will never know what answer to this question Epicurus would endorse, I explore various possibilities for what one would think about this question if one thought that death could not be bad for a person. Different possible accounts have different aspects that render them either inconsistent or implausible, but I come up with one account that passes muster. While this position might not have been the Epicurean position, I determine that it is the best position that is consistent with his claims about death, and thus I take the liberty to—perhaps charitably, to Epicurus—call it the Epicurean position. According to this position, goodness for a person is assessed by AU-type aggregation and in a current / forward-looking manner. Furthermore, according to this position, the notion “person” consists in an individual’s current conscious moments and those in the individual’s future. Temporarily leaving aside the question of which of these two accounts is more plausible, I argue that this characterization of the two positions, while seemingly straightforward, is an important development. Before my exploration and analysis, the debates about whether death is bad for the person who dies were far from clear; in fact, it wasn’t even clear what the debate about the badness of death was even really about. Thus, my conclusions constitute an important step 208 forward. More specifically, they are a step forward for the following reasons (and in the following ways): First of all, this seems to be the first attempt to characterize the camps’ answers to the more general question, rather than merely their answers to the specific question about the badness of death for the person who dies. Second of all, it seems that characterizing the camps’ answers to the more general question makes it much clearer what the camps disagree about. Third of all, and perhaps most importantly, equipped with this analysis, it now becomes clear why the consensus in favor of the deprivation theorist’s position and against the Epicurean position is premature. While perhaps the deprivation theorists responded successfully to Epicurus’s arguments themselves, there are other arguments that can be made in defense of the Epicurean position. The new points that the Epicurean should make are as follows: First: The deprivation theorists simply assume that TU-type aggregation is correct. Not only do they assume that TU- type aggregation is correct, but, for the most part, they don’t even recognize that there are alternative mechanisms for aggregation! Thus, in a sense, the assumption is even more gross. Second: The deprivation theorists for the most part completely ignore the question of which temporal component is appropriate for assessing goodness for a person. While Feldman briefly addresses it, the analysis is far from exhaustive. Further, the deprivation theorists assume that the two most salient temporal accounts (the temporally neutral account and the current / forward- looking account) are functionally equivalent. What’s more, as was the case with their non- treatment of the aggregation mechanisms, here too, they do not even make the assumption about the unimportance of addressing the temporal accounts, but, for the most part, they fail to recognize that they are even making an assumption of this sort. Thus, in sum, equipped with this new analysis, the Epicurean can (and should) now claim that not only has the debate not yet been resolved in favor of either camp, but that, in fact, the debate has not even really begun. Only now, with these different positions clarified, can we 209 engage in a debate about which camp’s view is more plausible. Of course, this in no way speaks to the question of which position is in fact more plausible. The Epicurean is not (in the above comments) saying that the deprivation theorists cannot come up with arguments against the Epicurean position, he is just saying that they have not yet done so. It is at this point that the real debate begins, and it is the task of both camps to provide arguments against the other camp’s position. Then, when the dust settles, we can determine who really has the best of the debate about the badness of death for the person who dies. 3.2. Of course, this is not where the chapter stops—this is merely the beginning. The next task is to consider the different positions substantively, and to make sense of who had the better of the debate—or, rather, which parties would have the better of the debates if their positions were articulated as coherently and plausibly as possible. I address the temporal component and the aggregating mechanism component separately, and, in short, my conclusions are as follows. First, as for the temporal component: I argue that there are at least three different temporal components that have some plausibility—the temporally neutral view, the current / forward-looking view, and the present moment view (although the present moment view might be the least plausible of the three by a good distance). Further, I argue that which temporal component one adopts in a particular instance is partially a function of whether one adopts Parfit’s “Extreme View” of personhood (which I described and endorsed in Part II Chapter 1 Section 1.5.E), and, similarly, the notion of personhood that one adopts toward the individual in question at the particular instance: Corresponding to the three temporal components are the following three notions of personhood, respectively: the temporally neutral view of personhood, the current / forward-looking view of personhood, and the present moment 210 view of personhood. Different people might employ different notions of personhood in different contexts and different times, and find different views of personhood plausible in different contexts and different times, and I argue that there is no fact of the matter about which is correct. A non-Empty-Question Theorist 166 would likely think that there is a fact of the matter, but it’s unclear how one would make an argument for any particular temporal view of personhood. According to the Empty Question view, however, which I argue is more plausible, there is no fact of the matter about which of the three notions of personhood is correct. Notwithstanding this, though, I say that this does not make the inquiry into goodness for a person unimportant, because, regardless of whether there is a fact of the matter about what constitutes a person, it is a fact that we generally think and act as if there is, and there are numerous reasons for why it is valuable to explore our concept of personhood and our concept of goodness for a person. Thus, although, I argue that there is no fact of the matter about which notion of personhood was correct, we can do an empirical investigation into how we generally use the concept of personhood. I argue that we ultimately use different notions at different times, and there is no clear answer. Nevertheless, I do indicate that I think that the current / forward-looking view seems to be more commonly employed, and it does seem to be the view that might be intended by all parties to the badness-of-death debate. For an Epicurean who says that death cannot be bad for the person who dies, and who thus is adopting AU-type aggregation, he must adopt a current / forward-looking view. For proponents of TU-type aggregation who say that death can be bad for the person who dies, for the types of cases that are being compared in the context of this question, it is moot whether they are in favor of a current / forward-looking view or a temporally neutral view, and they would happily stipulate a current / forward-looking view. 166 See Part II Chapter 1 Section 1.5.E for a description of Parfit’s discussion of the Empty Question view and the non- Empty Question View. 211 Next, as for the aggregating mechanism: As was the case with the temporal component, I argue that there is no single fact of the matter about whether AU-type or TU-type aggregation is correct for assessing the goodness for a person (regardless of how we understand the temporal notion of personhood). Also as with the temporal component, though, I argue that which view one finds most plausible will likely be partially a function of one’s views about the metaphysics of personhood. I argue that if one espouses Parfit’s Extreme View (of personhood), then AU-type aggregation seems considerably more plausible. On the other hand, while if one does not espouse Parfit’s Extreme View then TU-type aggregation might be plausible (and it might even seem, prima facie, as though TU-type aggregation is more plausible than AU-type aggregation if one denies Parfit’s Extreme View), it seems, after further thought, as though AU-type and TU-type aggregation might be on a par for the theorist who does not espouse Parfit’s Extreme View—and, if anything, with a slight nod to AU-type aggregation being the more plausible view even for the theorist who does not espouse Parfit’s Extreme View. (The Single-Life Repugnant Conclusion and the Haydn and the Oyster thought experiments provide some (but not all) of the support for this conclusion.) Thus, there is no fact of the matter about which is correct. Notwithstanding this, I argue that based on just the above comments, there is reason to find AU-type aggregation more plausible, and this is particularly the case if one takes into account my claim (and prior arguments) for the plausibility of Parfit’s Extreme View—though there is reason to find AU-type aggregation more plausible even if one does not take into account my prior arguments for the plausibility of Parfit’s Extreme View. Thus, what then is the overall conclusion? With respect to the broad question of what is good for a person, one can adopt any temporal notion of personhood—though the current / forward-looking version is often the most plausible—and while one could possibly adopt TU-type aggregation, one should strive to view what is good for that person as measured by AU-type 212 aggregation. As for the specific question of the badness of death for the person who dies, again, the answer is inconclusive on both dimensions, but I think there is reason to apply the above general conclusions, and thus there is reason to side with Epicurus: Death cannot be bad for the person who dies. Even if I think that this, reflectively, is the most plausible account, the deprivation theorists still might still be able to stake a claim to what people unreflectively—and often reflectively— think. Articulating this certainly has value, but the deprivation theorists were aiming for more. They were aiming to say that what in fact is good or bad for the person (in the context of death). With this project, I think that they are mistaken. Thus, despite the seeming consensus that they had the better position in the debate, I think that the contrary is true. 3.3. Implications of my conclusions regarding the plausibilities of the various accounts of goodness for a person: (1) The problem of deriving an “is” from an “ought” and (2) weakness of will 3.3.A. As I discuss in section 5.4.C of Part II Chapter 3, even if we believe intellectually that AU-type aggregation is correct for assessing the goodness of a life, many of us might still have a difficult time believing this at a non-intellectual level. While one might view oneself or other human beings with the third-person detached view here and there, these situations might be few and far between. As even Epicurus likely would concede, IMF seems to be generally true—we generally desire to maximize our happiness, as measured by TU-type aggregation in a current / forward-looking manner. As discussed, this situation regarding the third-person detached view is similar to the analogous situation in the context of free will, where even if one believes intellectually that human beings do not have free will, we still generally believe non-intellectually that we and other human beings do have free will—and the cases where we exhibit a non- 213 intellectual belief in the lack of free will are few and far between. The fact that Epicurus and I are suggesting both that (i) we should desire to maximize our happiness as measured by AU-type aggregation, and that (ii) we do, as a matter of fact, generally desire to maximize our happiness as measured by TU-type aggregation, leads to an interesting problem that I’ll call the problem of deriving an “is” from an “ought.” How can we enable our (intellectual) beliefs to affect our non-intellectual beliefs and our desires? How can we bring it about that we desire what we should desire (so that we can thus act in accordance with what we should desire)? It could be that this is an instance of the question of how one can bring oneself to be rational, in moments of (or in many moments of) irrationality. However, it’s also possible that the specific question we are confronting here is somewhat different. In this section, I thus discuss and elaborate upon deriving an “is” from an “ought” and related topics. In so doing, I also explore the interesting and tricky question of what kind of a desire IMF is (and, relatedly, the question of what kind of a desire the desire to continue to live is). A number of related questions arise in this inquiry: If IMF is a desire, perhaps it is not something that rationality speaks to (as Hume famously argued when he stated that we are the “slaves of our passions”). But perhaps IMF is not really a desire, but actually a belief about what is best for oneself? What, precisely, is the distinction between a desire and a belief? If IMF is a desire, is it what Frankfurt and others would call a higher-order desire? 167 Or is it a first-order desire? It seems to share some qualities of first-order desires but it also seems to share some qualities of second- order desires. Or is it another type of desire altogether? My inquiry into what IMF is will hopefully illuminate and contribute to our understanding of “higher-order desires” and related topics in the literature that Frankfurt’s famous essay gave rise to. 167 Harry Frankfurt, “Freedom of the Will and the Concept of Person” (Watson ed.). 214 As for the question of how to “derive the ‘is’ from the ‘ought,’” (i.e. how to bring it about that we espouse, adopt, and implement AU-type aggregation in our lives), I address a way in which we might think that there is hope in moving toward implementing AU-type aggregation in our lives: As I discuss in Part II Chapter 3, the “pure” pairings (of aggregation mechanism and temporal account), which naturally go together as a result of one’s views regarding whether or not Parfit’s Extreme View of personhood is true are (1) AU-type aggregation with the current / forward-looking or present moment view, and (2) TU-type aggregation with the temporally neutral view. 168 I consider these “pure,” because metaphysical beliefs might lead to these pairings. As for our moving toward “deriving the is” (i.e. moving toward adopting AU-type aggregation), however, it’s worth first taking note of the fact that a common view that we are familiar with is what I call the “first-person view” (the view that is a conjunction of TU-type aggregation and a current / forward-looking reference class). The fact that we are so familiar with the “first-person view” might—due to its hybridity—suggest that we have at least a foothold toward, or a foot in the door with, AU-type aggregation. It might be that the fact that we commonly hold this “impure” view might help us on our way to the goal of espousing AU-type aggregation, because it might be that it constitutes an intermediate stepping stone because of it having one foot in both pure camps (i.e. one foot in the TU-type and temporally neutral view and one foot in the AU-type and current / forward-looking view). Thus, perhaps recognizing that this foothold exists and strategically taking advantage of its existence might be helpful in various ways in the attempt to bring about the shift from the ought to the is. 168 Although these are what I call the “pure” pairings, we can also have (and often do have) non-pure pairings (such as (1) TU-type aggregation with the present moment view or the current / forward-looking view, and (2) AU-type aggregation with the temporally neutral view), and there is nothing inconsistent about holding espousing an impure view. It just seems as though they go together somewhat less naturally and that they thus might be espoused somewhat less frequently. 215 3.3.B. In this section, I address some further implications of my conclusions regarding the most plausible account of goodness for a person. In particular, I address the finding that the framework I introduce regarding what I call the “first-personal account of personhood,” the “second-personal account of personhood,” and the “third-personal account of personhood” provides interesting and important new insights that help shed light on and explain what is going on in instances of “weak will.” While I offer various findings regarding these three accounts of personhood and the phenomenon of weak will, in sum, I suggest that instances of the phenomenon of weak will occur when we switch, temporarily, to (implicitly) espousing either (1) an account that includes the present moment view of personhood as its temporal account, or (2) an account that includes a current / forward-looking account of personhood as its temporal account, and which includes an AU-type account of aggregation. I provide various discussions about both (1) what situations might cause these switches in the temporal accounts and aggregation mechanisms that we espouse, and (2) the interesting results that occur as a consequence of these switches. I argue that these findings are interesting and that they also provide novel insights in the area of philosophy that explores the phenomenon of weak will. Section 3.4: Implications of Part II Chapter 3 for AMU This is a very important section that will explain a number of the most important ways in which the material of this chapter indeed has an important relation to the overarching thesis of the dissertation—i.e., that AMU is the most plausible candidate for Theory X. There are various ways in which this chapter’s discussions are relevant to AMU, and in what follows, I mention four of these. These are the four most important ways (one of which is, by far, the most important of the four), but other ways that I do not address here exist as well. 216 3.4.A. Relation Number 1 By far, the most important implication of the material in this chapter is that the material helps support AMU by providing additional support for the rejection of IMI (the intrapersonal maximizing intuition—the intuition that, all else equal, it is agent-neutrally better for a happy person to live longer). IMI is incompatible with AMU, and the prima facie plausibility of IMI was seen as a threat to AMU. The goal of Part II Chapter 2 was to offer a reason to doubt the plausibility of IMI. In particular, Part II Chapter 2 considered a four-step argument that I thought was the reason people were attracted to IMI. It then cast doubt on two steps of the argument, the rejection of either of which would be sufficient to make the argument fail. An intermediate conclusion of the argument I was contesting was the claim that, all else equal, a longer happy life is better for a person than is a shorter happy life. This intermediate conclusion was arrived at by accepting a conditional claim (which was one of the two claims that I said should be rejected). However, even if I was right about this, there still could be other routes that one might offer to get to the intermediate conclusion (which, if successful would provide support for IMI). However, the arguments here in Part II Chapter 3 provide independent reasons for rejecting the intermediate conclusion. At the very least, I provide an explanation of why it’s an open question about whether TU-type aggregation is the most plausible measure of goodness for a person, because I suggest that AU-type aggregation is a plausible candidate as well. This would provide reason to doubt the intermediate conclusion. Further, I argue that AU-type aggregation is actually more plausible as a measure than is TU-type aggregation, and this thus provides extra support for the rejection of the intermediate conclusion. Thus, the arguments in this chapter provide additional reasons to doubt IMI, and, as a result, the arguments in this chapter provide support for AMU. 217 3.4.B. Relation Number 2 Another large set of implications: The framework, observations, and discussion in this chapter about goodness for a person will serve as the basis for many important points made in Part III of the dissertation (i.e., the Part that responds to three objections to the general umbrella theory of average utilitarianism—the Bottom Up Objection, the Top Down Objection, and the Two Hells Objection). It will be shown in Part III that many intuitions that might incline one to reject AU are actually mistaken and should themselves be rejected. The discussions in this chapter (Part II Chapter 3) provide the basis for explanations for why those commonly held intuitions should be rejected. Among other things, I’m referring to the two components of IMF: the desire to maximize according to TU-type aggregation, and the current / forward-looking view. I will ultimately be arguing for a temporally neutral and AU-type ethical theory for Theory X, and I will suggest that many of our contrary intuitions are actually being imported from our inescapable intuitions about intrapersonal maximizing. I will leave the details until Part III, but suffice it to say that without this chapter (Part II Chapter 3), Part III’s defense of average utilitarianism would be considerably weaker. 3.4.C. Relation Number 3 One of my main points in Part II Chapter 1 was to argue that the unit relevant for ethical inquiry is the conscious moment and not the person. While I hope to have argued successfully for this, this might have seemed prima facie false to some people, and thus additional support for this point can only be helpful. I don’t think that the points in Part II Chapter 3 directly support this point, but I do think that the considerations in this chapter could contribute to people being more open to the claim that the unit relevant for ethical inquiry is the conscious moment. There are a few ways in which this might work. 218 For one, with respect to the temporal component discussion, I think the fact that we frequently take a current / forward-looking view (e.g., consider the Parfit hospital case), shows that we are capable of finding the unit of personhood as being a subset of what we might have thought the unit of personhood was. In other words, once we move away from the temporally neutral view, we are shifting closer and closer to viewing persons as a collection of moments, or even a single moment. This might move one from the staunch position that it can’t be that conscious moments are the units relevant for ethical inquiry. Additionally, the discussion of the aggregating mechanism might provide similar support. If one does buy the fact that AU-type aggregation could be the appropriate mechanism for goodness for a person, this means that the unit relevant for agent-relative inquiry is the conscious moment and not the whole person. Of course this is the unit relevant for agent-relative inquiry and that doesn’t entail anything about the unit relevant for ethical inquiry, but one might think that there is or should be a connection between these notions in the agent-relative and agent-neutral domains, respectively. Thus, to the extent that the arguments on this point in this chapter are plausible, this might increase the plausibility of the claim that the unit relevant for ethical inquiry is the conscious moment, and thus this might increase the plausibility of AMU. 3.4.D. Relation Number 4 Another point that I think might come out of this chapter and Part II Chapter 2 is that we have a lot of views about goodness for a person that might not be quite what we thought they were. The Parfit examples bring a lot of this out for the first time, and I think that the discussions of the temporal component and the aggregating mechanism show us various ways in which we didn’t quite believe what we might have thought we believed. Thus, I think that this chapter reminds us that we can have confused views and that everything should be open to consideration 219 for possible revision. We should have less confidence in some of our allegedly fixed points in our intuitions, and we should thus be open to thoroughly examining all of our views. Thus, I think that if there people who, for one reason or another, find themselves highly averse to AMU, they should really try to stay open-minded. Conclusion to Part II Having argued elsewhere for the plausibility of average utilitarianism (having provided initial appeal for it in Part I Chapter 2, and defending it against objections in Part III), we are confronted with the question of what the relevant units are that should be averaged over. While the analogous question could theoretically arise for other theories, such as TU (where the question would be which units to take the additive sum of), in practice it would not, because the question would be a moot one for a theory like TU. The different sub-theories of TU would be coextensive. This, however, is not the case with AU, and thus there is a further question of which sub-version of AU is the most plausible. Part II addressed this question. A quick exploration into the matter at the end of Part I Chapter 2 suggested that ATLU is a possible candidate. It seemingly allows us to combine two core intuitions: ATLU maintains the population-size anti-more-is-better intuition (AMBP) (that, all else equal, it is not better that more people exist than fewer people) and it also maintains the intrapersonal maximizing intuition (IMI) (that, all else equal, it is better agent-neutrally that a happy person live longer). Accordingly, ATLU relies on the claim that, all else equal, it is not better for there to be more good conscious moments if they exist in an additional person, but it is better for there to be more good conscious moments if they exist in an already existing person. A question that confronts this view, however, is why the different location for additional good conscious moments is relevant. Part II argued that the location for additional good conscious moments is not relevant, and thus, taking AU as given, 220 it argued that AMU—and not ATLU—is the most plausible version of AU. Part II Chapter 1 offered a positive argument for why AMU is the most plausible version of AU. Part II Chapter 2 then considered various objections to AMU. Most importantly, Part II Chapter 2 explained why IMI, as stated, is false, and thus why AMU is true despite our IMI- friendly intuitions. Part II Chapter 3 considered and addressed the question of whether death is bad for the person who dies, and, more broadly, it considered and addressed what the most plausible account is of goodness for a person (i.e., agent-relative goodness). In so doing, Part II Chapter 3 provided significant further support for the main conclusion of Part II Chapter 2—that IMI if false. Thus, to the extent that Part II Chapter 3 provided support for the claim that IMI is false, Part II Chapter 3 provided strong support for AMU. Lastly, Part II Chapter 3 also considered and addressed some additional objections to AMU. In sum, Part II has argued that AMU is the most plausible version of AU. 169 169 It is worth reminding the reader, at this point, of some additional considerations that I originally articulated at the beginning of Part II: First: In Part III of the dissertation, it is noted that while I think that person-affecting intuitions are mere heuristics for (impersonal) utilitarian goals and thus that they should be rejected when in tension with the utilitarian goals themselves, in practice they end up giving us the same prescriptions in most cases. A related point was made and addressed in this Part: Despite the fact that AMU and ATLU appear to provide very different prescriptions, their prescriptions in practice will not diverge as much as one might think. In the vast majority of actual cases, the two will converge. This is because AMU will take as given the intuitions that people have, which are often, in many respects, aligned with ATLU, as opposed to with AMU. These ATLU-friendly intuitions would thus affect people’s actual experiences (and similarly, conscious moments’ experiences) in many cases, and thus any theory would need to take into consideration, as inputs, the empirical facts about the happiness or unhappiness that actually is brought about by certain actions. In other words, if people in fact have ATLU-friendly intuitions, then in many—if not most—cases, carrying out the prescriptions of ATLU would maximize the utility of people and of conscious moments, and, thus, carrying out the prescriptions of ATLU will be instrumentally valuable according to AMU. Carrying out the prescriptions of ATLU, in most cases, will be the best instrument for carrying out the prescriptions of AMU—the theory that I actually endorse as being the most plausible. Thus, only in fairly circumscribed cases will there be a divergence between the prescriptions of AMU and ATLU. The above-described way in which AMU relates, in practice, to ATLU is akin to utilitarianism’s accounts of various non-utilitarian intuitions: A utilitarian might say that (1) his theory actually agrees with non-utilitarian intuitions in “non-extreme” cases (not only because non-utilitarian intuitions might be helpful heuristics for maximizing utility, but also because, in non-extreme cases, acting in accordance with one’s non-utilitarian intuitions might itself bring about utility and thus further utilitarian prescriptions), and then say that (2) in the “extreme” cases, non-utilitarians are mistakenly influenced by the heuristics that work well for the “non-extreme” cases but which (a) do not work well as heuristics and maximize utility in extreme cases, and (b) do not bring about sufficient utility from acting in accordance with these intuitions (even when we take this utility as given and accept it as a fact for the purposes of calculating happiness levels) that it would be optimific to follow the non-utilitarian prescriptions in these extreme cases. 221 ===================================================================== Part III: A Defense of Average Utilitarianism (the umbrella theory): Handling three main objections to AU (as well as answering some additional objections and providing additional points and discussions) Having argued in Part II that AMU is the most plausible version of AU, Part III will provide a defense of all AU theories, more generally. In so doing, it picks up where Part I left off. In Part I Chapter 2, I offered a prima facie argument in favor of AU. First, I argued that Theory X had to be impersonal (and not person-affecting), and then I argued that it had to be a pure theory (and not a hybrid theory). This then left us with AU and TU as possible candidates for Theory X. I then argued that AMBP (the population-based anti-more-is-better intuition) is a very attractive intuition that provides prima facie plausibility of AU, and I argued that it is an intuition that many readers will share (though it certainly is the case that some will not share this intuition at all and some, while sharing it most of the time, will have some situations where they do not share it). I argued that this intuition, especially when taking into consideration various prima facie problems with TU that do not afflict AU (such as the Repugnant Conclusion), constitutes a strong prima facie reason to espouse (or to at least strongly consider espousing) AU. Even though many people would find this prima facie case I provided in favor of AU (in Part I Chapter 2) to be appealing or persuasive, seemingly everyone rejects AU, and this is the case even if they would, prima facie, find AMBP plausible. Why is this? Why is it that a seemingly coherent view that seemingly plausibly rejects the “creating more happy people, all else equal, is good” view is rejected by most philosophers? In short, seemingly everyone rejects AU Second: The conclusions of this Part might, at first glance, have been counterintuitive to many readers. After all, we are deeply accustomed to thinking of person-data as of paramount importance ethically. As I argued in Part I Chapter 2, however, it seems clear that we must reject person-affecting intuitions, and these were deeply rooted intuitions as well. The move that occurred in this Part is a different move than was the rejection of person-affecting intuitions, but I think the conclusions of this Part are less difficult to handle now that we have already rejected person- affecting intuitions. In a sense, the rejecting of person-affecting intuitions was the first step of a process of which this Part was the second step. 222 because of some well-known objections that people think are devastating objections against AU. Part III’s goal will be to answer these objections. If I am able to successfully answer all of these strong objections against AU, I think that I will be able to convince most readers that AU is correct. After all, as I stated, it seems to be that it is these objections that are what are preventing people from espousing AU. If not for these objections, we would be left with my prima facie argument in favor of AU from Part I, and I think that if a reader were provided with this argument and if the objections against AU discussed in this Part had been disposed of, the vast majority of readers would espouse AU. * * * * While various objections to AU are made, despite their differences, almost all of them 170 seem to be at least somewhat related to a single concern: Assuming that a particular action (e.g. bringing into existence a person with happiness level five) has no effects on others, why should the goodness of the action depend on anything extrinsic to those affected by the action (e.g. the happiness of the rest of society)? AU seems to dictate this conclusion, because the creation of an individual at level five would seemingly be bad if the prior population’s average happiness were level ten, but it would be good if the prior population’s average happiness were level three. This concern is manifested—with varying degrees of directness—in three different types of objections to AU. Most directly, it is manifested in what I’ll call the “bottom up” objection. Somewhat less directly, it is manifested in what I’ll call the “top down” objection. Lastly, although the Two Hells objection has various features, it, too, manifests the above concern to some degree. 170 It is not the case that all of the strong arguments against AU are related to this concern. The Adam and Eve objection is one example of an objection that largely arises out of different concerns. This objection will be addressed in Section 1.F.2, later in this chapter. Additionally, as stated in the paragraph in the main text accompanying this footnote, the Two Hells objection, while indeed somewhat related to the concern addressed in that paragraph, is substantially motivated by different concerns. I address the Two Hells objection in Part III Chapter 2. 223 First, the bottom up approach (which I address in Part III Chapter 1): One might point out that it is not at all clear which pool of individuals one should strive to maximize the happiness of: All sentient beings to ever exist? All people to ever exist? All people who currently exist? All people who exist currently or will exist in the future? It is unclear whether a non-arbitrary selection of any of these can be made. Further, and more importantly, according to this characterization of the concern, no choice of a relevant temporal population will yield plausible results for the AU theorist. Thus, according to the bottom up argument against AU, AU is implausible because there isn’t a single “reference class” 171 of individuals to average over that will render AU a plausible theory. The top down approach to objecting to AU (which I address in Part III Chapter 3) is to say that AU fails because it doesn’t always maximize what is of intrinsic value. The proponent of the top down objection says that since, according to utilitarianism, happiness is of intrinsic value and AU seeks to maximize something that often does not maximize the amount of happiness, AU fails to correspond with our intuitions. Parfit, McMahan, and Huemer are all examples of authors who provide a top down argument of this sort. Lastly, the objection to any version 172 of AU that is perhaps regarded as the strongest is one that Parfit raised and termed Two Hells. 173 Among other things, it addresses AU’s treatment of suffering. The objection goes as follows: Consider two possible hells. In Hell 1, ten people exist and suffer great agony. In Hell 2, there are ten million people who suffer ever so slightly less agony. AU will say that Hell 2 is better. Parfit and countless others, however, find it obvious that Hell 2 is far and away the worse of the two states of affairs. 171 I will define and explain this term in Part III Chapter 1. 172 Any version both (1) with respect to the relevant population and (2) with respect to the choice between AMU, ATLU, and any other AU theories that employ a different unit as the unit relevant for ethical inquiry and / or use a different aggregation mechanism to calculate the goodness of an instance of the unit relevant for ethical inquiry given the values of the building blocks. 173 Parfit, p.393. 224 While the three types of objections, just discussed, constitute the main objections against AU, this Part also addresses a handful of other objections to AU (both well-known and non-well- known objections). Two of these other objections that I address are Parfit’s Adam and Eve objection and Arrhenius’s Sadistic Conclusion objection. In sum, Part III will consider the main objections that are typically considered the strongest objections to AU theories—objections responsible for the general consensus that AU theories are implausible. My view is that, in Part III, I will successfully respond to each of these objections, and that I thus defray each of these objections against AU. 174 Thus, when this Part is taken in conjunction with the prima facie argument that I make for AU in Part I Chapter 2, the result is that the reader should be persuaded that AU is the most plausible candidate for Theory X. Thus, having argued in Part II that AMU is the most plausible version of AU, and since Part III provides a defense, more generally, of all AU theories (thus supporting the claim in Part I Chapter 2 that the most plausible Theory X is a version of AU), the conclusion after the first three parts of the dissertation will be that AMU is the most plausible candidate for Theory X. 175 ====================================================================== Part III Chapter 1. Parfit’s bottom up objection, a defense of AU with a temporally-neutral reference class, and an explanation of intuitions to the contrary 1.A. Introduction This chapter addresses a group of related objections that I call the “bottom up objection.” 174 It is also the case that in Part III (as is the case in the other Parts of this dissertation), in addition to making the various arguments and sub-arguments that support my theses and sub-theses, I also make and discuss a number of collateral findings that will hopefully be of interest in and of themselves—even leaving aside the ways in which they pertain to the main arguments and main theses of the chapters in which these findings are made, mentioned, and discussed. 175 Part IV of the dissertation will, of course, reinforce these findings. In fact, Part IV of the dissertation, if it is successful in its goals, would by itself support the conclusion that AMU is the most plausible candidate for Theory X, as long as it is taken together with my arguments in Part I Chapter 2 for why Theory X must be impersonal and non- hybrid. In other words, Part IV would support the conclusion that AMU is the most plausible candidate for Theory X even without Part II, Part III, and the portions of Part I that argue that AU is more plausible than TU. 225 One might point out that it is not at all clear which pool of individuals one should strive to maximize the happiness of: All sentient beings to ever exist? All people to ever exist? All people who currently exist? All people who exist currently or will exist in the future? It is unclear whether a non-arbitrary selection of any of these can be made. Further, and more importantly, according to this characterization of the concern, no choice of a relevant temporal population will yield plausible results for the AU theorist. Thus, according to the bottom up argument against AU, AU is implausible because there isn’t a single “reference class” 176 of individuals to average over that will render AU a plausible theory. In what follows, I first describe the bottom up objection in Section 1.B. In Section 1.C, I make some important preliminary points regarding axiology versus action and whether the bottom up objection should be treated as an objection to AU’S axiology or AU’s account of action. In Section 1.D, I then provide my main response to the bottom up objection, and I explain why I think that the objection fails. In short, I argue that person-affecting intuitions are to blame for the prima facie plausibility of the bottom up argument. Further, also in Section 1.D, I describe and defend what, in my view, is the most plausible version of AU (i.e. the most plausible reference class for AU to average over): a temporally neutral version of AU. In Section 1.E, I provide four additional considerations, each of which provides further (intuitive) support for my main explanation. In Section 1.F, I consider and respond to two additional (and related) objections to AU: First, I address Arrhenius’s Sadistic Conclusion objection and a related objection. Second, I address Parfit’s Adam and Eve objection. In Section 1.G, I explain a few ways in which AMU is even better equipped than is ATLU to fend off the bottom up objection against AU. I believe that my arguments in this chapter, which defend AU against the bottom up objection, successfully defend all versions of AU, but Section 1.G shows why AMU is even better off in responding to 176 I will define and explain this term a bit further below. 226 the top down objection than is ATLU. In Section 1.H, I provide a very brief discussion about how the bottom up objection pertains to questions regarding AU and issues it confronts when considering non-human animals. Finally, in Section 1.I, I briefly conclude. 1.B. Parfit and the Bottom Up Approach 1.B.1. Although the bottom up argument is an argument broad enough to be used against any version of AU, the versions of it found in the literature are all versions that are offered as objections to one specific form of AU—namely ATLU. 177 The reason for this is fairly straightforward: As discussed in chapter 2, aside from a few examples here and there, the vast majority of authors simply assume that ATLU is the only version of AU that exists. Furthermore, among those who recognize that other alternatives exist, it is either explicitly or implicitly assumed that ATLU is the most plausible—or least implausible—version of AU, and certainly the only one meriting consideration. In light of this, I will begin by describing the bottom up argument against ATLU in particular, since this is the version of the bottom up argument that is discussed in the literature. Specifically, I will begin by describing the bottom up argument against ATLU as offered by Parfit, McMahan, and Huemer. After addressing their bottom up argument against ATLU, I will then show how the bottom up argument also can be made against other versions of AU (such as AMU), and, in so doing, I will explain what I take to be the core features of the bottom up argument against the family of AU views, in general. 1.B.2. Parfit argues that we should reject AU, and he uses the bottom up approach to explain 177 Each of these (including ATLU) could, in turn, be further precisified by stipulating which temporal population is relevant. For in-depth discussions of these different possible sub-views, see the discussions throughout this chapter. 227 why. 178 He considers four temporal populations that we might consider to be the relevant ones, and he concludes that AU yields implausible results for each of them. While Parfit also offers his Two Hells objection as a reason to reject AU, I will defer a discussion of Parfit’s Two Hells objection until Part III Chapter 2, because it raises a different concern for AU. Parfit first considers the “temporally neutral” version of AU, according to which it is best to maximize the average happiness of all people to ever exist. Parfit thinks that this is absurd, because, as Jeff McMahan’s “Egyptology” example shows, a temporally neutral version of AU implies that whether one should have a child depends on facts about all previous lives, including those of the ancient Egyptians. He writes: “If the Ancient Egyptians had a very high quality of life, it is more likely to be bad to have a child now.” 179 But, as Parfit says, “research in Egyptology cannot be relevant to our decision whether to have children.” 180 Next, Parfit considers the possibility that what AU requires is that we maximize the average happiness of those who live after the action in question. This, however, dictates that, all else equal, we would bring about a better outcome by killing everyone except for the happiest people who are currently living. This, too, seems absurd. Third, Parfit considers the possibility that what AU requires is that we maximize the average happiness of all those who currently exist and all those who will exist in the future. This, however, according to Parfit, leads to what I will call Reverse Egyptology—because of its relation to McMahan’s Egyptology concern. The concern here is that according to a current / forward- looking version of AU, whether it is good or bad to have an additional child who will experience a particular level of predicted happiness will depend on the predicted levels of happiness of individuals who will live in the distant future. The happiness of these individuals in the future, 178 Parfit (1984, p. 420). 179 Parfit (1984, p. 420), citing McMahan (1981). 180 Parfit (1984, p. 420). 228 however, strikes Parfit as irrelevant to the goodness of a family’s current procreation decision. The decision can be illustrated as follows: Possible Future 1: 181 Possible Future 2: Have the child: 5, 5, 6, 6 | 18 AU = 8 Have the child: 5, 5, 6, 6 | 1 AU = 4.6 Don’t have the child: 5, 5, 5, x | 18 AU = 8.25 Don’t have the child: 5, 5, 5, x | 1 AU = 4 Thus, according to a current / forward-looking AU, it is better to not have the child if Future 1 will occur, but it is better to have the child if Future 2 will occur. Lastly, Parfit and Huemer both consider a version of AU that includes only currently existing individuals. According to them, even AU including just this temporal population gets the wrong result in Parfit’s case of “Mere Addition” 182 and Huemer’s similar case of “Benign Addition.” 183 These two cases, versions of which are depicted below, involve a comparison between world A, where one person exists, and world B, where the same person exists in addition to other individuals who are less well off, but who nevertheless have lives worth living. “Addition” is not meant to signify a temporal process, but just the imagination of a different world, similar to the first, but which includes additional people. According to AU, A would be better than B, but Parfit and Huemer think that AU gets the wrong result. Mere Addition: Benign Addition: A: (8, x, x, x) A: (8, x, x, x) B: (8, 5, 5, 5) B: (9, 5, 5, 5) Thus, according to Parfit and countless others, any specification of AU will fail, and thus AU fails as a theory. Thus, the bottom up objection has forced many philosophers to abandon both AU, and AU’s motivating intuition—the intuition that more happy people, all else equal, is not better. 181 This example doesn’t depict the individuals “who will live in the distant future” as being very far into the future— it doesn’t depict a string of values coming after the fourth value and before the fifth. This, however, is just for the sake of simplicity of presentation, and the point is the same. Also, note that the third individual is slightly better off in the cases where the child is had. This is supposed to represent the happiness a parent gains by having the child. A similar point, however, would apply even if the parent was not depicted as being made better off by having the child. 182 Parfit (1984, p. 419). 183 Huemer (2008). 229 1.B.3. While the bottom up argument found in the literature is offered (in the manner described above) against ATLU, the bottom up argument actually is a more general argument that can be offered against any version of AU. As an example, I will briefly address what the bottom up argument looks like when it is offered against AMU. The conversion of the bottom up argument against ATLU into the bottom up argument against AMU consists of just one simple alteration: In the ATLU version of the bottom up argument, the numerical values described in the examples above represent values of individuals’ entire lives; in the AMU version of the bottom up argument, however, the numerical values described in similar examples instead represent values of conscious moments. Thus, for example, instead of Egyptology referring to the values of lives of individuals who existed in the past, Egyptology in the context of AMU refers to the values of conscious moments that existed in the past. Aside from this difference, however, the bottom up argument against ATLU and the bottom up argument against AMU are the same. Not only are the two arguments the same aside from this difference, but it seems, at least prima facie, that the bottom up argument has both the same type of force and also the same extent of force in both contexts. (These two claims will, however, be further explored, and, in fact, denied, below, in Section 1.G.) Furthermore, the bottom up argument can be offered not only against ATLU and AMU, but also against any other version of AU, regardless of what the version of AU maintains is the unit relevant for ethical inquiry. Thus, the core of the bottom up objection against AU is that it attempts to show that, regardless of what one maintains is the unit relevant for ethical inquiry, each possible reference class of data to include in the averaging function results in implications that are implausible. Thus, according to the bottom up objection, AU fails, because—for any 230 version of AU, defined by its choice of the unit relevant for ethical inquiry—any specification of the appropriate reference class for the AU theory in question will provide implausible implications. For the ease of presentation, in what follows, unless I say otherwise, I will be describing the version of the bottom up objection that is offered against ATLU. The reader should understand that while my points, as stated, are about the bottom up argument against ATLU, there are analogous points to be made about the bottom up arguments against AMU. In most cases, it will be left to the reader to make the relevant substitutions to arrive at the points about the bottom up argument against AMU. However, in some cases, I will explicitly address how the points I make apply to the bottom up argument against AMU. Further, in Section 1.G, I will specifically compare and contrast the relative strength and plausibility of the bottom up arguments against ATLU and AMU, respectively, and of the responses that can be offered by ATLU and AMU, respectively. Section 1.C. Important preliminary points regarding axiology and action In this chapter, I will ultimately be arguing that a temporally neutral account of AU is plausible, and thus that the bottom up argument fails. Before diving into my account, however, there is an important preliminary question that must be faced. The first key task in attempting to reply to the bottom up argument against AU is to clarify what I take the argument to be about. After all, utilitarianism can be described as a family of theories. Among other things, it can be described as an axiological theory and it can also be described as a theory of moral action. Thus, what type of theory is the average utilitarianism that the bottom up argument seeks to reject? Is it an argument about axiology, an argument about action, or an argument about both? 231 I think that it is clear that it is an argument about both. It seems to me that the bottom up objection is taking average utilitarianism to be a theory that maintains (1) that the best state of affairs is the one that maximizes average happiness, (2) the best action is the action that brings about the best state of affairs, and thus (3) the best action is the action that brings about the state of affairs that maximizes happiness. In light of this, it’s important to be clear about a strategy of response that I could take that would not be very satisfying. What I could do would be to state that AU is an axiological theory, and that, as an axiological theory, it is able to successfully avoid the bottom up argument. It provides a temporally neutral AU axiological account, and this is it. This response then distances itself from the claim about action, and it does not take a stand on whether the AU theorist maintains that the best action is the action that brings about the temporally neutral state of affairs that maximizes average happiness. I could take this approach, but it seems to me that this is too simple of an approach. The case for a temporally neutral axiology is too easy, and the case for a temporally neutral AU axiology, while less easy, is not one that really is threatened by the bottom up argument. The bottom argument gets its force by how it seemingly shows that there is no reference class of data to be averaged over that is a plausible goal for one’s action. Further, while I doubt that the proponents of the bottom up objection would choose to espouse an AU axiology, it seems that it would not be overly surprising—and it certainly would not be inconsistent—for them to espouse an AU axiology. Their claim then simply would be that AU also is a theory about action, and while perhaps it can offer a plausible axiology, it is unable to offer a plausible account when the axiology is combined with the seemingly plausible utilitarian claim that the best action is the one that brings about the best state of affairs. Although these proponents do not state that AU has a plausible axiology (and although they also do not explicitly say that their claim that AU is 232 implausible amounts to an axiological claim), in my interpretation, it is the claim about action that is the core of the bottom up argument. In light of this, this chapter will proceed by lumping together axiology and action. In other words, for the purposes of this chapter, I will assume that the utilitarian is committed to saying that the best action is the action that brings about the best state of affairs. (If one wants to, one can blunt this claim by adding an “all else equal,” or some similar allowance for certain constraints, or some clause that makes clear that this is if the only facts we have are about happiness levels. These types of concessions are fine with me.) This is choosing the tougher route, but I think that it is the only route worth taking. In light of this, I will mean for axiological claims I make to imply the analogous claim about action, and also vice versa. I will strive to speak in consistent terms (most frequently in terms of action), but if I stray, this is not my intent. There is another way of describing the situation here regarding the bottom up argument and whether it is an argument against AU as an axiological theory or AU as a theory of action. A way to think of the bottom up argument is that it raises doubts about the conjunction of (1) AU as an axiological theory, and (2) consequentialism. Parfit and the others who offer the bottom up argument indeed seem to be focusing on action, and they thus seem to be assuming consequentialism. (Thus, it’s not exactly that the proponents of the bottom up argument are “lumping together” consequentialism and the axiological theory of AU, or that they are “lumping together” the axiological theory of AU and AU the action theory. Instead, it seems that they are just assuming consequentialism.) In light of this, the force of the bottom up argument could be taken in one of two ways. On the one hand, if one assumes consequentialism, it could be taken as an argument the axiological theory of AU (and here the thought would be that a doubt is raised about the axiological theory of AU that is not raised about the axiological theory of TU). On the other hand, if one assumes the axiological theory of AU, it could be taken as an argument against 233 consequentialism (and here the thought would be that AU raises a doubt about consequentialism that TU does not). Either way, though, what seems clear is that the bottom up argument’s force is steeped primarily in action and it attempts to show that the AU theory of action is not plausible. Parfit and the others who make the bottom up argument do in fact rely on action intuitions in making their bottom up argument. Thus, it does not seem to me that it is satisfactory to reply by simply arguing for a temporally neutral axiological theory of AU and leaving aside the assumption of consequentialism or the result of the conjunction of consequentialism and the axiological theory of AU. Thus, for the purposes of this chapter, I will take the more difficult approach in responding to the bottom up argument: Rather than just defend the axiological theory of AU, I will assume consequentialism during my defense of the axiological theory of AU—and, thus, in so doing, I will be defending AU as a theory of action against the bottom up argument. One other note is in order before proceeding: Parfit and the other authors who make bottom up arguments use temporal terms such as “past,” “present,” and “future,” as well as other similar terms. In my discussion, and following their lead, I do so too. What’s important to note, though, is that temporal terms such as these do not make sense if we are talking solely in terms of axiology. After all, whether an event is in the past, present, or future is a function of what time an action in question is occurring or what time some type of a determination is being made. This doesn’t mean that temporal statements can’t be made in descriptions of axiology—and indeed they can because times and relative times are a feature of axiology as well—, it’s just certain types of temporal statements will not make sense if one is talking about axiology. Thus, it’s important to note that when terms like these are used, if one wants to understand the statement in terms of axiology, one might then need to substitute some of the terms that are used in the statement for other terms that are compatible with speaking in terms of axiology. 234 1.D. My main response to the Bottom Up argument: Identifying the Relevant Population as the temporally neutral population, and an explanation of how person-affecting intuitions are to blame for the prima facie plausibility of the bottom up argument 1.D.1. Argument overview One of the main difficulties with AU that Parfit addresses, as discussed in Section 1.B, is the fact that it seems to be difficult to come up with a plausible account of the relevant temporal population for which to maximize average happiness. Ultimately, all of the different populations considered seemingly fail and this is what leads to the conclusion that AU fails as a theory— forcing many philosophers to abandon the intuition that more happy people, all else equal, is not better. In this section (Section 1.D), after reminding the reader about person-affecting intuitions and the difficulties associated with them, I will begin by redefining the challenge for AU as a need to articulate not a plausible temporal account, but rather, a plausible account of whether or not “unaffected individuals” 184 should be included in the relevant population. Next, I explain that the question of which population is relevant is distinct from the question of how to assess what is best for that population—as such, TU is confronted with the question as well. I suggest that an impersonal morality should both be temporally neutral and include unaffected individuals. I then take up Benign Addition, Reverse Egyptology, and Egyptology, and I argue that the fact that people find these objections persuasive is not evidence that TU is more plausible than AU, because their persuasive power is due to our person-affecting intuitions and the fact that person- affecting intuitions align with TU in these cases. I argue that there are just as many cases where person-affecting intuitions align with AU and thus, if we are to compare AU and TU, we should do so impersonally. Once we do so, the anti-more-is-better intuition comes into play, without 184 This is a technical term that I will introduce and define below. 235 much opposition, and it is difficult to reject AU. 1.D.2. A reminder regarding the problems with a population ethics theories that are either person- affecting or based on person-affecting intuitions 1.D.2.a. Part I Chapter 2 discussed person-affecting moral theories, at length, and the full discussion will not be replicated here, but in light of the fact that they will prove to play a large role later on in this chapter’s discussion, a brief refresher regarding person-affecting moral theories and the difficulties that afflict them will be useful. 1.D.2.b. Despite the initial plausibility of what Parfit calls “person-affecting moralities,” the vast majority of the population ethics literature rejects even utilitarian person-affecting theories and instead focuses on “impersonal” utilitarian theories. According to Parfit, a “person-affecting” morality is a moral theory according to which the part of morality concerned with human well- being “should be explained entirely in terms of what would be good or bad for those people whom our acts affect.” 185 As Parfit and many others argue, however, person-affecting moralities must be rejected because they are unable to solve the non-identity problem. 186 The non-identity problem is illustrated by Parfit’s “14-Year-Old Girl” case: A young girl has a child, who, due to her mother’s youth, has a life that is difficult, yet worth living. If, however, the girl had waited a number of years to have a child, she would have had a different child, but one that would have had a better life than that of her actual child. The problem is that it seems clear that it would be better if the girl waited before having a child, but we cannot explain this by saying the girl’s decision was worse 185 Parfit (1984, p. 394). 186 Parfit (1984, p. 359). 236 for her child, because if the girl had waited, a different child would have been born. 187 Thus, despite its initial plausibility and strong attempts to defend it, 188 a person-affecting morality seems ill-equipped to handle questions of population ethics. 189 Instead, and for good reason, Parfit and others turn to impersonal moralities—moralities that are not limited to a consideration of what is good or bad for those affected by an act—with the hopes of constructing a plausible account. Both AU and TU are impersonal moralities, and it is these views (and similarly impersonal views) that have been the primary subject of this dissertation. It’s important to note, however, that, as I will discuss below, despite their problems, person-affecting intuitions do serve an important heuristic function. Furthermore, there are other ways in which we must take person-affecting intuitions seriously in our theorizing—albeit not in an intrinsic manner, but, rather, in a merely instrumental manner. 1.D.3. Redefining the Challenge for AU With this refresher about person-affecting intuitions and their difficulties complete, I now jump into the meat of my response to the bottom up objection. While Parfit and seemingly everyone else who discusses the question of the relevant population for AU frame their discussions in terms of temporal populations, this is not the best way to address the issue. Rather, the better question is which temporal population of individuals unaffected by the action in question should be included in the relevant population, in addition to those, in any temporal population, who are either affected by or brought into or out of existence by the action in question. (Hereafter I will refer to the group of those who are either affected by or 187 Some would deny that this case gives reason to reject person-affecting moralities, arguing that the second child would be worse off in the state of affairs where the girl does not wait to have a child. However, it seems odd to say that this child is harmed by not being brought into existence, and maintaining this position would require one to accept additional implausible positions as well. For further discussion, see Broome (2004). 188 See Roberts (2004). 189 See Holtug (2004) for a discussion of the difficulty—in fact, the impossibility—of constructing a plausible person- affecting morality. 237 brought into or out of existence by the action in question as individuals that are “not-unaffected.”) Everyone seems to think that not-unaffected individuals should be included in the relevant population for AU, regardless of when they exist. Parfit’s own discussion of strategies of conservation and depletion, 190 among others, indicates that there doesn’t seem to be anything counterintuitive about including future individuals’ happiness in the relevant population if they are not-unaffected. What the relevant populations addressed by Mere Addition, Reverse Egyptology, and Egyptology have in common is that they include a group of not-unaffected individuals. They differ with respect to the particular temporal population of unaffected individuals that they include: Mere Addition includes unaffected individuals in the present, Reverse Egyptology includes unaffected individuals in the present and in the future, and Egyptology includes unaffected individuals in the present, future, and past. Thus, while Parfit and others are right to suggest that the alleged lack of appeal of AU is due to the inability to articulate a plausible temporal population for which to calculate average happiness, what people actually find unappealing about the various populations considered seems to be just that they include individuals who are unaffected by the action in question. Further, what is allegedly unappealing about including unaffected individuals in the relevant population seems to be equally the case for all of the unaffected individuals, regardless of whether they exist in the past, present, or future. Thus, while, above, I stated that a better question than asking which temporal population is relevant is to ask which temporal population of unaffected individuals is relevant, the best question seems to be whether the relevant population should include only not-unaffected individuals or whether it should include both not-unaffected individuals and unaffected individuals. If any unaffected individuals should be excluded, there 190 Parfit (1984), p.361. 238 seems to be no reason to include those from other temporal populations, and vice versa. While perhaps the past seems less able to be affected than the future or the present, what is relevant is that it has been stipulated that the future and present individuals in question would be unaffected by the current action as well. Thus, the challenge still remains for AU, but the challenge seems to be that it must articulate whether the relevant population for which to maximize happiness includes both unaffected individuals and not-unaffected individuals, or whether it only includes not-unaffected individuals. 191 1.D.4. Both AU and TU Must Articulate the Relevant Population It is important to note something that seems to have gone somewhat ignored 192 in the population ethics literature that is critical of AU: This task of articulating the population relevant for ethical inquiry is not a question only for AU. It is also a question for TU—AU and TU are both impersonal moralities that seek to maximize an abstract value for a population. As such, the question of which population is relevant for ethical inquiry is a question that is distinct and separable from the question of how to assess what is best for that population. Just as an AU account is incomplete without a description of the relevant population, so too is a TU account. TU’s common prescription to “maximize happiness” or to “maximize happiness of the population in question” leaves undetermined which population’s happiness should be maximized. Of course it is no surprise that the task of articulating the relevant population is thought to 191 Note the fact that this past point (that the relevant differences appear to be between what is affected and what is not affected) should put us on notice about the fact that person-affecting intuitions might be underlying the bottom up objection. In light of this, we should not be surprised when it turns out that our intuitions are actually largely affected by person-affecting intuitions in the cases at issue here. 192 There certainly has been discussion about the relevance of future persons, merely possible persons, and persons that are contingent relative to the set of options, but the fact that TU must answer the question of which population is relevant does not seem to get mentioned in critiques of AU. Discussions of AU’s shortcomings do not make clear that the choice of the relevant population is a question that should be answered prior to the debate between AU and TU, and thus should be answered without being influenced by a thinker’s commitment to AU or TU. 239 be a task only for AU. The inclusion of various temporal populations of unaffected individuals will often affect AU’s prescriptions, but for TU, these unaffected individuals in any temporal population will just serve as a constant that does not affect TU’s prescriptions. Thus, while examples like Parfit’s case of depletion and conservation show that it matters to TU which not- unaffected individuals are included in a population, it seems as though TU does not confront a question of whether unaffected individuals are relevant. In the cases of Benign Addition, Egyptology, and Reverse Egyptology, no matter how many unaffected individuals are included in the relevant population, TU’s prescription about whether the individual in question should procreate will remain the same. Nevertheless, TU is an impersonal morality that seeks to maximize happiness for a population, and as such, it is left under-described if it does not articulate whether unaffected individuals are part of the relevant population for ethical inquiry. Even if different versions of TU provided identical prescriptions, it would be important to articulate which version one subscribed to. This notwithstanding, the fact that TU is confronted with the same task of selecting the relevant population as is AU does not mean that AU is off the hook. If TU were plausible for at least one population while AU were not plausible for any population, this would indicate that TU is a plausible theory and that AU is not. Nevertheless, the question of which population is relevant and the AU versus TU question are conceptually independent. 1.D.5. The Relevant Population Should Include Unaffected Individuals and Unaffected Values Up until this point I have used the term “unaffected individual” as an umbrella concept to refer to two separate cases that I will now distinguish: cases of what I will more properly call “unaffected individuals” and cases of what I will call “unaffected values.” I argue that both 240 unaffected individuals and unaffected values must be included in the relevant population, and given that the question of relevant population is distinct from the question of AU versus TU, this result will apply to both theories. A common thought might be that the relevant population for ethical inquiry should exclude individuals who exist in each of two states of affairs and who will be equally well off in each—i.e. “unaffected individuals.” However, the relevant population should not be restricted in this way. As a result of the difficulties exposed by Parfit’s two-child case, it was necessary to adopt an impersonal morality that would treat the two following cases alike: (1) A case of two different individuals, one in s1 and at value “3” and the other in s2 and at value “7” and (2) A case where an individual exists in s1 at value “3,” and where the same individual exists in s2 at value “7.” However, if we are going to ignore identity in cases like this, it seems as though an impersonal morality should also treat the following two cases in the same way: (1) A case of two different individuals, one in s1 and at value “5” and the other in s2 and at value “5” and (2) A case where an individual exists in s1 at value “5,” and where the same individual exists in s2 at value “5.” Thus, if an impersonal morality includes in the relevant population for ethical inquiry two different individuals with the same value (cases of what I call “unaffected value”), it should also include in the population one individual who is equally well off in both states of affairs (cases of what I call “unaffected individual”). Further, as I argue below, it seems as though an impersonal morality should include unaffected values in the relevant population, and thus, it should include unaffected individuals. Someone who thinks that unaffected values should not be included in the relevant population would think that the way to compare states of affairs would be to consider only the differences between their strings of values. For example, in comparing A and B, below, we might match up four of the “5”s in each state of affairs, thus finding that only the “x” and one “5” in A 241 and the “6” and the “2” in B are the population relevant for ethical inquiry. A: 5, 5, 5, 5, 5, x B: 5, 5, 5, 5, 6, 2 While perhaps plausible at first, consider the possibility that the option of bringing about state of affairs C arises, and the decision becomes whether to bring about A, B or C: C: 3, 7, 3, 7, 8, 1 C doesn’t share any values with A or B, and thus it is clear that all of the “5”s that had not been considered part of the relevant population now must be considered part of the relevant population. Thus, while on TU, the comparison between A and B seemingly could leave out the first four “5”s, this is not because these “5”s are not relevant to ethical inquiry, but it’s just that in this case they happen to not affect the result. 193 One might say, however, that unaffected values are relative to all available options. However, even if we can say that unaffected values are relative to all available options, this shows that the notion of an unaffected value is not a fundamental part of an impersonal morality. In order to know which values are unaffected, we have to be told what the consequences of the available actions are, and only from this can we deduce which values are affected and which not. Thus, it would be wrong to say that we have some prior notion of which values are affected. Rather, this is a derived judgment from something more fundamental—the raw data about all of the values themselves. The broader point here is that the causal consequence of a choice about how to act is the entire resulting state of affairs, and it is this that must be compared to the causal consequence of a different choice—also an entire resulting state of affairs. Further, as discussed above, the two possible causal consequences of a choice are the same, regardless of the identities of the 193 In other words, while the following is not the case for AU, TU satisfies what one might call a monotonicity condition, according to which population X is better than population Y if and only if the combined population X & Z is better than Y & Z. However, this property of TU notwithstanding, as just argued, even unaffected values (e.g. population Z) are still part of the population relevant to ethical inquiry. 242 individuals and how directly or indirectly the agent brings about the various components of the resulting states of affairs. Thus, an impersonal morality should be temporally neutral and should include both unaffected individuals and not-unaffected individuals. 194 1.D.6. The Person-affecting-intuitions explanation for why Benign Addition, Reverse Egyptology, and Egyptology all miss the mark Now that we understand better what TU and AU entail, we are better equipped to see why Benign Addition, Reverse Egyptology, and Egyptology miss the mark. I will argue that the fact that people find these objections persuasive is not evidence that TU is more plausible than AU, because their persuasive power is due to our person-affecting intuitions and the fact that person- affecting intuitions align with TU in these cases. To illustrate this, I show that versions of these examples that are slightly altered to produce alignment of person-affecting intuitions with AU result in AU likely seeming more attractive. Thus, I suggest that the result is a wash and that TU and AU must be compared in a manner divorced from the complicating person-affecting intuitions. 195 1.D.6.a. Benign Addition Parfit and Huemer both argue that Benign Addition and the related cases give us reason to think that TU is more plausible than AU. How, they ask, can it be worse for there to be additional 194 I’m open to the idea of including other conscious creatures as well, but this would not change the structure of the theory. I discuss non-human creatures in Section 1.H. 195 I owe thanks to an anonymous reviewer for making the following point and for stressing its importance: Although I refer throughout this paper to ”person-affecting intuitions,” it is important to recognize the differences between various types of person-affecting intuitions. Among the different possible formulations are the following two. First, x is better than y, if (x is better than y for some actual person and worse for no one). Second, x is better than y, only if x is better than y for some actual person. Unless stated otherwise in what follows, I will intend a third formulation—the conjunction of the first and second formulations. It’s helpful to notice when making comparisons, though, which conjunct of the conjunction is operative in making a particular claim. In most cases, however, it is the second formulation that is doing the work. 243 happy people (Mere Addition), and how can it still be worse to add additional happy people when the original person is made better off (Benign Addition)? To avoid confusion, in what follows I will discuss Benign Addition, but similar points apply to Mere Addition. Benign Addition: Benign Addition Variation: A: (8, x, x, x) A: (8, x, x, x) B: (9, 5, 5, 5) C: (5, 5, 5, 9) The objection says that in Benign Addition, B is better than A, and the idea is that this is because of TU. My thought is that this has nothing to do with TU. Rather, I think this is likely because of the person-affecting intuition, which here suggests that B is better. One way to test this might be to consider Benign Addition Variation. Would it be better to reduce the welfare of an existing individual to bring into existence three additional happy individuals? AU and TU are both impersonal moralities that seek to maximize average and total happiness, respectively, for the relevant population, and we know that all of the values must be included in the population. Further, the fact that whether to include unaffected individuals in the relevant population does not alter TU’s prescriptions seems to have masked the fact that TU is an impersonal morality and not a person-affecting morality. In the above cases, AU and TU will have the same prescriptions in comparisons between A and B and between A and C—with AU saying A is better in both cases and TU saying it is worse in both cases. While a proponent of TU will maintain that C is better than A, I think that many people who prefer B to A and who think that this is because they find TU plausible 196 will in fact find that they prefer A to C in Benign Addition Variation. If so, it seems as though the alleged force of the Benign Addition example is not a function of intuitions in favor of TU, but rather a function of our 196 It’s important to note here (and throughout Section 1.D when I make claims analogous to this) that I am not saying that a committed proponent of TU would find A preferable to C. This is certainly not the case since TU holds that C is preferable to A. Those about whom I’m speaking are those that find B preferable to A and who then conclude that what must be influencing them and what must be doing the work in an explanation of their intuitions is a sympathy for TU. 244 person-affecting intuitions—and the same could be said about the force of Benign Addition Variation if it were used to show that AU were more plausible than TU. In Benign Addition, person-affecting intuitions align with TU, and in Benign Addition Variation they align with AU. 1.D.6.b. Reverse Egyptology The same thing at work in Benign Addition seems to be at work in Reverse Egyptology as well: Person-affecting intuitions align with TU in Reverse Egyptology, but in similar cases that TU and AU would treat in the same way as Reverse Egyptology, person-affecting intuitions align with AU. Recall the Reverse Egyptology example: Possible Future 1: 1A: Have the child: 5, 5, 6, 6 | 18 AU = 8, TU = 40 1B: Don’t have the child: 5, 5, 5, x | 18 AU = 8.25, TU = 33 Possible Future 2: 2A: Have the child: 5, 5, 6, 6 | 1 AU = 4.6, TU = 23 2B: Don’t have the child: 5, 5, 5, x | 1 AU = 4, TU = 16 The objection aims to draw out our intuitions that, regardless of which of the two possible futures obtains, the same choice should be made (it is better to have the child). This is presented as an objection to AU and as evidence in favor of TU, because TU gives us this prescription, whereas AU does not. AU says that we should have the child if Future 2 will obtain, but that we should not have the child if Future 1 obtains. However, as with Benign Addition, I think that if we have the intuitions that Reverse Egyptology suggests we do, this is not because of TU, but rather because of person-affecting intuitions, which here suggest that having the child is better in both cases. One way to test this might be to consider the following cases: Case 3: 3A (rearranged values for 1A, above): 5, 5, 6, 6 | 18 AU = 8, TU = 40 3B (rearranged values for 1B, above): 5, 5, 5,18 | x AU = 8.25, TU = 33 Case 4: 4A (rearranged values for 2A, above): 5, 5, 6, 1 | 6 AU = 4.6, TU = 23 4B (rearranged values for 2B, above): 5, 5, 5, 1 | x AU = 4, TU = 16 245 In these cases, two individuals are deciding whether to create a zygote that will be frozen for one thousand years before developing into a child. As the tables show, since there are corresponding values in 1A and 3A, 1B and 3B, 2A and 4A, and in 2B and 4B, AU and TU will have the same prescriptions in this new example. However, I think that many people who claim to prefer TU would find that in case 3 they prefer 3B and that in case 4 they prefer 4A. While 4A is the option TU espouses, 3B is not. Both 3B and 4A, however, are espoused by AU. Similarly, in these two cases, the person-affecting intuitions suggest that 3B and 4A are better. Thus, as with Benign Addition, it seems as though common intuitions in Reverse Egyptology might be due to the fact that, in that example, TU’s prescriptions align with the prescriptions of person-affecting intuitions. 197 1.D.6.c. Egyptology So as not to repeat myself, I will not explicitly apply the above analysis to the Egyptology case and the Egyptology variations, depicted below, but the same points apply. Happy Egypt: 1A: Have the child: 18 | 5, 5, 6, 6 AU = 8, TU = 40 1B: Don’t have the child: 18 | 5, 5, 5, x AU = 8.25, TU = 33 Unhappy Egypt: 2A: Have the child: 1 | 5, 5, 6, 6 AU = 4.6, TU = 23 2B: Don’t have the child: 1 | 5, 5, 5, x AU = 4, TU = 16 Case 3: 3A (rearranged values for 1A, above): 18 | 5, 5, 6, 6 AU = 8, TU = 40 3B (rearranged values for 1B, above): x | 5, 5, 5,18 AU = 8.25, TU = 33 Case 4: 4A (rearranged values for 2A, above): 6 | 5, 5, 6, 1 AU = 4.6, TU = 23 4B (rearranged values for 2B, above): x | 5, 5, 5, 1 AU = 4, TU = 16 While perhaps it might seem odd to some to include the past in the population relevant for 197 In the versions of Reverse Egyptology and Egyptology discussed here, a parent experiences a gain in happiness as a result of having the child in question. However, even if the version analyzed were one in which a parent did not experience a gain in happiness, the analysis would be similar, and the conclusion the same: Our intuitions in these cases are not dictated by AU or TU, but by person-affecting intuitions. 246 ethical inquiry, recall that I have already argued that those who are unaffected by an action seem to be similarly situated regardless of when they exist temporally. Further, I also argued that all unaffected values should be included in the population relevant for ethical inquiry for an impersonal morality. Even if one accepts this, however, one might still think that there is a difficulty with these examples, because we seemingly would never be confronted with a choice between 3A and 3B or a choice between 4A and 4B. Both seem to suggest that we can change the past. However, since only the values matter for an impersonal morality, it was not crucial that the “18,” “x,” “6,” and “x” in the past in Cases 3 and 4 were put in the past—the examples would have sufficiently illustrated the same point about the Egyptology objection even if they were put after the “ | ”. Further, even if these values were kept before the “ | ” (and thus in the past), the example maintains its force, because the point being made here is an axiological one—when is x better than y? This question can be asked even if one is unable to make a choice that will bring about x or y, due to x or y being in the past or for other reasons. Not only is one able to make betterness assessments of this sort, but one can also wish that things in the past were or were not the case. For example, one can wish that it was the case that the ancient Egyptians lived fulfilling lives. Further, and very interestingly, these types of wishes or hopes are not limited to comparisons between possible past events. Parfit’s discussion in Part II of Reasons and Persons illustrates this nicely: First, Parfit plausibly argues (via his Hospital Example and its variations 198 ) that when it comes to oneself at a particular time, one will prefer that it be the case that one’s life’s remainder contains as little suffering as possible, even if this means that his life will have had a much larger quantity of past suffering, and thus more total suffering during one’s lifetime. Despite this, Parfit argues, when it comes to individuals other than oneself—be they near or distant—one wishes foremost not that it be the case that they have the 198 Parfit (1984), p.165. 247 least amount of suffering remaining, but that their entire life (past and future both included) have as little total suffering as possible. What this shows is that, when it comes to thinking about individuals other than oneself, we often hope that future suffering be greater, even if it means that past suffering was less by a greater amount. This provides strong additional support for the claim that the past is and should be included in the population relevant for ethical inquiry. These points should be sufficient, but there is also an explanation of how these points regarding Egyptology can be understood not only as being axiological points, but also as points that apply to the context of decision-making. Since it still might be difficult for some to accept that past values are relevant for ethical inquiry, I will provide an example to give further intuitive support for this claim—an example which will also show that sense can be made of Cases 3 and 4, even as they are written. This example will illustrate two points: First, a reader who is sympathetic to evidential decision theory (EDT) will take the example to show that choices between 3A and 3B and between 4A and 4B are choices that can confront us. According to EDT, even TU, in practice, might at times provide different prescriptions based on whether it takes a temporally neutral or current / forward-looking form, and the examples suggests that a proponent of EDT should espouse a temporally neutral morality. This illustrates the important point that any persuasive power that the Egyptology objection provides over and above the persuasive power of Reverse Egyptology is, at best, limited to impressing causal decision theorists. The example, however, is not introduced only for the eyes of proponents of EDT. Its implications are broader: Even if temporally neutral and current / forward-looking versions of TU never actually have diverging prescriptions, if they did, a temporally neutral version would be more plausible. As such, even in cases where the values of individuals in the past are “unaffected,” 248 they are as relevant and important as the unaffected values in the future. Although one might try to draw a distinction between the relevance and importance of happiness that will be experienced and happiness that was experienced, no such distinction can be made. Genetic Twin Example: You live in a small city, and there’s a possibility that you have a gene that causes two behaviors—neither of which would be expressed by many individuals without the gene: 1: In one’s thirties, one will work extraordinarily hard to bring about a 20% gain of happiness for each of one’s fellow citizens, and 2: In one’s fifties one will be destructive and bring about a 80% loss of happiness for each of one’s fellow citizens. You know that, a century ago, your genetic twin lived in a city with five times the population of yours. However, you do not know whether he had the particular bad gene, and thus whether he manifested these two behaviors. You do, however, know that he lived to age ninety. You, on the other hand, have been infected with a virus that will kill you in your forties, before you could become destructive to your citizens. What should you do? Should you bring about the 20% gain to your population before your death? Since bringing about the 20% gain to your population is strong evidence that your brother brought about a net 60% loss to his population, it seems as though your choice is between two states of affairs: Brother’s city 199 The present (your city) A: Help your citizens: (….10,10,10 | 6, 6, 6, 6, 6 | 10, 10, 10 | 12…..) B: Don’t help your citizens: (….10,10,10 | 10,10,10,10,10 | 10, 10, 10 | 10…..) Presented with this situation (and assuming EDT), a current / forward-looking version of TU would provide a different prescription than would a temporally neutral version of TU. The current / forward-looking version would prescribe helping one’s citizens, but the temporally neutral version would prescribe not helping one’s citizens: According to EDT, what’s relevant is the probability of what occurred in the past conditional on your doing an act. Conditional on your not helping your citizens it’s probable that your brother did not have the gene and that his citizens had values of ten. Conditional on your helping your citizens it’s probable that your brother did have the gene and that his citizens had values of six. So according to EDT and a temporally neutral morality, one should not help one’s citizens. It seems plausible to me that it would be better to not help one’s citizens, and thus to bring about state of affairs B. (Of course, while some find this plausible, others find this claim implausible, and thus view this as a reductio against EDT. People’s intuitions on this are far from unanimous.) 199 The numbers here don't capture the fact that your brother’s city was so much larger than yours. But it indeed is the intent of the example for your brother’s city to be much larger than yours. 249 Wherever one’s intuitions lie on this question, this example should provide further intuitive support for the claim that past individuals should be included in the population relevant for ethical inquiry—even to a causal decision theorist who denies that a temporally neutral and current / forward-looking TU would, in practice, ever diverge. This is for the following reason: Just as we can hope for future individuals to be happy, so too can we hope that past individuals were happy. Thus, even in cases unlike the above example, cases where current assessments place identical strings of values for past people in each of the states of affairs being compared, these values should still be taken into consideration by an impersonal morality—be it AU or TU. Once the scope is widened beyond not-unaffected individuals, there do not seem to be grounds to exclude from the relevant population any individuals, past, present, or future. As argued earlier in Section 1.D, the fact that unaffected individuals in the past cannot be affected does not distinguish them from individuals in the present or future whom, it has been stipulated, are unaffected by an act. Thus, in sum, and regardless of whether one focuses on these points merely as axiological points or whether one also views these points through the lens of decision-making, once the scope is widened beyond not-unaffected individuals, there do not seem to be grounds to exclude from the relevant population any individuals, past, present, or future. 200 For additional explanation of why inclusion of the past is not as implausible as it might initially seem, see the discussion below, in Section 1.E.1, about heuristics. 1.D.7. Conclusions regarding my main response, and some possible objections 1.D.7.a. Conclusions regarding my main response While Benign Addition, Reverse Egyptology, and Egyptology are examples of cases where 200 Below, in Section 1.F.1, I will consider and respond to two additional objections to my temporally neutral account of AU (objections which could also be considered to be objections to AU, more generally). One of these objections is Arrhenius’s “Sadistic Conclusion” objection, and the other is a related objection raised by an anonymous reviewer of an article of mine. 250 commonly held person-affecting intuitions align with TU, I argue that there are just as many cases where they align with AU. Given that these person-affecting intuitions affect our assessments in both directions, if we want to compare AU and TU, we should describe worlds in a manner divorced from these complicating person-affecting intuitions. We should compare worlds without information of whether an individual in one world will exist in the other, and of which value he would have if he did. This information is not relevant to the comparison between AU and TU—we should only consider the strings of values. Thus, if we are to come to a conclusion about whether AU or TU is more plausible, I think we need to consider this impersonal case. The consideration of this impersonal case might lead some to favor TU and some to favor AU. But this is where the anti-more-is-better intuition (see Part I Chapter 2) comes into play, and in my opinion, without much opposition. Seemingly everyone who rejects AU—in the context of positive utility—rejects it because of absurdities it allegedly entails in cases where AU conflicts with person-affecting intuitions. Once AU and TU are compared impersonally, as they should be, AU’s treatment of cases of positive utility is difficult to reject. 201 1.D.7.b. Some possible objections At this point, a few objections might be made. Objection #1: One might think that to the extent TU coincides more with person-affecting intuitions than 201 It should be noted that the answers that I give to the bottom up objection against AU do not (fully) speak to the intuitions that many people believe are elicited by Parfit’s case of Adam and Eve, which I raised in Part I Chapter 2 and will discuss in depth in Section 1.F.2. The Adam and Eve case is a case where person-affecting intuitions do not seem to fully infiltrate one’s intuitions in the way that I identify in the examples discussed in the section on the bottom up objection. Further, even if we consider an impersonally described version of the Adam and Eve case, people seemingly have the intuition that aligns with TU. Thus, this does seem to be evidence that there is at least somewhat of a counterweight, even in the impersonal case, to the anti-more-is-better intuition—and furthermore, that sometimes (as is the case for many people in the Adam and Eve case), the counterweight will outweigh the anti-more-is-better intuition. I discuss the Adam and Eve objection later in this chapter, in Section 1.F.2. 251 does AU, this is reason to prefer TU to AU. Response #1: While this might be true if it were the case that TU coincided more with person-affecting morality than does AU, the sets of cases described above are typical: Cases where person-affecting intuitions coincide with TU can be made to coincide with AU by rearranging the slots, and vice versa. As such, there doesn’t seem to be greater overlap between person-affecting intuitions and TU than there is between person-affecting intuitions and AU. 202 Objection #2: One might object to my conclusions and say that the real takeaway here is not about the relative merits of AU and TU, but rather, that a person-affecting morality is superior to both. Relatedly, one might say that we shouldn’t be abstracting from situations to exclude seemingly relevant facts from our analysis. Response #2: While this type of an objection to my main account might have some prima facie appeal, 202 As I said, person-affecting intuitions correspond equally frequently with AU as they do with TU. In light of this, it is moot whether or not it would be the case that person-affecting intuitions aligning more frequently with TU than with AU would provide (1) support for TU as compared to AU, or even (2) support for TU tout court. Further, if one is persuaded that person-affecting moralities must be rejected as candidates for Theory X (for the reasons I gave in Part I Chapter 2—i.e. because of their inability to solve the non-identity problem and thus provide a tractable account of population ethics), but if one finds them particularly tempting, perhaps one might argue that the only criterion for choosing an impersonal theory (be it AU or TU) is that it be the impersonal theory that most closely approximates person-affecting intuitions. This type of a pseudo-hybrid position, offered in the previous sentence, seems not only extreme, but also terribly implausible for various reasons, but questions (1) and (2), above, still are interesting ones. With respect to these two questions, however, I think that the answer is “no” to both, and that to provide any support at all for AU or TU, there would have to be a consideration other than it more frequently aligning with person- affecting intuitions. It seems to me that greater alignment of either theory with person-affecting intuitions would, all else equal, simply be happenstance, because it has been shown that neither AU nor TU is fully compatible with person-affecting intuitions. (Of course, the matter would be different if AU or TU aligned completely with one of the impersonal theories and they thus were coextensive. If this were the case, then I think that the impersonal theory that corresponded perfectly with the person-affecting morality would thereby be provided with strong support. If that were the case, then the impersonal theory would be, for practical purposes, the same as the person-affecting theory, and both would thus have one’s full support if one fully endorsed person-affecting moralities. However, as discussed at length, neither AU nor TU is coextensive with person-affecting intuitions, and thus the hypothetical situation discussed here does not apply.) Notwithstanding this entire foregoing discussion, however, we must still remember that the questions in this footnote are, for the purpose of the dissertation, moot, because, as I have argued, neither AU nor TU aligns more frequently with person-affecting intuitions than does the other. 252 the difficulties that make a person-affecting morality intractable as a theory of population ethics— discussed at length in Part I Chapter 2, and revisited briefly earlier in this chapter—are so great that this is not a good option. Despite the fact that we do have person-affecting intuitions in many cases, a person-affecting morality is not a viable candidate for Theory X. While this response (in conjunction with my discussions that show that person-affecting moralities are not viable options) is sufficient, there is more that that I can and will say about person-affecting intuitions that will make this an easier pill to swallow. The following section (Section 1.E) provides a number of additional considerations that provide further (intuitive) support of my main account. The first of these additional considerations will be an explanation of why we might have person-affecting intuitions, and why they might be useful even if they themselves do not constitute a viable Theory X. In short, that section will provide an account of how person-affecting intuitions are heuristics that help simplify our thinking regarding goodness- assessments and regarding decisions, and how these heuristics are consistent with, and in fact, further the goals of an impersonal Theory X. 1.E. Additional Considerations that provide further (intuitive) support for my main explanation While I believe that my main account (which I offered in Section 1.D) provides a response on behalf of the AU theorist that is sufficient to defray the bottom up argument against AU, it may be the case that some readers still have lingering intuitions to the contrary. Within this group of readers who have lingering intuitions to the contrary, it seems as though there might be two camps—on the one hand, those whose intuitions to the contrary are strong enough that they do not accept my account, and, on the other hand, those whose intuitions to the contrary are less robust and thus who do accept my account, yet do so only despite some lingering degree of internal turmoil. 253 This section (Section 1.E) aims to speak to readers in both of these camps. While I maintain that my main account is a sufficient response to the bottom up argument, I aim, in this section, to provide further intuitive support for my point. While I attempt to do so in a few ways, I do so primarily by attempting to provide various explanations of what might lead readers to have intuitions contrary to my position. My hope is that the points in this section will thus enable the reader to recognize that any lingering intuitions that they have that are contrary to my position are a function of the mechanisms at work that I describe in this section (or mechanisms that are similar to those that I describe), and that these lingering intuitions are thus not evidence of weakness of my position. Thus, my hope is that readers in both of the camps that I refer to in the previous paragraph will be persuaded by my arguments in this section—meaning that those who currently reject my account will find that they accept my account, and that those who currently accept my account, but with difficulty, will find that they are able to do so with less difficulty. In Section 1.E.1, I explain how person-affecting intuitions are heuristics that further betterness assessments and prescriptions that are in accordance with those made by impersonal moralities. In Section 1.E.2, I explain that there is additional uncertainty regarding hedonic data that afflicts AU in particular (and, especially, a temporally neutral AU), and I show (1) how AU’s seeming impracticality is mistaken for implausibility, and (2) how hedonic data regarding the unaffected and unknown are mistaken for being irrelevant. In Section 1.E.3, I explain how intuitions to the contrary of the assessments and prescriptions of a temporally neutral AU are likely a result of the infiltration of intuitions arising as a result of IMF (the “Intrapersonal Maximizing Fact”), which was introduced in Part II Chapter 2. In Section 1.E.4, I argue that the seeming implausibilities of a temporally neutral AU account are actually features of impersonal moralities more broadly, and thus are not attributable to AU or to a temporally neutral version of AU. I also provide an additional argument, in Section 1.E.4, for the claim that any reference class 254 other than a temporally neutral reference class would be arbitrary and thus implausible. 1.E.1. Person-affecting intuitions as heuristics that further betterness assessments and prescriptions that are in accordance with those made by impersonal moralities 1.E.1.a. The fact that we have some person-affecting intuitions is not necessarily evidence that a person-affecting morality is correct, and the reason for this is that there are possible explanations, even within an impersonal morality, for why we might have person-affecting intuitions. In this section, I offer an account of this sort. Specifically, I argue that the fact that we have person- affecting intuitions can be explained by the fact that, in most situations that confront us, these intuitions are helpful heuristics that simplify our thinking (regarding goodness assessments and decision making). More specifically still, and this is what’s crucial, I argue that person-affecting intuitions are heuristics that help us with thinking that helps us further the goals of impersonal moralities. The reason that person-affecting moralities are helpful heuristics for us in furthering the goals of impersonal moralities is that the vast majority of decisions (or, similarly, goodness assessments) we make in life are examples of decisions in same numbers cases. Further, not only are most of our decisions in same numbers cases, but most of our decisions also involve states of affairs involving different outcomes for the very same individual (or individuals). It is these features of our decisions that make it the case that person-affecting moralities are helpful heuristics for us in furthering the goals of impersonal moralities. While what’s key is that these are same numbers cases, there are two versions of these same numbers cases, and, since the person-affecting-intuition heuristics operate slightly different in the these two versions, I will discuss them separately. One type of same numbers case is one in which the actor (or assessor) knows exactly which individual (or individuals) is (or are) affected 255 by the action in question. In the other type of same numbers case, the actor might know some of the individuals who will be affected, but he does not know whether there will be others who are affected, and, if so, how these other individuals will be affected. Two preliminary points are in order before continuing: First, there is not a bright line that separates these two cases. After all, it seems that we could never or almost never be in a position to know that there won’t be other individuals who are affected by an action (albeit potentially very indirectly and potentially years in the future). Thus, the two cases are not as distinct as I suggest they are. Notwithstanding this, we can make sense of there being cases that diverge along the dimension of whether or not we know which individuals are affected. Second, I’m assuming that despite not knowing all of the effects of one’s action in the second type of case, we do still know, or at least assume, that it is a same numbers case. Of course, it could be that many if not most of our decisions do end up having effects that result in a case being a different numbers case. Notwithstanding this, however, I will consider these cases to be same numbers cases, because those whom we know we affect, even in the second type of case, are of the same number in either possible state of affairs being considered. While a comparison between the two states of affairs might be a different numbers case if we had all the information that I’m stipulating we do not have, I’m leaving aside this possibility because it could affect things in either direction and there is no reason to think that it would have a net effect on the examples I consider. 1.E.1.b. Same numbers cases where we know who is affected In decisions of this sort, a person-affecting morality will coincide perfectly with both AU and TU, and, importantly, focusing merely on what is good or bad for the individual (or 256 individuals) affected will provide a shortcut for decision-making. In determining which action or state of affairs is best overall, one needs merely to decide which state of affairs or action is best for the individual (or individuals) who are affected by the potential action. Using this shortcut will be just as accurate as carrying out a greater exploration into hedonic facts of the world, and it will have the benefits both of taking much less time and effort to carry out, and also of lowering the risks of making calculating errors by drastically shrinking the relevant data set. Thus, if one acts in accordance with person-affecting intuitions in same numbers cases where we know who is affected, one will also be acting in accordance with AU and TU. Further, this is the case for any reference class of AU or TU as long as the affected individuals are included. In other words, acting in accordance with person-affecting intuitions in these cases would also be acting in accordance with, for example, all of the following: a current / forward- looking AU, a temporally neutral AU, a current / forward-looking TU, and a temporally neutral TU. 1.E.1.c. Same numbers cases where we know some who are affected but where we do not know if there are other individuals who are affected Person-affecting intuition heuristics will operate somewhat differently in same numbers cases where we know who is affected than they will operate in same numbers cases where we know some who are affected but where we do not know if there are other individuals who are affected. Despite this, however, person-affecting intuition heuristics in the latter case will still also aim to further the goals of impersonal moralities. The difference between how the person-affecting intuitions work as heuristics in the two cases is as follows: In the same numbers cases where we know who is affected, discussed above, acting in accordance with person-affecting intuitions would also constitute acting in accordance with AU 257 and TU (with one’s reference class of choice, subject to the minimal constraints above). In other words, acting in accordance with person-affecting intuitions would definitely be acting in accordance with AU and TU (with one’s reference class of choice, subject to the minimal constraints above). In the same numbers cases where we know some who are affected but where we do not know if there are other individuals who are affected, however, it is not true that acting in accordance with person-affecting intuitions would definitely be acting in accordance with AU and TU. Rather, in these cases, it would merely be the case that acting in accordance with person- affecting intuitions would probabilistically, or all else equal, be the action that is in accordance with AU and TU. While using these person-affecting intuitions as heuristics in these cases does not definitely bring about the action that is in accordance with AU and TU, this does not, however, mean that the heuristics are not useful. The fact that the heuristic brings about the action in accordance with AU and TU at least with all else equal is itself an important accomplishment. Given one’s limited knowledge about the welfare of the spatiotemporally distant, what is good according to a person-affecting morality, all else equal, will also be what is good on AU or TU: Not only will the treatment of individuals whom we know are affected serve as a proxy for the current-and-future population that includes both affected individuals and unaffected individuals, but this current-and-future population will serve as a proxy for a temporally neutral population including both affected individuals and unaffected individuals—and, thus, the treatment of affected individuals will serve as a proxy for this temporally neutral population (of affected and unaffected individuals) as well. Or, said differently, since we seemingly are unable to affect the past, the best state of affairs for a current / forward-looking reference class will be what is the best state of affairs for a temporally neutral reference class. Further, the person-affecting intuition heuristics in same numbers cases where we know 258 some who are affected but where we do not know if there are other individuals who are affected have similar benefits to the same heuristics when used in the first type of same numbers case: Using the shortcut will take much less time and effort to carry out than a full exploration into the hedonic facts of the world. Further, since a non-omniscient individual likely would be unable to ever obtain all hedonic facts, spatiotemporally, of the worlds one is choosing between, some sort of a heuristic will seemingly need to be used to make any estimate at all about goodness assessments or any estimate about the relative goodness of different decisions. Thus, it is this limited knowledge about the hedonic facts of the world that makes it the case that person-affecting intuitions are useful heuristics even in same numbers cases where we know some who are affected but where we do not know if there are other individuals who are affected. 1.E.1.d. Population ethics and different number cases: Cases where person-affecting intuition heuristics do not work in the way that we would hope they do While same numbers cases constitute the vast majority of decisions that we confront on a daily basis, there are, of course, different numbers cases that we could confront and that we do in fact confront, and population ethics questions frequently involve different numbers cases. It is these types of cases where person-affecting intuitions do not provide the value, as heuristics that further the goals of impersonal moralities, that they do in same numbers cases. As we’ve seen in the examples in section III, the assessments of both AU and TU can easily diverge from the assessments of person-affecting intuitions. Further, there does not seem to be any evidence that acting in accordance is even “all else equal” likely to further the goals of AU or TU. As we have seen, whether person-affecting intuitions align with AU or with TU in a different numbers case—even where all the hedonic facts are known—is a function of details of the examples about personhood that are unrelated to the goodness of the states of affairs (as 259 assessed by AU or TU). This is why the person-affecting intuitions, if used as heuristics in different numbers cases, do not makes us more likely to get the result in accordance with the assessments of AU or of TU. Thus, in questions of population ethics, the person-affecting heuristic does not provide the value that it might in everyday cases. 1.E.1.e. An additional point An additional note regarding person-affecting intuitions is in order. As with all heuristics, there are cases where the shortcut gets the wrong result, and it is these cases that have been under the microscope in this chapter. Even in these cases, though, where person-affecting intuitions seem to diverge in their prescriptions from AU, and thus provide the wrong answer to moral problems, an AU theorist—or even a TU theorist, for that matter—should not ignore person- affecting intuitions. This is the case because even if the person-affecting intuitions are mistaken, the fact that we have a “taste” for person-affecting intuitions needs to be taken into account when carrying out calculations for AU or TU. In other words, it will sometimes be best—according to AU or TU—to act in accordance with a person-affecting principle because it is instrumentally valuable. 203 If people hold person-affecting intuitions, it might cause real disvalue to act in contravention of these intuitions, and avoiding this disvalue can, at times, be what’s best according to either AU or TU. 204 However, while this is an important consideration, it is not one that played 203 And these person-affecting intuitions will have particularly strong force in the short term, in decisions that affect people we know and otherwise in the present. As such, they will be more instrumentally relevant in the present. And they will likely be less instrumentally relevant as we address populations that are more distant temporally (though, while perhaps less so than temporally, it seems that spatial distance will also often, as a practical matter, reduce the instrumental relevance of person-affecting intuitions). Similarly, the flip side of this coin is that when we are considering spatiotemporally distant populations, we will be less enticed by person-affecting intuitions, and it will likely be easier and more natural for us to think in terms of impersonal moralities. 204 It’s also worth noting and or remembering that throughout the dissertation (and, in particular, in Part II Chapter 4), there are many discussions of theories and heuristics that are structured similarly to the discussion in this subsection. 260 into the examples given in section III, and this is because all of the hedonic facts in those examples were stipulated. Thus, while perhaps tastes for person-affecting intuitions could have been at play in these examples, this would already have been taken into consideration in the formulation of the numerical values in the charts, and they would not be additional considerations to be considered after the numerical values had been determined. 1.E.1.f. Summary In sum, I think that all of my foregoing discussion provides a plausible account of how person-affecting intuitions operate as heuristics that, on the whole, further the goals of impersonal moralities. In light of this, I do not think that the fact that we have some person-affecting intuitions necessarily constitutes evidence that a person-affecting morality is correct and that an impersonal morality is incorrect. Thus, it seems that the discussion about heuristics helps support my point for why we should stay within the impersonal framework, and, further, why the reader should accept the main account that I offer, in section III, as a reply (on behalf of the AU theorist) against the bottom up argument against AU. In this subsection, I argue that person-affecting intuitions are heuristics that further goals of impersonal moralities. In other words, they are shortcuts that in most cases—and, in fact, in almost all cases that we do in fact confront in life—do bring about results that are coextensive with impersonal moralities, and thus which further impersonal moralities. I say, however, that, as with all heuristics, heuristics sometimes get the wrong result, and, in these cases, we must remember that heuristics are merely heuristics and that they should not be followed when they get the wrong result. Instead, the actual theory that the heuristics further is the theory that should be followed in these cases. However, this does not mean that the person-affecting intuitions should be completely ignored, however, because to the extent that they constitute actual tastes, these tastes must be taken at face-value—albeit for merely instrumental purposes. In other parts of the dissertation, I argue that various other theories function in ways analogous to the way in which person-affecting intuitions have been shown to function with respect to impersonal moralities in this section. For example, I argue that ATLU can and should be treated similarly with respect to AMU. Further still, I argue that deontological theories in normative ethics can and should be treated similarly with respect to utilitarianism. In all of these cases, as with person-affecting intuitions with respect to impersonal moralities, the former is a heuristic that generally furthers the latter. Furthermore, there is less divergence between prescriptions between the two theories within a pair than one might think prima facie, in light of the fact that tastes for the heuristic do have to be taken into consideration (instrumentally). In some cases, however, divergences remain, and, in these cases, the latter of the two theories in the pair is the one that should be followed. 261 1.E.2. The additional uncertainty regarding hedonic data that afflicts AU in particular, and especially a temporally neutral AU: How AU’s seeming impracticality is mistaken for implausibility and how the unaffected and unknown are mistaken for being irrelevant 1.E.2.a. Introduction As was the case in Section 1.E.1, this section (Section 1.E.2) attempts to speak to readers who have lingering intuitions contrary to various aspects of the main account I offered in Section 1.D. While I maintain that that account is a sufficient response to the bottom up argument, I aim, in this section, to provide further intuitive support for my point. This section, like the last one, attempts to do so by providing an explanation of what might lead readers to have intuitions contrary to my position. While the points in this section certainly will not constitute an attempt to fully explain people’s difficulty accepting a temporally neutral version of AU, I think that the points here do illuminate two forms of confusion that likely contribute, at least somewhat, to people rejecting a temporally neutral version of AU. Furthermore, as I have already argued, the temporally neutral version of AU is the most plausible version of AU, and thus, a rejection of the temporally neutral version of AU is likely to render the bottom up argument against AU successful. To the extent that these points are plausible, this section contributes somewhat to supporting my main account (from Section 1.D)—and thus to disarming the bottom up objection against AU. The two separate confusions that are illuminated in this section are the following: First of all, compared to the TU theorist (and also compared to the person-affecting morality theorist), the AU theorist is terribly clueless (about data) in different numbers cases. Furthermore, the temporally neutral AU theorist is especially clueless. What this means is that it is very hard to implement AU, and it is especially hard to implement a temporally neutral version of AU. Many people might find that this feature of a temporally neutral AU (and AU in general) makes the theory less plausible. I do not think that this makes the theory less plausible, however, and I think 262 that the perhaps-common belief (be it held explicitly, implicitly, or even unconsciously) that it does make the theory less plausible explains at least part of people’s resistance to a temporally neutral AU. Second of all, we rarely have much data about unaffected individuals, and we particularly rarely have much data on spatiotemporally distant unaffected individuals. In light of this, this information (which we do not have) is unable to be used in our calculations. Since we lack this data, we need to make estimates based both on what we do know and based on heuristics (be they person-affecting heuristics or other heuristics). This lack of use of data regarding the unaffected (and especially the unaffected in the past) might easily make it seem as though this data is irrelevant. Despite this, however, this data is not irrelevant. It’s just that it’s hard, practically speaking, to get this information. Thus, this confusion of the unaffected (or unknown) and the irrelevant is what mistakenly leads people to think that a temporally neutral reference class, and, also, relatedly, AU (which requires data regarding the unaffected to make a prescription), are implausible features of a Theory X. 1.E.2.b. Over the years, an argument against consequentialism, called the Epistemic Argument against consequentialism, has received a fair amount of attention in the normative ethics literature. The argument has been made by a number of authors, including Shelly Kagan. Kagan’s account of the argument goes as follows: Consequentialism is wrong because it is impossible to know the future. Therefore we will never be “absolutely certain as to what all the consequences of your act will be.” Acts that seem likely to lead to good results can and often do lead to disastrous results, and vice versa. There will be causal ramifications deep into the future, and most of these results will be both unforeseen and unforeseeable. Thus, “How can we tell which act will lead to the best 263 results overall—counting all the results?” And, “If it is impossible to tell whether any act is morally right or wrong, how can consequentialism possibly be a correct moral theory?” 205 While many, including Kagan, do not think that this argument poses a serious threat to consequentialism, James Lenman has recently argued that it does. Whether or not the Epistemic Argument threatens consequentialism is a matter of debate, but one thing that is not a matter of debate is the fact that consequentialism is data-driven, and we as agents—or as assessors of the goodness of possible states of affairs—are generally drastically under-informed about all of the data that would be relevant to our assessments of states of affairs. After all, we do not know what the future holds. Thus, even if we can still meaningfully carry out relative goodness assessments among different total state of affairs, it certainly seems that we must be aware that our assessments are generally highly probabilistic and that our goodness assessments, while perhaps the best possible estimates, should be offered without necessarily articulating that there is high confidence in the goodness assessment turning out to be accurate. While this vast degree of uncertainty in our calculation affects consequentialism on the whole, it is interesting to note that, in addition to this form of uncertainty, there are other forms of uncertainty that afflict consequentialism, and these forms of uncertainty do not equally afflict different versions of consequentialism. For example, in addition to the uncertainty due to the Kagan-Lenman argument that afflicts both AU and TU, there is a respect in which AU confronts significantly more uncertainty than does TU—and this is in different numbers cases, the cases that make up the majority of the challenging cases in population ethics. (For reasons discussed in the previous section on heuristics (1.E.1), the uncertainty asymmetry between TU and AU that I am about to discuss is limited to different numbers cases. Thus for the purpose of the coming discussion, assume that we are discussing only different numbers cases—the bread and butter of 205 Shelly Kagan, Normative Ethics (Boulder: Westview Press, 1998), p. 64. 264 population ethics.) In assessing the relative goodness of two states of affairs that could result from an action (in different numbers cases), TU needs only compare the individuals that are affected by the action (though of course the Epistemic Argument points out that this itself is wrought with difficulty). AU, however, needs not only data about the individuals that are affected, but also the data on any unaffected individuals that are included in one’s reference class (assuming that unaffected individuals are in fact included in the reference class). For the TU theorist, this additional data is a constant that is unnecessary for making the relative goodness assessment. The AU theorist, however, has a tougher time using his theory, because there is more data that is necessary for the calculation, and thus, due to huge gaps in our data set, AU is afflicted with a vastly greater amount of uncertainty than is TU. As discussed in Section 1.D.7, AU theorists (and, similarly TU theorists, to the extent they confront uncertainty) are not utterly handcuffed—there are ways to use heuristics to arrive at relative goodness assessments, even if there is not a high confidence level associated with their accuracy. Notwithstanding the fact that AU is not utterly handcuffed, however, it seems to certainly be a strike against 206 AU that—at least in this respect 207 —it is even more difficult to apply successfully than TU. 206 While this might be a strike against the theory in some ways, it’s not clear that difficulty applying the theory is a strike against the plausibility of the theory. Furthermore, this dissertation is concerned primarily with what constitutes the most plausible Theory X—and this dissertation does not shed nearly as much ink on questions related to applying the most plausible Theory X. 207 There are ways in which TU might be thought to be more difficult to apply. One way in which this is thought to be the case is that TU has a challenge of determining where the “zero point” is located. I myself, as I argue in Part IV, do not think that this specific problem is indeed a way in which TU is more difficult to apply, because, in my view, zero points do not even exist. However, as I discuss in Part IV Chapter 1, the common discussion about zero points and the difficulty that they pose for TU is not that zero points might not exist (as I argue is the case), but, rather, that there is a challenging problem of identifying where the zero point is located (even though, on the view of these authors, the zero point indeed exists). If it were the case, as these authors think, that a zero point exists but is difficult to locate, this would constitute a substantial way in which TU is difficult to apply. AU, however, would not have a difficulty applying its theory due to the uncertainty of where the zero point is located, because, for AU to operate, unlike for TU, no zero point is required. There might also be other examples of ways in which it can be argued that TU has features that make it difficult to apply in ways that do not afflict AU. 265 Interestingly, while there is a disparity between TU and AU with respect to how much data is needed to make a relative goodness assessment, there is also a disparity of this type among versions of AU that employ more and less comprehensive temporal reference classes. For example, a temporally neutral version of AU would require more data to carry out a relative goodness assessment than would a current / forward-looking version of AU. Thus, in light of the uncertainty that comes with our lack of data, it seems that there would be even more uncertainty associated with an assessment made by a temporally neutral version of AU than by a current / forward-looking version of AU, and thus it would be even more difficult to use a temporally neutral version of AU—in other words, the confidence level one would have in one’s assessment would be even lower with a temporally neutral version of AU than with a current / forward- looking version of AU. Similarly, one could compare other versions of AU in this respect based on how narrow or broad their reference class of choice is. While the uncertainty that confronts these various types of theories (in increasing order of the uncertainty that afflicts them: starting with TU, which experiences just Kagan-Lenman type uncertainty, then circumscribed reference class versions of AU, and then a temporally neutral version of AU) is massive, the uncertainty we have is even more massive than it is if we focus merely on human beings on our planet. The following topic will be elaborated on below, in Section 1.H, but it’s important to briefly point out here the extent of our lack of knowledge. A utilitarian will want to include all sentient life in a goodness assessment, and this will include both non-human creatures on our planet and any sentient life that exists elsewhere in the universe. We have no idea how much life exists elsewhere in the universe, nor do we have any way of knowing how happy it is. If there are lots of planets full of beings capable of much more happiness than we are, than perhaps any additional individuals or conscious moments on our planet would make for a worse state of affairs, whereas if we’re among the beings who are capable of the most happiness, 266 then the opposite would hold true. Similarly, while perhaps we have some sense of how much sentient life exists on our planet, and perhaps we can begin to make some guesses about the happiness levels of these other creatures—based on their biological similarities and differences from us—even these estimates about numbers of creatures and these estimates of their happiness levels may be wildly off. Furthermore, a temporally neutral temporal account would multiply these uncertainties—perhaps ad infinitum—according to the expansive continuum of time, both forwards and backwards. 1.E.2.c. With the vast amount of uncertainty that exists having been illustrated, the upshot is as follows. I think it’s possible that the issue of uncertainty has gotten tangled up with the issue of irrelevance. Furthermore, I think that heuristics, discussed in Section 1.E.1, are playing a key role in the process. As discussed in section 1.E.1, despite great uncertainty—of all of the types discussed above—we are forced to make goodness assessments and decisions about actions with the limited data we have. To do so, we use various heuristics, among the most helpful of which is the person- affecting heuristic. Furthermore, with respect to AU, we also employ other heuristics and proxies, including the heuristic of a circumscribed reference to act as a proxy for a more comprehensive one—for example a reference class of affected people being a proxy for those that are affected and unaffected in the present (but for whom we have data), or this second reference class as a proxy for all current / forward-looking individuals, (including those who are affected, those who are unaffected and for whom we have data, and also those who are unaffected and those for whom we have no data), and this third reference class as a proxy for the temporally neutral reference class. While these heuristics likely successfully provide us with helpful guidance, I think that our 267 need to rely on them has transformed our view of the data themselves about which there is such great uncertainty and on which we cannot rely when making our assessments. While the unaffected—be the data about them known or unknown—are data on which we do not focus, because of our limitations, we have come to see this data as irrelevant, as opposed to merely unknowable or indecipherable with our limited capacities. This, I think, is partly responsible for our viewing the unaffected as irrelevant—and particularly the unaffected, the data of which are unknown, and even more particularly, the unaffected, the data of which are unknown and who exist distantly, be it spatially or be it temporally (be it in the past or future). In light of this, it is difficult for people to view people other than the affected as being part of the relevant reference class, and this, I think, is what, in part, leads to the prima facie plausibility of the bottom up objection. With respect to AU, we seemingly get the wrong result if we focus only on those who are affected by an action, but, due to the above, we don’t view others as being relevant. With TU, however, no such problem arises. Furthermore, prima facie, not only does this make AU seem implausible because its various possible reference classes that capture more than just the affected seem mistaken, but the temporally neutral reference class for AU is particularly difficult for many to accept because we know even less and are even more in the dark about the data of those in the distant past. What we must do, then, is realize that uncertainty is not the same as irrelevance, and we must realize this even when, in order to function, we use heuristics all the time that treat the uncertain as though they are irrelevant. Once we see this, it is clear how we might have come to view the uncertain as irrelevant. This dialectic is akin to that where the utilitarian maintains that a deontologist’s rules are merely tools or heuristics that help us maximize utility in the face of uncertainty. According to the utilitarian, even if we think that we are best off never violating any of these rules, this is not because the rules are primitive and that happiness is not the only good, 268 and that the happiness brought about by following the rules is only a collateral benefit—which could not grow to a large enough sum to outweigh the violation of the rule, and thus which would in a sense be akin to being irrelevant (as used in my discussion)—but rather, we should not violate the rules because of the uncertainty of the happiness that would arise from breaking them, and thus our past evidence that following the rules is in fact optimific. Thus, while the use of heuristics that I have described generally leads to the ability to make assessments more easily, it might be distorting our understanding of the individuals who are unaffected and of the individuals whose data is unknown. This distortion does not affect the prima facie plausibility of a TU account, but it does affect the prima facie plausibility of AU, and, even more so, it especially affects the prima facie plausibility of a temporally neutral version of AU. 1.E.2.d. Conclusion It thus seems that there are two separate upshots for us here in section 1.E.2 (though, despite their separate importance, one of them did help us arrive at the second one). First of all, compared to the TU theorist (and also compared to the person-affecting morality theorist), the AU theorist is terribly clueless in different numbers cases. Furthermore, the temporally neutral AU theorist is especially clueless. What this means is that it is very hard to implement AU, and it is especially hard to implement a temporally neutral version of AU. Many people might find that this feature of a temporally neutral AU (and AU in general) makes the theory less plausible. I do not think that this makes the theory less plausible, however, and I think that the perhaps-common belief (be it held explicitly, implicitly, or even unconsciously) that it does make the theory less plausible explains at least part of people’s resistance to a temporally neutral AU. Second of all, we rarely have much data about unaffected individuals, and we particularly 269 rarely have much data on spatiotemporally distant unaffected individuals. In light of this, this information (which we do not have) is unable to be used in our calculations. Since we lack this data, we need to make estimates based both on what we do know and based on heuristics (be they person-affecting heuristics or other heuristics). This lack of use of data regarding the unaffected (and especially the unaffected in the past) might easily make it seem as though this data is irrelevant. Despite this, however, this data is not irrelevant. It’s just that it’s hard, practically speaking, to get this information. Thus, this confusion of the unaffected (or unknown) and the irrelevant is what mistakenly leads people to think that a temporally neutral reference class, and, also, relatedly, AU (which requires data regarding the unaffected to make a prescription), are implausible features of a Theory X. 1.E.3. The IMF (Intrapersonal Maximizing Fact) infiltration explanation Despite my arguments for a temporally neutral AU, many readers might still find it particularly counterintuitive that past individuals (or past conscious moments) are relevant for decisions, even if and when these values are unaffected by actions in the present. In addition to the points that I have already made, though, I think that there is another big reason for why we might find it difficult to espouse a temporally neutral reference class and reject a current / forward- looking reference class. I think that the point that I will make in this coming section gets to the core of many tensions that afflict our intuitions in various sub-areas of population ethics. I think that many of the tensions in population ethics can be traced to two views that are in some tension with one another: The first is the anti-more-is-better intuition when it comes to society at large (and the intuition that we should maximize average when it comes to society at large)—regardless of whether the unit relevant for ethical inquiry is the person or the conscious moment. Second, as discussed in depth in Part II of this dissertation, when it comes to oneself, a 270 person is generally constrained by IMF (the “intrapersonal maximizing fact)”: One inescapably seeks to maximize the amount of accrued happiness going forward—it follows an analogue of TU and with a current / forward-looking reference class. This is generally a fact about our psychology, regardless of whether one espouses AU-type or TU-type aggregation as the aggregation mechanism for the goodness of one’s life, and regardless of which temporal reference class one espouses for the determination of the goodness of one’s life. 208 However, the existence of IMF will incline many—both unreflectively, and even, perhaps, reflectively—to think that (1) the most plausible aggregation mechanism of goodness of a life is TU-type aggregation (and not AU-type aggregation), and (2) that the most plausible reference class is a current / forward-looking one (and not a temporally neutral one). Furthermore, even those who espouse AU-type aggregation as the aggregation mechanism for goodness of a life will likely feel at least some prima facie pull toward the plausibility of TU-type aggregation, and even those who espouse a temporally neutral reference class will likely feel at least some prima pull toward the plausibility of a current / forward-looking reference class. This is because even these individuals who espouse AU-type aggregation or a temporally neutral reference class (or who espouse both) will experience IMF. I think that many of the difficulties we encounter in the search for Theory X are due to the tension between these two intuitions. More specifically, I think that many of the difficulties we encounter in the search for Theory X are due to intuitions about one’s own life infiltrating assessments for society at large. I’ll show, below, how this occurs and why we have reason to reject intuitions related to intrapersonal maximizing when it comes to our search for Theory X. Consider again Parfit’s “My Past or Future Operations” example (discussed earlier in this chapter and, more at length, in Part II Chapters 2 and 3). I think that, when it comes to oneself, one will adopt a current / forward-looking view and prefer and hope that the suffering has already been 208 For an in-depth discussion of these concepts, refer to the discussions in Part II Chapters 2 and 3. 271 endured. While which implications and conclusions can be drawn from this can be disputed, what seems to be clear is that this is an instance of IMF at work. If, however, one considers what would be best agent-neutrally, and if one is told that the situation, instead of involving oneself, involves a spatially and or temporally distant individual, named Joe, whom one does not know, I think that the common intuition would be consistent with a temporally neutral view—and not with a current / forward-looking view. The fact that we would take a temporally neutral view about this agent- neutral assessment of the two possible states of affairs for Joe, I believe, is consistent with my claim that Theory X should be temporally neutral and also consistent with my arguments in past sections. The importance of re-introducing IMF and Parfit’s “My Past or Future Operations” example, though, is that I think it provides insight into our hesitance to adopt a temporally neutral Theory X when provided with cases that differ in certain ways from the above case of Joe. There are two different mechanisms by which this works. First of all, I think IMF can infiltrate our agent neutral assessments, and it can do so in the following way. As described in the discussion of the Parfit example, IMF brings it about that we often think that what is best for ourselves is to be assessed with a current / forward-looking reference class. Further, although we might not implement the current / forward-looking reference class when assessing what is best for someone who is spatially, temporally, and or emotionally distant, it often is the case that the more spatially, temporally, or emotionally close another person is to oneself, the more likely one is to put oneself in his shoes, and use for the assessment of what is best for him the same current / forward-looking reference class that one likely uses for oneself. Then, supposing that, we then are asked to assess the agent-neutral goodness of two possible states of affairs (one where the suffering of person X is in the past and is ten units, and the one where the suffering of person X is in the future and is five units—and where nothing else is different in the two states of affairs), one is then likely to use the Pareto principle to infer that the better state of 272 affairs is the one where the better result occurs for person X. Further, in cases where person X either is oneself or one who is spatially, temporally, and or emotionally close to oneself—and thus a person in whose shoes one puts oneself—one is likely to use IMF to conclude that what is best for the person in question is to have the suffering be in the past, and, when combining that intermediate conclusion with the Pareto principle, conclude that what is best agent neutrally is the state of affairs where the suffering is in the past—albeit a greater suffering. This is the most straightforward way in which this infiltration occurs. However, I think that our views about the first-person framework (and, similarly, our views about cases of spatial, temporal, and or emotional proximity, where the first-person perspective infiltrates) infiltrates even further. We are so used to addressing questions, situations, and problems that pertain to ourselves and those whom are close to us, that I think we frequently unconsciously extrapolate and employ the current / forward-looking reference class in most cases where we make agent-neutral assessments. Thus, the first-person framework infiltrates a large quantity of the cases that were addressed in the original exploration of temporal reference classes. We, as persons, trying to maximize the happiness going forward for ourselves, apply this framework to many of the decision-making choices provided in this chapter, and I think that this accounts for many of the intuitions contrary to those I’m arguing for. Furthermore, there is an additional way in which the first-person perspective’s current / forward-looking reference class infiltrates both our agent-relative and our agent-neutral goodness assessments. In just about every case where we make relative goodness assessments (and, assuming one subscribes to causal decision theory, certainly in all of the cases where we are making relative goodness assessment in order to inform practical decisions), we are making goodness assessments between states of affairs that include identical pasts. Thus, as a shortcut, we have grown accustomed to only considering the present and the future in our relative goodness 273 assessments. Thus, this is yet another way in which our first-person perspective’s current / forward-looking reference class has infiltrated our relative goodness assessments—be they agent- relative or agent-neutral. Further still, I think that we might even at times catch ourselves adopting a current / forward-looking reference for class for agent-neutral and / or agent-relative goodness assessments even if, upon greater reflection, we might intellectually espouse a temporally neutral view. Similar in various ways to the problem of free will (and to aspects of my discussion about personhood in Part II Chapter 3), we might intellectually adopt a temporally neutral reference class, but despite this, still be unable to escape the grasp upon our thinking of a current / forward-looking reference class. Thus, to hark back to Strawson’s “Freedom and Resentment,” 209 we might be capable, at times, to occupy an objective standpoint, but in many—if not most—cases we might not, and we exhibit reactive attitudes—the equivalent here being the first-person perspective’s current / forward-looking view. While the first-person perspective and its current / forward-looking reference class originates with IMF in cases regarding oneself, my hypothesis is that its reach extends to others, far beyond cases of merely oneself, and that its reach extends to questions of agent-neutral goodness, far beyond questions of agent-relative goodness. Importantly, however, there are two aspects of the first-person perspective that are extending to these other cases: First, it is the current / forward-looking reference class, and second there is the TU-type method of aggregation. With respect to the bottom up argument and my response to it, by far the main point of the current discussion is to focus on the infiltration into agent neutral assessments of the current / forward-looking reference class. The main point here is to provide an additional hypothesis to explain our apparently strong attachment to the current / 209 Strawson, “Freedom and Resentment.” 274 forward-looking reference class. Notwithstanding this, however, the current infiltration hypothesis also provides reason to think that TU-type aggregation—which is espoused by many for goodness assessments of their own lives first personally, and also, often, for goodness assessments of the lives of others—is infiltrating agent-neutral goodness assessments in a similar manner. While this is not the main point of this section, this provides additional strong support for the plausibility of AU. It provides an account of why people might mistakenly believe that TU-type aggregation is plausible for agent-neutral (i.e. population-wide) goodness assessments. And, further, it provides an account of why people might mistakenly believe that AU is implausible for any (i.e., all) of AU’s possible reference classes. All of the discussion in this section would not support my points, however, unless it were shown that the current / forward-looking reference (and, secondarily, TU-type aggregation)—be they plausible or not in the context of one’s agent-relative goodness assessments for oneself—are mistakenly being imported into the context of agent-neutral goodness assessments. As I showed, at length, in Part II Chapters 2 and 3, though, the reasoning chain that starts with IMF and that ends with claims about agent-relative and claims about gent-neutral goodness, is a flawed one. 1.E.4. Summary In sum, the foregoing three subsections provide three additional considerations that each, on its own, provides strong further support for the conclusions of my main argument in response to the bottom up argument (in Section 1.D)—and for the sub-arguments I make within that argument. 1.F. Two additional types of objections to AU In this section, I raise and address two additional types of objections to AU. First, in 275 Section 1.F.1, I address a pair of related objections that raise a similar concern with AU: One is an objection introduced by Arrhenius and entitled the “Sadistic Conclusion” objection. The second related objection was raised by an anonymous reviewer of an article mine, 210 and, since it does not have a name, I will refer to it as the “reviewer’s objection.” Next, in Section 1.F.2, I raise and discuss Parfit’s “Adam and Eve” objection. I raise all of these objections here in the chapter on the bottom up objection because they are all, to varying degrees, related to the bottom up objection. This is particularly the case for the Sadistic Conclusion objection and the reviewer’s objection— both of which fall squarely within the purview of the bottom up objection. (I nevertheless raise them both here in their own sections, rather than in earlier sections in this chapter, because it seems that a separate discussion of them will be fruitful.) The Adam and Eve objection, on the other hand, is only tangentially connected to the bottom up objection, but here is as good of a location as any to address it. 1.F.1 The Sadistic Conclusion objection and the reviewer’s objection, responses to these two objections, and discussions of related insights In this section, I discuss and respond to two related objections that can be raised against AU. One of these objections is Arrhenius’s “Sadistic Conclusion objection.” The other objection is the reviewer’s objection, and this objection raises a concern similar to the one raised by the Sadistic Conclusion objection. What these two objections have in common is that they both elicit the intuition that bringing about the best state of affairs according to a temporally neutral AU will bring about unpalatable results, and that, if we are to espouse AU, there is some reference class that is a subset of the temporally neutral population that would instead be a more plausible reference class over which to maximize average happiness. This allegedly more plausible subset is either a current / forward-looking reference class, or, alternatively, the reference class of those 210 Michael Pressman, “A Defense of Average Utilitarianism,” Utilitas 27 (4): 389-424 (2015). 276 who are affected by the decision (or action) in question. In what follows, I describe both of these objections and I explain why they do not successfully cast doubt on AU. 1.F.1.a The reviewer’s objection An additional concern that one might raise (and that an anonymous referee did raise) for a temporally neutral AU view is that it “implies variants of the Repugnant Conclusion.” The reviewer’s words are worth providing in full, and, thus, the reviewer’s objection is as follows: Suppose that, in the past, there existed 10 billion people with lives barely worth living. Each had a lifetime wellbeing of 1. We have two choices regarding who will exist in the future. In outcome A, another 10 billion people will exist, with only marginally better lives than the people in the past. Their lifetime wellbeing will be 2. In outcome B, 50 million people will exist, with an extremely high lifetime wellbeing of 100. This means that the average wellbeing is 1.5 in A, and 1.49 in B. AU therefore favors A. If a main intuition behind AU is ‘the idea that while we want people to be as happy as possible, we do not necessarily have reason to bring into existence additional happy people’ and that ‘when it comes to happy people, more is not better,’ the implication that A is better than B seems troubling. If a smaller and very happy total population is preferable to a larger and less happy one, why shouldn’t we have this preference also regarding future populations? This is a reasonable concern that is important to raise. The temporally neutral AU theorist does, however, have the following response: The average happiness must be considered for the whole, temporally neutral, state of affairs, and choosing to focus on some subset of the entire state of affairs will always enable one to describe cases such as the above. Consider a variation of the case offered, but where the world is just beginning, and both options A and B also include the creation of the original 10 billion people at wellbeing level 1. In this case, it seems that the AU intuition is that A is better than B. However, as I have argued in this chapter, the only difference between choice one and choice two is whether the unaffected (and unaffectable!) people are in the past or in the future. It seems that this difference is immaterial to which state of affairs is better, and thus it seems that the two cases should be treated alike. I have argued that the temporally neutral version is (perhaps counter to initial intuitions) the more plausible version. Thus, as attractive as the intuition in the reviewer’s objection may be, it seems that revising it is the least 277 difficult of various routes to reflective equilibrium. 1.F.1.b. The Sadistic Conclusion Objection The Sadistic Conclusion objection, introduced by Arrhenius, is an objection that can be raised against AU and against hybrid theories that have at least some features of AU. Arrhenius’s Sadistic Conclusion is as follows: “For any number of lives with any level of negative welfare, there are situations in which it would be better to add these lives to a population rather than to add some number of lives with positive welfare.” 211 According to the Sadistic Conclusion objection, this is an implication of AU, yet it is an unpalatable implication, and thus this provides a reason to reject AU. To see why an AU theorist might be committed to accepting the Sadistic Conclusion, consider the following: Suppose w1 contains ten people with a happiness level of ten. According to AU, the value of w1 is ten. Now, compare the results brought about by adding either one hundred people at level two versus adding one person at level negative five. If we add to w1 the one hundred people at level two, we will get a world, w2, the value of which, according to AU, is 2.73. If, however, we had to w1 the one person at level negative ten, we will get a world, w3, the value of which, according to AU, is 8.18. Thus, according to AU, this is a case where a world that is brought about by adding people with negative welfare is better than a world that is brought about by adding people with positive welfare. This is an implication of AU, but, according to the proponent of the Sadistic Conclusion objection, this is sadistic. Thus, the proponent of the objection argues that this is a reason to reject AU. My answer is as follows: The Sadistic Conclusion objection is really just a variant on the reviewer’s objection, 211 Arrhenius. 278 although it perhaps is a more specific (and narrower) version of the reviewer’s objection. Unlike the reviewer’s general objection, which was couched in terms of populations with positive happiness, the Sadistic Conclusion objection involves a population with negative happiness, and this arguably might give it stronger intuitive pull, thus making it arguably a stronger objection. What the two objections have in common is that they show that when adding two potential populations (p2 and p3, where p2 has a higher average happiness level than does p3 (and regardless of whether p2 and or p3 have positive or negative happiness values)) to a currently existing population p1 (this need not be a temporal addition; it could be modal), it can be the case that AU will prescribe adding p3, even though p3 has a lower average happiness level than p2. In the reviewer’s objection, p2 and p3 both have positive happiness levels. In the Sadistic Conclusion objection, p2 has a positive happiness level and p3 has a negative happiness level. The intuition underlying these objections is that if we care about AU, we should care about AU for circumscribed populations as well (or instead) of caring about AU temporally neutrally. Or, more specifically, the intuition perhaps is that we should care about the AU of those who are affected (e.g., those in the future, or the population that we are considering introducing). Even with the “future” subset (as opposed to (and potentially broader than) the “affected” subset), the intuitive idea behind the alleged importance of caring exclusively about (and thus carrying out the AU analysis over) the future seems to be steeped in the general idea that it’s the future that we can affect. Thus, as discussed earlier in this chapter, intuitions regarding the importance of the future do seem to be traceable to the notion that the future can be affected and things / people not in the future cannot be affected. This intuition seems to be that the future is somehow open in a way that the past is not, even if there are people in the future who aren’t affected, and this suggests that the common intuition is that there is a relevant distinction between the unaffected in the future and the unaffected in the past. But I have already explained above, in connection with the reviewer’s 279 objection, and at length throughout my main response to the bottom up objection (in Section 1.D) why we must espouse a temporally neutral AU (even if it might, prima facie, seem implausible). Thus, my response to the Sadistic Conclusion objection is simply to reinforce that I’ve already shown that we cannot pick and choose subsets of the full population to care about for the purposes of applying our Theory X (such as AU). Instead, we need to stick with temporally neutral account, and this is for the reasons that I’ve articulated earlier in this chapter. Additionally, various examples (some of which have already been discussed in this chapter) can be given that draw out the intuition that the temporally neutral account is what we must stick to. Thus, in sum, in answering the Sadistic Conclusion objection and the reviewer’s objection, I rely on the arguments and points that I have made earlier in this chapter. Of course, the Sadistic Conclusion objection does bring in the added element of suffering (negative happiness values), which are not present in the reviewer’s objection. Further, many readers will likely feel that the force of the Sadistic Conclusion objection is stronger than that of the reviewer’s objection as a result of the element of suffering being present in the Sadistic Conclusion objection but not in the reviewer’s objection. While I recognize that many readers might feel this way, and I myself do feel that the Sadistic Conclusion objection might have somewhat more force, I nevertheless think that the two objections are the same in relevant respects. I defer an explanation of why until Part III Chapter 2 (on the Two Hells objection), which is where I address head on the topic of suffering. That chapter will provide a number of strong reasons to reject the position that cases of suffering merit treatment that is any different from the treatment of positive happiness cases, and it will provide reasons to reject the view that cases of suffering provide any additional reasons to doubt the plausibility of AU. 212 I think that 212 Among the various reasons that I offer for why we should reject the position that cases of suffering merit treatment that is any different from the treatment of positive happiness cases, one of these reasons is that, as I argue in Part IV, “zero points” of value do not exist. If I am right about this, this seemingly would mean that there is no meaningful 280 there already is a sufficient response to the Sadistic Conclusion objection here, but, to the extent that one feels that the dimension of suffering renders the objection not fully warded off, I point the reader to the Two Hells chapter to provide considerable additional support for the response that I provide here to the Sadistic Conclusion objection. In sum, my arguments here and elsewhere in the dissertation show that neither the reviewer’s objection nor the Sadistic Conclusion objection successfully casts doubt on the plausibility of AU. Although these objections might have some prima facie intuitive pull, it seems that, upon closer examination, the intuitions that these objections elicit should, upon reflection, be rejected. 1.F.1.c. Causation and how what one causes is an entire state of affairs This chapter has argued for a temporally neutral reference class for one’s candidate of choice for Theory X. I myself espouse a version of AU, but I have argued that, even for TU, a temporally neutral reference class is what one should espouse and employ. To the extent that a reader has difficulty accepting this view of mine, I think that it might be less a result of difficulty being on board with a temporally neutral axiology, and more a result of difficulty being on board with the view that bringing about the best state of affairs temporally neutrally is what we should strive to do when we act. In my view, though, it indeed is the case that the best state of affairs according to a temporally neutral axiology is what is relevant for assessing actions. I think that perhaps the main obstacle I might have in trying to persuade a reader of this position of mine is a seemingly common belief about action: The idea that, loosely speaking, and with some exceptions, what someone causes when he acts are only the things that would have distinction between positive happiness and negative happiness. If this is the case, then there seemingly is no meaningful distinction between the reviewer’s objection and the Sadistic Conclusion objection. 281 been different had he not acted in the way that he did—or something somewhat like that. In my view, however, and as discussed in Section 1.D.5, above, the correct way to think about causation and action is that what one causes when one acts is the entire temporally state of affairs that results. This would include everything, including, for example, the fact that the planet Mars exists and is x number of miles from the earth on date y, and that World War II ended in 1945. Although it might seem counterintuitive to some, I think that the only workable account of causation is a temporally neutral account on which everything temporally neutrally is a cause of one’s act. While it might seem as though we only cause some smaller subset of things or events by carrying out a particular act, I think, as discussed in Section 1.D.5, that the determination of what these “affected,” or “caused” things are is the result of a process of comparing two possible entire states of affairs, and the determination of what is “affected” or “caused” is merely a derived determination. In order to know which things or events are unaffected, we have to be told what the consequences of the available actions are, and only from this can we deduce which values are affected and which not. Thus, it would be wrong to say that we have some prior notion of which values are affected. Rather, this is a derived judgment from something more fundamental—the raw data about all of the values themselves. My opponent would likely be quick to point out that we do, all the time, speak and think about causation in the more common way: We might speak about how (or think that) my making a shot in basketball caused my team to win the game, or caused me to be late for dinner, or caused my friend to be happy, but we do not speak about how (or think that) this shot caused Mars to exist and be x number of miles from earth on date y, or that it caused World War II to end in 1945. I agree that we do not speak or think this way, but I maintain that this is just because we typically operate with a shorthand (and a shorthand that, in most cases, is helpful). Thus, despite our typically speaking and thinking this way, I do still think that the correct way to speak and think 282 about action is to view the causal effects of an act to be the entire temporally neutral state of affairs that obtains as a result of the action having been carried out. In sum, the causal consequence of a choice about how to act is the entire resulting state of affairs, and it is this that must be compared to the causal consequence of a different choice—also an entire resulting state of affairs. My point here in this section is as follows: I think that if a reader does not accept my view that a temporally neutral reference class is one that we should espouse and employ, I think that this is less a result of difficulty being on board with a temporally neutral axiology, and more a result of difficulty being on board with the view that bringing about the best state of affairs temporally neutrally is what we should strive to do when we act. Further, to the extent that this is the case, I think that the main obstacle I face in getting a reader on board with my account is the fact that people commonly think that their actions only cause some small subset of an entire temporally neutral state of affairs. I think that it is this view that might be preventing some from espousing my view. If a reader comes around to the view that all actions cause entire temporally neutral states of affairs, then I think that my conclusions in the bottom up chapter will be much easier to accept. Because this dissertation is not about the metaphysics of causation, though, I will not argue for my view of causation any further (beyond what I have said here and in Section 1.5.D). Notwithstanding this, I do think it is useful to point out what my view is on causation, and also point out that I think that a reader’s espousing of a different account of causation could be what is preventing him from getting on board with my arguments and conclusions in this chapter regarding a temporally neutral reference class. 283 1.F.2. The Adam and Eve objection 1.F.2.a. The objection Parfit introduces the Adam and Eve objection as a reductio against AU. 213 It goes as follows: Suppose that Adam and Eve exist and have fantastic lives, say at a happiness level of 100. Suppose they can choose to either procreate or not procreate and that, if they procreate, their descendants—namely the human race—will also live fantastic lives, though their descendants lives will be ever so slightly less fantastic than the lives of Adam and Eve. Suppose that the lives of the descendants would be at a happiness level of 99. AU seems to be committed to the position that it would be better if Adam and Eve do not procreate, because doing so would lead to a state of affairs with lower average happiness than the state of affairs in which they do not procreate. According to Parfit and other proponents of the Adam and Eve objection, though, it seems that the state of affairs in which Adam and Eve procreate would be much much better than the one in which they do not. Even more strongly, though, it seems as though the state of affairs in which they do procreate certainly could not be worse than the state of affairs in which they procreate (or, to couch the objection in terms of action, it certainly could not be wrong for them to procreate when confronted with the choice between bringing about each of the two states of affairs just described. Thus, because AU seemingly is committed to a seemingly implausible position regarding the Adam and Eve example, this example is thought to constitute a strong objection against AU. 1.F.2.b. My analysis First of all, it’s important to be clear that, as I stated in a footnote in Section 1.D.7, the answers that I gave to the bottom up objection against AU do not speak to the intuitions that many 213 Parfit, Reasons and Persons, page 401. 284 people believe are elicited by Parfit’s case of Adam and Eve. 214 The Adam and Eve case is a case where person-affecting intuitions do not seem to infiltrate one’s intuitions in the way that I identify in the examples discussed in the section on the bottom up objection. Further, even if we consider an impersonally described version of the Adam and Eve case, people seemingly have the intuition that aligns with TU. Thus, this does seem to be evidence that there is at least somewhat of a counterweight, even in the impersonal case, to the anti-more-is-better intuition—and furthermore, that sometimes (as is the case for many people in the Adam and Eve case), the counterweight will outweigh the anti-more-is-better intuition. Another way of describing the dialectic here is as follows: The Adam and Eve objection elicits intuitions contrary to AU, but it does so in a way that is different from just about every single other objection to AU that I consider in this dissertation. The Adam and eve objection elicits intuitions that are contrary to AMBP (the (population-wide) anti-more-is-better intuition), and AMBP’s role in my arguments for AU was to play a large role in what I called the “prima facie case” for AU, which I offered in Part I Chapter 2. Just about all of the other objections to AU that I consider in the dissertation are not objections to the prima facie case for AU, but, rather, they are objections to AU that attempt to show why AU is implausible even if one is attracted to the prima facie case for AU. Thus, these other objections, for the most part, do not (directly) cast doubt on AMBP itself. What, though, is my response to the Adam and Eve objection? I have a number of different 214 A question might arise as to why I’m addressing the Adam and Eve objection here in the chapter on the bottom up argument if the Adam and Eve objection isn’t entirely related to the issues raised by the bottom up objection or to the issues I raise in my response to the bottom up argument. The answer is that (1) although the Adam and Eve objection is not entirely related to these issues, it does indeed still relate to these issues, and, as I explain below, I believe that my person-affecting intuition explanation to the bottom up objection actually fully disposes of one of the two intuitions that are elicited by the Adam and Eve objection, (2) there isn’t another part of the dissertation that involves issues that are more closely related to those in the Adam and Eve objection than does the bottom up chapter, and (3) the Adam and Eve example also shares the feature of the bottom up argument whereby both objections provide certain specific cases with numbers of people and their happiness values where betterness comparisons need to be made, and they both show purport to show that TU aligns with our intuitions in the case that is provided and that AU does not. 285 points to make in response. First: I do actually think that my main response to the bottom up argument plays a role here with the Adam and Eve objection as well—although only a partial role. According to AU, it is wrong for Adam and Eve to procreate and it is right for them to not procreate. According to TU, it is wrong to not procreate, and it is right to procreate. According to the intuitions that the objection seemingly attempts to elicit, it seems that one such intuition is definitely that it is not wrong to procreate. Arguably, a second intuition that it elicits is that it is better (i.e. right) if they do procreate (I’ll leave it vague regarding whether there is a moral requirement that they procreate or whether it’s just that it would be better if they do.) I think that person-affecting intuitions can explain the allegedly TU-friendly intuition that it is not wrong to procreate. As we would see if I were to draw up a chart similar to those I used in my main response to the bottom up argument (or as we just by thinking about the case), no person is worse off or harmed as a result of Adam and Eve procreating. Thus, according to person-affecting intuitions, it is not wrong to procreate. Thus, in this case, person-affecting intuitions align with TU and not with AU. As I argue in my main response to the bottom up argument, our common sense intuitions often track person-affecting intuitions, and, thus, the mere alignment of person-affecting intuitions with TU in a particular case does not provide any support for TU over AU. Rather, what our intuitions are a function of in the case is person-affecting intuitions, and it’s random whether the specific case involves an alignment of person-affecting intuitions with AU or with TU. Thus, in my view, in this case, this explanation of mine explains why we might think it is not wrong to procreate. My person-affecting intuition explanation, however, does not explain the possible common sense intuition in the Adam and Eve case that it is wrong to not procreate (or that it better, or right, to procreate). My person-affecting intuition explanation does not explain this specific intuition. And this specific intuition will be what I now will proceed to attempt to explain. I think it’s very 286 important to note, though, that I think that this intuition, which I will now seek to explain, is a considerably weaker intuition than is the one that I already explained. While I do not deny that people will feel the pull of the intuition that it is better for Adam and Eve to procreate, I think people would feel the other intuition much more strongly—that it is I not wrong to procreate. I think that of people who feel both intuitions, people will feel the second intuition less strongly, and I think that there will be some people who only feel the first intuition. And, I articulated in the previous paragraph what, in my view, is a persuasive explanation of the first intuition. But, since there certainly are people who will feel the pull of the second intuition (that it is wrong for Adam and Eve to not procreate, and that it is right / better if they do procreate), I now proceed to address this intuition. Next, I think it’s important to point out that I think that for most people who feel the pull of the intuition that it’s wrong for Adam and Eve to not procreate, I do not think that they would feel these anti-AMBP 215 intuitions across the board in most cases. For example, most people would not feel the anti-AMBP intuition in a case where the initial population (the one that would exist if the parents in question did not procreate) is sufficiently large (and not composed of only two people). I think that most people would only feel the anti-AMBP intuition when a population is sufficiently small. Of course, there will indeed be people who subscribe to the anti-AMBP intuition across the board: Indeed, a committed TU theorist would be someone who does. (I distinguish the “committed TU theorist” from a TU theorist who doesn’t wholeheartedly (if at all) feel the pull of the anti-AMBP intuition, but who nevertheless espouses TU because he perceives it to be the most plausible account of Theory X, all things considered. This latter person’s 215 Since AMBP stands for “the (population-wide) anti-more-is-better intuition,” it seems a little awkward to refer to an intuition as the “anti-anti-more-is-better-intuition.” Instead, one could perhaps just call the anti-AMBP intuition “the (population-wide) more-is-better intuition,” or the MBP intuition. First of all, I need not create a new acronym because I only will be discussing this intuition in this sub-section, but, secondly, this seems like an odd formulation because MBP are the three initials of my name (my middle name being Benjamin), and the MBP intuition is certainly not an intuition that I hold! 287 assessments would be consistent with the anti-AMBP intuition’s assessments, but he might not feel the pull of the anti-AMBP intuition, or he might only feel a mild pull.) As for what to say to the “committed TU theorist,” there isn’t a huge amount that I can say to someone like this, and someone who truly is a committed TU theorist will likely not be moved by most of the arguments in this dissertation. (I think that the dissertation does most of its work in capturing people within its orbit if they anyone other than someone who is a staunchly committed TU theorist, but I think that there are very few staunchly committed TU theorists, and this is because of the Repugnant Conclusion and the various other objections against TU that I raise in Part I of the dissertation.) But there are a few things that I can say even to this staunchly committed TU theorist: I can raise the Repugnant Conclusion, the Negative Repugnant Conclusion, and the Mere Addition Paradox. These objections to TU are powerful, and these are the objections that render most people who, all things considered are a TU theorist, to be TU theorists who are not the staunchly committed TU theorist. In any event, for the person who espouses the anti-AMBP intuition across the board, I can simply direct him to the Repugnant Conclusion and its variants and suggest that they raise concerns that should render him either not a TU theorist, or, at the very least, less of a staunchly committed TU theorist. As I said, though, I think that there are not many people in this camp (whole-hearted endorsers of the anti-AMBP intuition). Thus, I turn my attention to those who (1) have an intuition elicited by the Adam and Eve example that it is wrong to not procreate, and (2) who only feel this anti-AMBP intuition in the Adam and Eve case and cases like the Adam and Eve case where the population is sufficiently small. In response to this final group, I think that a few things can be said: First, the very fact that this group of people only feel the anti-AMBP intuition in these extreme or semi-extreme cases where the populations are very small should itself raise a red flag and suggest that they are not truly committed to the anti-AMBP intuition (or that they are, in the 288 cases with larger populations where they come out the other way, not truly committed to AMBP intuition). As I discussed when I discussed hybrid views, extreme implications of a view do at times have to be accepted, and we cannot just hybridize our views by letting our theory track our prima facie intuitions in all cases. Thus, this consideration doesn’t necessarily cut in favor of or against un-hybridizing by adopting the AMBP intuition and rejecting the anti-AMBP intuition, as opposed to vice versa. There are, however, some reasons to think that it is the anti-AMBP intuitions in the cases of small populations (such as the Adam and Eve case or a case where there is a world of one person) that must be rejected. It seems as though there are some reasons to think that other considerations might be infiltrating the view of the person we are considering who says that Adam and Eve are acting wrongly if they decide to not procreate. Here is what we should say to this person: First, as we did with the staunch TU theorist, we should remind this person here about the Repugnant Conclusion and its variants. These are powerful objections that seemingly have the ability to fully overpower and out-muscle (and certainly, at the very least, significantly weaken) the intuitions that Adam and Eve act wrongly if they choose to not procreate. Thus, even if one thinks that the Adam and Eve objection is a problem for AU, it seems that one should realize that there are seemingly significantly greater problems that afflict TU, and, at the very very least, the problems that afflict TU are not problems that are less significant than those afflicting AU that are raised by the Adam and Eve objection. 216 Second, as for which considerations might be infiltrating, there are a few that might be doing so: (1) My opponent would not have existed if Adam and Eve had failed to procreate, and that 216 And, to the extent that the Adam and Eve objection is tempting one to abandon AU in favor something other than TU (e.g., a hybrid view or a person-affecting view), the problems that afflict those views are even more significant than those that afflict either TU or AU (in my view, at least). 289 might make him disinclined toward the decision to not procreate. After all, he probably likes being alive. Similarly, if there was no procreation, my opponent’s loved ones and acquaintances would not exist, and all of this would further disincline him toward the decision to not procreate. The non-existence of oneself and one’s loved ones seems unappealing. In other words, personal and self-interested considerations might be infiltrating my opponent’s intuitions. (2) There typically are positive practical effects of having more people in a population. If the population is greater than two people, there seemingly might be greater happiness in the community due to social interactions and various other factors. It seems that a world of two people (and certainly a world of one person!) might be a lonely existence that lacks the happiness that could be cultivated in a larger population. These considerations, of course, would be inappropriate infiltrations, though, because the happiness levels in the Adam and Eve example have been stipulated. (3) There might be non-welfarist considerations (such as perfectionist considerations) infiltrating my opponent’s assessment as well. Perhaps one might have perfectionist reasons for thinking that a population of many people is better than a population of two people (and thus that Adam and Eve should procreate), and perhaps these perfectionist reasons are inappropriately infiltrating one’s intuitions in the Adam and Eve case (which is supposed to be just about welfarist reasons). The Adam and Eve objection is being used as an objection to AU, which I’m presenting as a theory of welfare-based goods, and thus non-welfarist considerations, such as perfectionist considerations, would be inappropriate infiltrations if they are infiltrating one’s assessment of the Adam and Eve objection against AU. * * * * Thus, in sum, I have provided a host of responses to the Adam and Eve objection: (1) There are two versions of the intuition that are elicited by the Adam and Eve objection, and my response to the bottom up objection, steeped in person-affecting intuitions, disposes of the first of 290 the two versions. (2) I think that the version of the intuition that my response to the bottom up objection disposes of is actually the more widely and strongly held of the two intuitions elicited by the Adam and Eve objection. Moving on, now, to the (weaker) intuition that I have not yet explained away: (3) Most people will not feel who feel the pull of this intuition will not feel this anti-AMBP intuition across the board, but, rather, they will only feel it in the Adam and Eve case and other cases with sufficiently small populations. (a) For the staunch supporter of the anti-AMBP intuition and TU, I point him to the Repugnant Conclusion and its variants. If he still isn’t moved, there’s nothing further I can do. (b) Most people will not be “staunch supporters.” I argue that these people should have their anti-AMBP intuition (and their AMBP intuition) cast in doubt because of their failure to consistently espouse one of them, leading to their at least implicitly espousing a hybrid account. (i.) Next, as for how to persuade the these non-staunch supporters that it is the anti- AMBP intuition that should be rejected, and not the AMBP intuition that should be rejected, (assuming that my arguments elsewhere against hybrid views are persuasive, and thus importing my non-hybridity assumption here), I (1) point them to the Repugnant Conclusion and its variants, and I argue that those objections are stronger than the Adam and Eve objection, and (2) I point to the various inappropriate infiltrations (three of which I describe above) that, in my view, are likely to be occurring in my opponent’s assessment of the Adam and Eve case. In sum, I think that these responses provide strong reasons to ward off (or, at the very very least, significantly weaken) the Adam and Eve objection against AU. 1.G. How AMU is better equipped to fend of the bottom up objection than is ATLU 1.G.1. Introduction As mentioned elsewhere in the dissertation, and as it has already become and will continue to become clear in various chapters, the two main Parts of the dissertation—Parts II and III— reinforce each other. The defense of AU in Part III is strengthened by the choice in Part II to endorse AMU as the most plausible version of AU. While the arguments in Part III apply to ATLU as well—and while I believe they are sufficient to provide a defense of ATLU—AMU is even better equipped to respond to the three objections than is ATLU. It has additional resources in this regard. Similarly, the flip side of this coin is that these aspects of Part III provide additional 291 support for the choice, in Part II, of AMU as the most plausible version of AU. These considerations are not among the points addressed in Part II, but they supplement the reasons for thinking that AMU is the most plausible version of AU. The current chapter on the bottom up argument against AU is a perfect example of this mutual reinforcement of Parts II and III of the dissertation at work. The points made throughout the chapter so far apply equally to ATLU and to AMU. I believe I successfully show in this chapter how ATLU and AMU are both able to respond successfully to the bottom up argument against AU. Notwithstanding this, however, I believe that AMU is even better equipped to respond to the bottom up objection than is ATLU. AMU has additional firepower in this regard. As such, not only is AMU even better able to respond to the bottom up objection than ATLU, but this very fact provides support for AMU over ATLU as the most plausible version of AU—support independent from the conditions provided in Part II of the dissertation. Section 1.G.2 will address AMU’s superiority in this regard with respect to person affecting intuitions. Section 1.G.3 will address AMU’s superiority in this regard with respect to the infiltration discussion of Section 1.E.3. As the reader will see in the upcoming sections, there is another way in which the superiority of AMU in handling the bottom up objection falls into two categories. On the one hand, some of the points in the upcoming sections do not show that there is a way in which AMU as a theory is itself better equipped logically than is ATLU to handle the bottom up objection, but, rather these points show that an AMU theorist is psychologically better equipped to handle the bottom up objection. In other words, for examples, an AMU theorist might be better equipped than an ATLU theorist to resist the appeal of person-affecting intuitions. This, however, if true, is a psychological feature of an AMU theorist, and not, strictly speaking, an advantage of the theory itself (AMU) in fending off the bottom up objection. On the other hand, 292 however, some of the points in the upcoming sections are indeed logical points that point to ways in which AMU as a theory is itself logically better equipped to fend off the bottom up argument. Thirdly, there are some points that involve some overlap between these two categories. There are some ways in AMU has a psychological advantage over ATLU, but where this psychological advantage arises specifically due to some logical commitments of the theory. Although, in a sense, this could be said about all of these psychological advantages, there is a sense in which, for some of them, there is more direct connection between a logical commitment and the psychological advantage. Thus, it does seem that this is a third category. In the following sections, I do not point out which arguments fall in which categories, but the reader should keep in mind these distinctions throughout the following sections. 1.G.2. AMU vs ATLU with respect to person-affecting intuitions Adopting a framework according to which the unit relevant for ethical inquiry is the conscious moment (such as that of AMU, but not that of ATLU) does not directly militate in favor of TU or AU, but it does potentially weaken arguments that have been partially responsible for AU’s widespread rejection. One of these arguments is the bottom up argument against AU. As I’ve argued already, a common feature of many people’s intuitions—be they explicitly or merely implicitly espoused—is that a morality should be person-affecting. While the nonidentity problem provides strong reasons for jettisoning these intuitions, they often seep into our theorizing even when we explicitly aim to adopt non-person-affecting moralities. In particular, as I argued in Section 1.D, in many examples—including Benign Addition, Reverse Egyptology, and Egyptology—that allegedly show that AU is mistaken and that TU is more plausible, I think that person-affecting intuitions are doing the lion’s share of the work. If, however, conscious moments—and not persons—are the unit relevant for ethical inquiry, then we should discount, 293 even more strongly than we currently do, person-affecting intuitions and arguments whose force relies crucially on them. After all, if conscious moments are the units relevant for ethical inquiry, how conscious moments are divided among persons—and which persons are constituted by which conscious moments—is a matter not of intrinsic importance, but merely of instrumental importance. Thus, a theory for which the unit relevant for ethical inquiry is the conscious moment—such as AMU—has an even stronger reason to reject and discount person-affecting intuitions than does a theory—such as ATLU—for which the unit relevant for ethical inquiry is the person. This point has two different aspects to it. First, as a matter of fact, an AMU theorist (or others for whom the unit relevant for ethical inquiry is the conscious moment) is probably less likely than is an ATLU theorist (or others for whom the unit relevant for ethical inquiry is the person) to find himself swayed by the person-affecting intuitions and arguments whose force relies crucially on them—of which I have argued the bottom up argument is one. Secondly, however, and perhaps more importantly, an AMU theorist (or others for whom the unit relevant for ethical inquiry is the person), when he reflects, will have a strong additional reason to reject arguments for person-affecting intuitions and arguments whose force relies crucially on them—of which I have argued the bottom up argument is one. While it is disputable which of these two aspects is more important, I think the latter is, because it speaks to a reflective judgment regarding the coherence and plausibility of a theory, whereas the former speaks merely to an intuition that might only be found plausible prima facie. At this point, the ATLU theorist might offer a rebuttal. As mentioned above, in section 1.D.2, the AMU theorist confronts an analogue of person-affecting intuitions—namely conscious- 294 moment-affecting intuitions (or CMAIs). 217 Thus, the ATLU theorist might point out, while perhaps the AMU theorist might not be swayed by person-affecting intuitions in the way that the ATLU theorist would, the AMU theorist might have the corresponding problem of being swayed by CMAIs. Thus, the ATLU theorist might say, when the AMU theorist confronts the conscious- moment-based analogues of Benign Addition, Reverse Egyptology, and Egyptology, he will be swayed by CMAIs to the same extent that ATLU theorists would be swayed by person-affecting intuitions. In theory, this is true, and the ATLU theorist is correct. The AMU theorist confronts an analogous problem, and it is a problem that should not be ignored. This notwithstanding, however, the AMU theorist confronts a substantially less serious problem, and this is for two reasons. First of all, it’s important to remember that the reason that the person-affecting issue even is a problem at all is not that person-affecting intuitions might be plausible, but it’s because they are ingrained in us and seep unconsciously into our theorizing. (If they were consciously seeping in, then we have already addressed the fact that the nonidentity problem provides sufficient reason to consciously reject them.) In light of the fact that the problem is a function of an unconscious process, it is highly relevant how strongly the person-affecting intuitions are seeping in. And here is what distinguishes AMU from ATLU. Person-affecting intuitions are deeply ingrained in our psychology, but CMAIs are not. Yes, it’s true that CMAIs do have some intuitive pull, but this is not the way we generally think. 218 Regardless of what our Theory X of choice is, it is a fact about our psychology that we think in terms of persons. Thus, we aren’t as likely to be unconsciously swayed by CMAIs as we are to be unconsciously swayed by person-affecting intuitions. Moving from the person to the conscious moment is precisely what removes some of the unconscious pull 217 CMAIs are just as intractable as a theory as are PAIs, and for the same reason. Just as in the context of persons, we must also aggregate impersonally in the context of moments. 218 In fact, it seems that, in many cases, we do not hold CMAIs, and we actually often hold anti-CMAI intuitions. After all, most people think that, for example, death is bad (even agent-neutrally). 295 of the person-affecting / conscious-moment-affecting intuitions. The second reason that CMAIs are less likely to unconsciously sway us than are person- affecting intuitions is related to the first explanation, but slightly different. Or, perhaps, the second reason could be couched as a partial explanation for the first reason. Regardless of which of these two understandings of the following point one prefers, the thought is as follows: While someone who speaks of person-affecting intuitions has a (perhaps imprecise) sense of what constitutes the person that is being affected, it’s less clear that the same can be said when we are speaking about a conscious moment that is affected. What is it that even makes a conscious moment the same conscious moment that is thus capable of being affected? This question is somewhat more difficult to articulate than the corresponding person-based question is for non-reductionists of personhood, and even for most reductionists of personhood. It’s not clear that an account can be provided of what constitutes the same conscious moment as something else. Even if such an account can be provided, however, I do think that the notion is sufficiently opaque that this opaqueness might prevent us from feeling the pull of CMAIs. Thus, in light of these reasons, I don’t think that AMU confronts a force (in CMAIs) that is as strong as the force of person-affecting intuitions for ATLU. Additionally, I think that the force of person-affecting intuitions on AMU will be significantly less strong than the force they have on ATLU. In light of both of these considerations, I think that AMU is even better equipped to respond to the bottom up argument than is ATLU—even though ATLU itself has a response that I believe to be sufficient to turn away the bottom up argument. 1.G.3. AMU vs ATLU with respect to the IMF explanation (provided in Section 1.E.3) In addition to AMU being better equipped to handle the bottom up objection than ATLU is for the reasons discussed in the previous section, there is an additional way in which AMU is 296 better equipped to handle the bottom up objection. Just as the AMU theorist is better able to avoid the infiltration of person-affecting intuitions into prima facie intuitions and post-reflective theorizing, so to the AMU theorist is better able (than the ATLU theorist) to avoid the infiltration of the IMF-related intuitions, discussed in Section 1.E.5, that can cause a potential AU theorist to have difficulty warding off the bottom up objection against AU. In Section 1.E.3, I provided an additional explanation for why a potential AU theorist might find it difficult to accept a temporally neutral temporal account and reject a current / forward-looking temporal account. The thought was that, regardless of what our agent-neutral theory of goodness is and even regardless of what our theory of how to assess the goodness of an individual’s life, we are all (for the most part) constrained by IMF: We seek, in our lives, to maximize happiness aggregated according to TU-type aggregation and with a current / forward- looking reference class. This thus inclines most of us toward accepting a view of goodness for our own life that is assessed by aggregating according to TU-type aggregation and with a current / forward-looking reference class. Further, I said that this reasoning also infiltrates our assessments of those whom are close to us, and perhaps even those who are distant from us. Then, explicitly or implicitly using the Pareto principle, these agent-relative assessments are translated into agent- neutral assessments. Thus, I argued, IMF infiltrates our agent-neutral theorizing. I showed in Part II Chapters 2 and 3, however, how this reasoning was flawed. Although, this hypothesis from Section 1.E.5 applies to all versions of AU, I think that an AMU theorist is better able to avoid the infiltration mechanism that is the subject of the hypothesis. Further, I think that this is the case even though even an AMU theorist will be constrained by IMF. The reason for this is that ATLU as a theory is explicitly committed to a number of the claims that are made on the above chain of faulty reasoning. In light of this, there are fewer steps that must be assumed for the ATLU theorist to have his agent-neutral theory be 297 infiltrated by a current / forward-looking reference class. For example, in addition to being constrained by IMF, an ATLU theorist is (even reflectively) committed to the view that given a stipulated past, a person’s life is best if it maximizes its happiness as measured by TU-type aggregation with a current / forward-looking reference class. Yes, ATLU espouses a temporally neutral reference class for the assessment of the goodness of a person’s life, but, given that uses TU-type aggregation, if a person’s past is stipulated, the two reference classes will yield coextensive assessments. This, however, is not the case for AMU, because it is an open question how the AMU theorist should assess the goodness of an individual’s life, and, furthermore, in Part II Chapter 3, I argued that there are strong reasons for an AMU theorist to adopt AU-type aggregation for the goodness of an individual’s life. Since AU-type aggregation does not necessarily yield the coextensiveness that TU-type aggregation would in the case just described, the fact that at an AMU theorist either would favor AU-type aggregation for assessment of the goodness of an individual’s life, or that, at worst, he might do so, means that the AMU theorist does not necessarily agree with the claim that the ATLU theorist would—namely, that, given a stipulated past, maximizing one’s happiness with a current / forward-looking reference class will result in the best life for the individual in question. Furthermore, since the ATLU theorist is on board with the above claim about agent- relative goodness, and in light of the ATLU theorist’s choice of the person as the unit relevant for ethical inquiry, there is an additional step of the reasoning in the flawed infiltration chain that the ATLU theorist is committed to accepting. In light of the Pareto principle, all else equal, the ATLU theorist will also think that, assuming a stipulated past, increasing a person’s happiness going forward will increase the goodness of the state of affairs agent neutrally. For the AMU theorist, however, this is step does not follow. Even if an AMU theorist were on board (which he is not) with the claim that, stipulating a fixed past, maximizing one’s happiness with a current / forward- 298 looking reference class would result in the best life for an individual, the AMU theorist still would not be committed to the claim that this would lead to the better state of affairs agent-neutrally. The reason for this is that the unit relevant for ethical inquiry for the AMU theorist is the conscious moment—not the person—and thus the Pareto principle would need to be applied to conscious moments—not persons—and thus, the inference that the ATLU theorist is able to make would not follow for the AMU theorist. Thus, in sum, while the potential ATLU theorist believes in a temporally neutral reference class for agent-relative and for agent-neutral goodness, ATLU’s other commitments (namely its commitment to TU-type aggregation for the goodness of an individual’s life, and its choice of the person as the unit relevant for ethical inquiry) result in ATLU making prescriptions in many cases that are coextensive with those of theories with a current / forward-looking reference. Thus, despite the difference between ATLU and such theories, it is quite likely that that this coextensiveness and also the relative simplicity of focusing merely on the present and future will lead intuitions in favor of the current / looking reference class to seep into the potential ATLU theorist’s agent-neutral theorizing. The AMU theorist, however, due both to his choice of the conscious moment as the unit relevant for ethical inquiry, and due either to espousing AU-type aggregation for assessments of agent-relative goodness or to having it be an open question whether he should espouse AU-type or TU-type aggregation for assessments of agent-relative goodness, does not end up necessarily having prescriptions that are coextensive with theories that have a current / forward-looking temporal component. In light of this, although a potential AMU theorist is constrained by IMF, and although he might well fall prey to the faulty chain of reasoning that would lead the intuitions in favor of a current / forward-looking temporal component to seep into his account, it is much less likely that this result will occur for the AMU theorist than that it will for the ATLU theorist. 299 The AMU theorist will need to sign onto numerous steps of reasoning to bring about this result, whereas the ATLU theorist’s commitments already get him most of the way to this result, and it is thus much easier for the current / forward-looking temporal component of agent-neutral (and of agent-relative) goodness assessments to infiltrate the ATLU theorist’s intuitions. To the extent that the foregoing is plausible, this provides an additional reason for why AMU is better equipped to stay committed to a temporally neutral temporal component, despite prima facie temptations—and perhaps somewhat reflective temptations—to reject a temporally neutral account as implausible. 1.G.4. Summary In sum, although I have argued that AMU and ATLU are both able to successfully respond to the bottom up argument, I have now argued, in Section 1.G, that there are various ways in which it seems that the AMU is even better equipped (than the ATLU theorist) to respond to the bottom up objection. Some of these ways in which the AMU theorist is better equipped are due to logical features of his theory, some are merely psychological advantages that he has as an AMU theorist, and some of the ways in which he is better equipped involve overlaps between these two categories. 1.H. Extension section on animals and the bottom up objection Utilitarianism concerns itself with the ability to feel pleasure and pain, and, thus, it seems clear, following Bentham, Singer, and others, that it would be speciesist to not have Theory X apply to non-human sentient creatures as well. Although there are many interesting questions that arise in the context of animals and candidates for Theory X, and a full exploration of these questions could occupy an entire chapter (or more), here I just flag a few of these topics for the 300 reader. In what follows, I consider a few topics, I provide a few comments, and I raise some questions that I will return to in future projects. I do so at this juncture, because the questions that arise relate to the bottom up objection against AU. Although TU does not seem to confront particularly interesting questions in the context of animals, interesting questions do arise in the context of how AMU and ATLU handle animals. Further, one might find AMU’s and ATLU’s treatment of animals to be implausible, and this might thus lead one to doubt the plausibility of AMU and ATLU. I raise these objections in what follows, I explain that they rely on various empirical assumptions (albeit perhaps quite plausible assumptions), and I explain that to the extent that these topics constitute objections against AMU and ATLU, they are akin to the bottom up objection and they thus can and should be replied to in the same way as is the bottom up objection. Consider a topic that arises in the context of animals and ATLU: ATLU will seem to say (though AMU will not) that it will be better off to prevent the existence of sentient animals with life spans that are quite short compared to the range of life spans of all sentient creatures. This implication, by itself, might be seen to be a strong reason to have doubts about ATLU as a theory. Though various attempts could be made to relativize the theory to different species, it seems initially that any such attempt would be ad hoc and not track what seems to be of value. Thus, one might think that this constitutes a reason to doubt the plausibility of ATLU. AMU, however, is confronted with difficulties as well when non-human creatures are taken into consideration. Of course, there are huge practical obstacles in estimating which creatures are sentient, to what extent those that are sentient are sentient, and what the sentience is like for these creatures. Leaving these issues aside, though, it seems quite possible that some creatures that lack highly developed nervous systems might have considerably different (and, quite 301 possibly, dramatically lower) welfare potentials. Mill certainly thought that this was the case. 219 If this is so (and if creatures more often than not have net positive happiness), both AMU and ATLU would seem to suggest that species with lower potentials, if possible, should be prevented from coming into existence. In other words, for example, perhaps we should prevent lobsters from mating and propagating their species. These concerns might lead some readers to doubt the plausibility of AMU and ATLU. Although interesting questions abound here, in short, I do not think that these considerations cast doubt on AU theories. After raising a set of possible empirical responses, I’ll explain that the concerns here are akin to those in the bottom up objection. One response the AMU theorist can give is that all conscious creatures are likely to have the same welfare potential, and that this is the case because of similarities in the patterns in which humans and other creatures feel pleasure and pain. Another (related) response could be that despite great differences in conscious experience, welfare potential in the relevant sense is the same in lower creatures because their experience might differ simply in the frequency of conscious moments—as opposed to in the intensity of the conscious moments themselves. Empirical questions abound here. (So, too, however, do theoretical questions.) To the extent, though, that it is the case for one of various reasons that non-human animals have lower happiness levels (be it according to ATLU or AMU), and, thus, that it is the case that these creatures should be prevented as much as possible from being brought into existence, this might seem prima facie counterintuitive or wrong to many readers. My answer to this, though, is simply my answer to the bottom up argument, because the issues are largely the same. This bullet should indeed be bitten and for the same reasons as is the case for why AU says that people who would be happy but below the average level of happiness should not be brought into existence. 219 Mill. 302 The intuitions that might be leading one to think that AU gets the wrong result when it prescribes the prevention of the propagation of species with either lower welfare potentials (relevant for AMU or ATLU) or with shorter lives (relevant for ATLU but not AMU), are the same intuitions underlying the bottom up objection that has been the topic of this chapter of the dissertation. My response to the objection to AMU and ATLU here in the context of animals is thus the same as my answer to the bottom up objection more generally. There are a handful of other concerns that arise for AU theories in the context of animals, but, in my view, most of these other issues also boil down to variations of the bottom up objection, and thus these issues can be resolved in a similar manner. Regardless of whether or not the context of animals poses problems for certain candidates for Theory X, a number of interesting questions arise regarding animals and regarding the application of a candidate theory to all sentient creatures at once. Many of these questions are empirical, but many are also theoretical, and there is a large overlap between both types of questions. Going forward, it’s important to remember that (in my view, and in the view of most utilitarians) any plausible account of Theory X will need to be plausible not only across human beings, but also across all sentient life. Further, considering candidate theories in this broad context will potentially give rise to newly discovered objections to theories, and this might result in new challenges, but it might also help bring about greater clarity—clarity that might, in some cases, be hampered by focusing only on human beings. 1.I. Conclusion This chapter has described and addressed the group of related objections to AU that I call the “bottom up objection”—an objection to AU that has persuaded many philosophers that AU must be false. In my view, however, the bottom up objection, despite its prima facie plausibility, 303 does not succeed in its attempt to show that AU is false. In this chapter, I explained how the AU theorist can successfully respond to the bottom up objection. In my main response, which I provided in Section 1.D, I argued that the AU theorist should endorse a temporally neutral reference class, and that, if he does so, he can successfully ward off the bottom up objection. Further, I argued that the AU theorist can show that the bottom up objection’s prima facie plausibility turns out not to be due to greater plausibility of TU than AU, but, rather, to be due to an alignment of the prescriptions of person-affecting intuitions with those of TU (and not with those of AU) in the cases that are offered in the bottom up objection—when there actually are an equal number of cases (which the proponent of the bottom up objection doesn’t introduce) where the prescriptions of person-affecting intuitions align the prescriptions of AU (and not with those of TU). In light of this and the various considerations that relate to this, which I discuss in the chapter, the appropriate conclusion to draw is that the bottom up objection does not successfully cast doubt on the plausibility of AU. Further, after providing my main response to the bottom up argument, I then provided a number of additional considerations that provided further support for my conclusion that AU is able to successfully respond to the bottom up objection. ====================================================================== Part III Chapter 2. Two Hells 2.1. Introduction Perhaps the objection to any version of AU that is regarded as the strongest is one that Parfit raised and termed Two Hells. 220 I have yet to discuss it because it raises concerns of a different type from the objections and alleged counterexamples already discussed: It addresses AU’s treatment of suffering. The objection goes as follows: Consider two possible hells. In Hell 1, ten people exist and suffer great agony. In Hell 2, there are ten million people who suffer ever so 220 Parfit, p.393. 304 slightly less agony. AU will say that Hell 2 is better. Parfit and countless others, however, find it obvious that Hell 2 is far and away the worse of the two states of affairs. This chapter will address Parfit’s Two Hells objection to AU, and it will also consider some related issues. First, in Section 2.2, I explain that the simplest responses to the Two Hells objection are to simply rely on my arguments elsewhere in this dissertation to diffuse it. One such option is to just baldly state that there is no reason for cases of suffering (i.e., negative happiness) to be treated any differently from cases of pleasure (i.e., positive happiness). More importantly, however, the second such option is to rely on Part IV of this dissertation, which argues that zero points do not exist, and thus there is no meaningful sense in which a happiness level can be labeled as positive or negative (and thus, respectively, better than or worse than non-existence). If my arguments in Part IV succeed, this completely pulls the rug out from under the Two Hells objection, because it crucially requires the notion of negative wellbeing. After Section 2.2, however, I proceed by assuming (counterfactually) that my arguments in Section 2.2 were not successful and that the Two Hells objection has not already been set aside. In other words, I proceed under the assumption that either my claim that negative happiness does not raise any additional issues to be resolved by AU than do cases of positive happiness, and under the assumption either that zero points do exist or that their non-existence does not remove the power of the Two Hells objection. I then proceed to provide arguments that show that, even if we make the assumptions articulated in this paragraph, AU still is able to successfully respond to the Two Hells objection. In Section 2.3, I offer a response to the Two Hells objection, which, if we leave aside the arguments I make in Section 2.2 (e.g., about zero points), constitutes my “main response” to the Two Hells objection. I then offer an additional response in Section 2.4. In conclusion, despite the fact that the Two Hells objection might, prima facie, be a devastating objection to AU (and perhaps, some might argue, the strongest of all objections 305 against AU), I argue that I provide at least two responses that are independently sufficient for the AU theorist to use to successfully respond to the Two Hells objection (one being the one based on the lack of zero points, and the other being my “main response”). 2.2. Arguments from other portions of my dissertation preventing the Two Hells objection from having force 2.2.A. A preliminary argument that the Two Hells cases (i.e., cases of suffering) do not require treatment different from cases of positive happiness My simplest response (though likely not my most persuasive response) to the Two Hells objection is to baldly state that there is no reason for cases of suffering (i.e., negative happiness) to be treated any differently from cases of pleasure (i.e., positive happiness), and to state that cases of negative happiness do not raise any additional issues to be resolved by AU beyond those that are raised in the cases of positive happiness. In my view, this is completely correct. I am certainly in the minority, but I think that others have been mistaken, and that AU provides an adequate assessment of the Two Hells case. It seems plausible to me that there is nothing wrong with its symmetrical treatment of cases of positive happiness and negative happiness. Narveson and Boonin-Vail, themselves, however, do not think that their slogans about the addition of happy people apply to the addition of unhappy people. But this asymmetry seems unwarranted. It’s my view that (and to use the words that these authors used in the context of positive happiness), with unhappy people, just as with happy people, “How large a population you like is purely a matter of taste,” and “there is no moral argument at issue here.” 221 In my view, it is not bad that unhappy people exist because their lives contain unhappiness; unhappiness is bad because it is bad for people. We should aim to reduce unhappiness for people, not reduce the 221 Narveson, p.68. 306 number of people for a reduction of unhappiness. 222 Another way of stating this view of mine is to say that I think that if one accepts (1) my prima facie argument for AU (which I provided in Part I Chapter 2), (2) my responses to the bottom up objection against AU (which I provided in Part III Chapter 1), and (3) my responses to the top down objection against AU (which I will provide in Part III Chapter 3), then I think that one should be persuaded to accept AU even without the need for a specific response to the Two Hells objection. Another way of saying this is to say that I think that the arguments that I provide in these other chapters are broad enough to cover all cases, and thus they are broad enough to apply to cases of positive happiness and negative happiness—those arguments are not in any way limited to cases of positive happiness. Thus, in my view, even without saying anything further, and without speaking specifically to the Two Hells cases, I think that my arguments in Part I Chapter 2, Part III Chapter 1, and Part III Chapter 3 provide an AU theorist with a defense of AU sufficient to ward off the Two Hells objection. As I said, I’m in the minority, though, and most people probably do feel (1) that the Two Hells objection elicits intuitions against AU that are different from the general anti-AU intuitions AU that might be elicited by other objections against AU, and (2) that these intuitions are strong enough to cast serious doubt on the plausibility of AU. Thus, the response that I have given in this sub-section will not satisfy many readers. Although I find the points here sufficient, most readers will not, and thus more needs to be said to show that the Two Hells objection is not decisive against AU. Further, even I (who am of the view that my general arguments in favor of AU are sufficiently broad to handle the Two Hells cases) do agree that there is a different (and additional) 222 These authors’ statements are made within the framework of a person-affecting morality, but, as discussed in Part I, the general gist of the intuitions that these authors refer to can also be understood in a framework that is not a person-affecting one (and which, instead, is impersonal). Thus, my purpose in using these quotations is to evoke the intuition itself. I do not espouse a person-affecting framework. 307 type of intuition that is elicited by cases of suffering and I do feel the force of these intuitions. Even though I think that my arguments are sufficiently broad to handle the Two Hells cases, even someone of my view will benefit from being able articulate an explanation of why it is that the Two Hells cases elicit the intuitions that they do (even if, all things considered, these intuitions are outweighed by other intuitions that one has and thus do not move one, all things considered, to reject AU). The arguments regarding zero points, in the next subsection, do not articulate an explanation of why it is that the Two Hells cases elicit the intuitions that they do, but the responses that I give to the Two Hells objection in Sections 2.3 and 2.4 do articulate explanations of why it is that the Two Hells objection elicits the intuitions that they do. Thus, I think that the arguments in Sections 2.3 and 2.4 not only will be valuable for the purpose of persuading a reader, who at this juncture is moved by the Two Hells objection to reject AU, to realize that he instead should see that AU has a successful response to the Two Hells objection, but, also, the arguments in those sections will be valuable even for someone like me, who is already persuaded by the considerations in this subsection, because the arguments in those sections will provide an understanding of why the Two Hells objection elicits the intuitions that it does, and thus these explanations will further entrench someone like me in the camp that views AU as successfully warding off the Two Hells objection. 2.2.B. The Two Hells objection requires the existence of zero points, but, as I argue in Part IV of the dissertation, zero points do not exist There is a second way in which I think my arguments elsewhere in the dissertation offer a successful warding off of the Two Hells objection, and I think that this second way is significantly more persuasive than the first—indeed, it seemingly completely disposes of the Two Hells objection. This second way is as follows: In Part IV of the dissertation, I argue that what I call “zero points” do not exist. I’ll defer a full explanation of what the term “zero point” means until 308 Part IV itself, but for now it will suffice simply to say that if no zero point exists, then there is no meaningful sense in which a happiness level can be labeled as positive or negative (and thus, respectively, better than or worse than non-existence). If my arguments in Part IV that there are no zero points succeed, as I believe they do, this seemingly pulls the rug out from under the Two Hells objection, because the Two Hells objection crucially requires the notion of negative wellbeing. If there is no meaningful distinction between positive and negative levels of wellbeing, this seemingly removes the force of the Two Hells objection. As a result of there not being zero points, the Two Hells case wouldn’t arise as described, and this is because (1) it wouldn’t be the case that people in the two hells cases are assigned (non-arbitrarily, at least) a negative happiness level, and, similarly, (2) it wouldn’t be the case that it would be correct to describe the two hells cases as involving worlds where it would be better if the people in the worlds (i.e. the hells) did not exist at all. Someone who wants to argue (against me) that that the Two Hells objection does have force against AU will need to take at least one of two routes: (1) argue that I fail in my attempt to show that there are no zero points, and argue that there are indeed zero points, or (2) argue that even if there are no zero points, the Two Hells objection still maintains its force against AU. I won’t take up option number one here, and I will simply point the reader to Part IV of the dissertation and let him or her decide based on my points there. I will, however, briefly address option number two. In connection with option number two, perhaps one might argue as follows: Yes, if it’s the case that there are no zero points, then it would be incorrect to label the two hells cases as cases of negative happiness or, similarly, as cases where it would be better if the suffering individuals did were not alive. Notwithstanding this, however, this does not make us unable to describe the hells as including individuals who are “undergoing terrible pain and suffering.” Saying this seemingly would not be inconsistent with zero points (of a monistic property of happiness, 309 which is what is directly relevant to questions in population ethics) not existing. 223 Or, similarly, even if there are no zero points of happiness, one seemingly could describe the Two Hells cases in terms of what events occur there: For example, we could define one of the hells as involving lives in which people are constantly pricked by sharp needles. Thus, one might think that the force of the Two Hells objection against AU would still remain even if there are no zero points, because the two above formulations could seemingly be provided even if there are no zero points, and these (newly described) cases seemingly would elicit the same intuitions. I don’t think that my opponent’s argument works, though. It seems to me that it really is the existence of a zero point that is what is convincing people that it’s so bad that for people to be existing in the two hells cases and thus that having more people living with lives of this sort is a bad thing. If there is no zero point, there then simply is no fact of the matter about whether a person with suffering being brought into existence is a good, bad, or neutral thing. If there is no zero point, then the comparison between a life’s existence and its non-existence is an undefined comparison. 224 If there is no zero point, relative comparisons can be made between happiness levels, but the happiness levels will not have absolute values. Thus, to the extent we are making determinations based solely on happiness / wellbeing, we would be left only with the ability to make comparisons in the way that AU makes comparisons. Thus, in sum, it does seem as though the Two Hells objection indeed requires zero points. Thus, if my arguments in Part IV for the non-existence of zero points succeed (as I believe they do), this fully pulls the rug out from the Two Hells objection, and the Two Hells objections will thus not cast any doubt at all on AU. 223 See Part IV Chapter 8 of the dissertation for discussions regarding the differences between, on the one hand, possible zero points of pleasure and pain, and, on the other hand, possible zero points of a monistic property of happiness that is a single metric that takes as inputs pleasure and pain. The discussion there also addresses the ways in which these notions interact and it addresses which properties are relevant for the purposes of determining whether zero points for AU or TU, and what the existence of zero points of pleasure and or pain might mean for whether there the monistic property “happiness” has zero points. 224 See Part IV of the dissertation for a discussion of this and related points, and for an explanation of why these comparisons would be undefined. 310 2.2.C. Summary At this point, given the two arguments in Section 2.2 (and especially given the very strong argument regarding zero points in Section 2.2.B), it is my view that the Two Hells objection has been fully warded off as a result of arguments that I provide elsewhere in this dissertation, and there thus is no need to further address the Two Hells objection. Notwithstanding this, though, there might be some readers who are not persuaded by my claims that (1) cases of negative happiness do not raise any additional issues to be resolved by AU than do cases of positive happiness, (2a) that zero points do not exist, or (2b) that the non-existence of zero points removes the force of the Two Hells objection. In light of the fact that some readers might not be persuaded by these claims, I will now, in what follows, assume that I’ve been wrong about claims 1, 2a, and 2b, and I will proceed to specifically consider the Two Hells objection. My view is that, for the reasons set forth in Sections 2.3 and 2.4, even granting these unfriendly assumptions, I can show that the Two Hells objection does not successfully cast doubt on AU. (Going forward, though, do keep in mind that the easiest route to warding off the Two Hells objection will still be to make the argument based on the non- existence of zero points.) 2.3. My main response: The Limited Force of the Two Hells Example Although Larry Temkin is not a utilitarian, and although he makes it clear that he finds average utilitarianism to be particularly implausible, 225 I think that a discussion of his, in Rethinking the Good, provides fodder for an AU theorist to respond to the Two Hells objection. 226 225 Temkin, pp.319-23. 226 As Temkin says, though, the original source of the sort of example that he discusses is Stuart Rachels. Rachels’s original example was described in his unpublished Philosophy, Politics, and Economics thesis, “A Theory of 311 According to Temkin, when comparing two states of affairs in which a burden of a very similar quality is imposed on two different groups, the number of people affected in each state of affairs is relevant—or, as he says, “additive aggregationist” reasoning is plausible. For example, according to Temkin, in a case like Two Hells, Hell 2 is much worse than Hell 1, and this is because the difference in quality of the burdens is very small, and this enables the number of individuals in Hell 2 to make the badness of Hell 2 outweigh the badness of Hell 1. Whether or not the badness of Hell 2 and Hell 1—at least with respect to the utility ideal, leaving open the possibility that ideals other than utility are valuable—is directly proportional to the total units of utility is an open question, but at the very least, Temkin argues that it is plausible that the number of affected parties is relevant. Further, he argues, a large enough disparity of people can outweigh the slight disparity in quality of the burden that cuts in the opposite direction. According to Temkin, however, additive aggregationist reasoning is not plausible for comparisons of two states of affairs in which there is a large disparity in the quality of a burden. Thus, what Temkin means is that in comparisons of this sort, it might be the case that no number of people affected by the lesser burden will make that state of affairs worse than a state of affairs with a particular number of people affected by a severe burden. To use his example, there is no number—however large—of people afflicted by the mild annoyance of mosquito bites that would make such a state of affairs worse than a different state of affairs in which some small number of people undergo severe torture. Temkin ultimately argues that the two above claims, combined with a continuity claim (asserting that a sequence of steps between cases with merely slight quality disparities would ultimately get one from the case of torture to a case of mosquito bites), and the claim that the Beneficence” (Oxford University, 1993), and Rachels further discusses the topic in his “Counterexamples to the Transitivity of Better Than,” Australian Journal of Philosophy 76 (1998): 71-83. In my discussion, I will continue to refer to Temkin’s example and discussion, but it should be remembered throughout that Rachels is the original source. 312 better-than relation is transitive are inconsistent. He laments the need to reject any of these four claims, but he argues that we must reject the transitivity of the betterness relation. There are various reasons to doubt the plausibility of this position, but they are beyond the scope of this chapter’s analysis. Instead, let’s focus on the implications of Temkin’s first and second claims for the Two Hells objection—assuming that his claims about people’s intuitions are fairly accurate, as seems likely to be the case. Temkin’s discussion suggests that while people will generally side with the TU theorist in Parfit’s Two Hells case and in similar cases, people will generally side with the AU theorist in cases of suffering where either Hell 1 (or Hell 2) is compared to a Hell 3, where there are an enormous number of individuals each of whom has a utility level that is only slightly negative (e.g. due to mosquito bites or mild headaches). Thus, it seems that comparisons between states of affairs involving suffering do not in fact only support judgments in favor of TU and against AU, but, rather, there will be cases that support the opposite judgment. Further, since it’s not clear exactly where people will implicitly draw the line between severe and slight quality disparities, it’s far from clear whether these cases, collectively, would provide more support for AU or for TU. Thus, the point here bears some structural resemblance to the points about reference class made in Part III Chapter 1: While it might be the case that intuitions side with TU in the objections given, this is not due to the merits of TU, but rather to the fact that these are cases where the principle that is doing the work aligns with TU, and there seem to be equally many cases where this principle aligns with AU. As was the question in Part III Chapter 1 Section 1.D.7.b, though, the question then becomes whether this other principle that is more tailored to our intuitions is in fact the principle that should be adopted—instead of either AU or TU. In Part III Chapter 1, it was argued that person-affecting principles cannot be adopted because of the great difficulties they face. Similarly 313 here, it seems that we cannot maintain Temkin’s two principles about different types of comparisons. Unless one is willing to accept his claims about transitivity, or unless one rejects some other seemingly innocuous claims, it is inconsistent to hold both of his claims about numbers mattering and not mattering in their respective contexts. Thus, it seems that one is in fact required to have a single theory of whether the number of people matters or does not matter. This does not necessarily mean, however, that one must either accept a pure version of TU or AU for the purposes of aggregation. One could adopt a hybrid view of some sort where, perhaps, value is equal to the quality of the burden times the quantity (i.e. the quality times (the quality times the number of people)). This, however, would simply shift the relative weights of quality and quantity, and it would not avoid the problem of Temkin’s spectrum. (I, of course, argue in Part I Chapter 2, though, that all hybrid views should be rejected.) Thus, ultimately, it seems that one must either reject the absolute claim about the comparisons between cases of large disparity in quality or the absolute claim about the comparisons between cases of slight disparity in quality, or both, and it seems that people will find rejecting either of them to be difficult. As such, it seems that cases of suffering do not provide strong support in favor of TU or in favor of AU. As with cases of positive utility, there are cases where our intuitions support one and there are cases where our intuitions support the other. 2.4. The Relevance of Person-Affecting Intuitions Given that the context of suffering was supposed to provide strong reason to reject AU alone, it is no small feat to have shown that this context provides potentially equally strong considerations in favor of AU and TU respectively. If the above considerations bring us anywhere close to equipoise between AU and TU, this is a success for the purposes of this dissertation, because this dissertation has proceeded by first motivating AU, and then attempting to show that 314 the main objections against AU are not as strong as they otherwise might have seemed, and not strong enough to merit abandonment of AU. Thus, the discussion of suffering could stop here. Nevertheless, it’s worth continuing for two reasons: First, one might think that the intuition in the Two Hells case is stronger and harder to abandon than the intuitions in cases of comparisons involving large quality disparities. Second, even if we truly are in equipoise for the cases of suffering, it’s worth hazarding somewhat of a positive argument (in favor of AU’s account of suffering) to attempt to explain people’s intuitions in the Two Hells case. Thus, let’s consider a case that is similar to Two Hells, but in which the larger population is at the exact same (negative) utility level as is the smaller population. (It should be noted that this is the type of case that is at one end of Temkin’s spectrum, because there is not even a slight quality disparity that favors the larger population.) Further, let’s compare this case to its analogue where everything is the same except that the utility values are positive. Though there certainly are readers who will think that the state of affairs with the larger happy population is better than the one with the smaller population, there will be many readers who will be willing to concede that population size is a matter of taste in the positive utility context, but who will deny the analogous claim in the negative utility context. What might explain this? I think that this difference can be explained by person-affecting intuitions. Let me explain. In the positive utility context, it might be thought that if the additional individuals are brought into existence they are thereby benefited, but that if they are not brought into existence, there is no individual that is harmed. These individuals brought into existence wouldn’t be benefited in the sense described in Part I, but at least there would be people to point to who could at least in some sense be described as having benefited, whereas if they are not brought into existence, there is no such individual. In the negative utility case, for the same reason, one might think that there are individuals who have been harmed if the additional people are brought into existence, but that no 315 one has been benefited if they are not brought into existence. 227 Thus, through a quasi-person- affecting lens, the positive utility case involves either a benefit or nothing, and the negative utility case involves either a harm or nothing. This observation alone, however, does not explain the intuition that numbers might matter in the negative utility case but not in the positive utility case. There still is a benefit in one case and a harm in the other, and the intuition we are trying to explain is that we are neutral about the additional benefit, but not neutral about the additional harm. An unexplained asymmetry remains. However, once we see that person-affecting intuitions likely influence us to view the positive utility case solely in terms of a benefit or no benefit and the negative utility case solely in terms of a harm or no harm, the heavy lifting has been finished. It is well known that our common sense moral intuitions take into account distinctions such as those between doing and allowing and between conferring harms and failing to confer benefits. Thus, once viewed through this lens, it is unsurprising that even those who think population size is a matter of taste in the positive utility context will find it to be a moral matter—and not merely a matter of taste—in the negative utility context. So far so good, but how does this support AU? This supports AU because it casts doubt on our intuitions regarding Two Hells and cases similar to Two Hells. First, it seems that the intuition against AU in cases like Two Hells, is quite possibly a function of person-affecting intuitions, and it has already been shown why they are problematic in the population ethics context. Further, it seems that the intuitions on which they rely in this context are quite possibly ones rooted in distinctions—such as the distinction between harming and failure to benefit—that are components of deontological theory, and which not only are contestable in their own right, but at the very least seem inapplicable in our current inquiry (These are inapplicable to our current inquiry because, as 227 It is the additional people who have, in a sense, been “harmed.” 316 I stated at the outset of Part III, the objections against AU that I’m considering in Part III (1) implicitly rely on intuitions associated with action, and (2) implicitly assume consequentialism.) Thus, these considerations might provide the beginnings of an explanation of why someone who is inclined to be in favor of AU in the positive utility context is perhaps less willing to be on board with AU in the Two Hells-type context. 228 Further, as argued in Section 2.3, above, even if people might be inclined to have anti-AU intuitions in the Two Hells case itself (and thus reject the arguments in this section), suffering cases on the whole provide no more support for TU than they do for AU. For these reasons, Parfit’s Two Hells objection—at the absolute least— should not be considered a strong reason for rejecting AU. 2.5. An additional possible explanation of anti-AMU intuitions in the Two Hells case In addition to the foregoing account of Two Hells, I think that there is an additional explanation of anti-AMU intuitions in the Two Hells case. I think that the following is a secondary explanation, but, notwithstanding this, I do think that there are some insights in this explanation that are worth taking note of. Here is the explanation. 228 Not only has the foregoing discussion attempted to provide an explanation of why one might have asymmetrical intuitions in the positive and negative utility contexts, but it has aimed to show that—once made of aware of this explanation—one should revise one’s view and maintain a position that has a symmetrical treatment of the two contexts. It’s important to note, however, that even this—if accepted—is not (by itself) enough to show that we should reject our prima facie TU-friendly intuitions in the Two Hells context to make them consistent with AU-friendly intuitions in the positive utility context. One could consistently make the opposite adjustment instead. While I do not have enough space here to fully articulate my explanation, the following is my view in a nutshell. First, I believe I have provided strong reasons (throughout this paper) to accept average utilitarianism in the positive utility context. If plausible, these provide strong reasons to maintain AU-friendly intuitions in the positive utility context and to revise one’s TU-friendly intuitions in the Two Hells context—as opposed to making the opposite adjustment. Second, my explanation for the different intuitions in the positive utility context and the Two Hells context operates by showing how person-affecting intuitions are seeping in and how our person-affecting intuitions provide different responses in the context of benefits and in the context of harms. The difference, however, in brief, is that we care more about the effect for a person in the harm context than in the benefit context. Thus, since I have argued that person-affecting intuitions should be rejected, it is our intuitions in the context of benefits—and thus, in this case, the positive utility context—that more closely approximate an impersonal morality, and, thus, it is these that we should maintain. For this reason, our intuitions in the Two Hells context are more suspect, and it is the intuitions in this context that should be rejected. I owe great thanks to an anonymous reviewer for pointing out the need to provide an explanation for why one should revise one’s conflicting intuitions in the manner I suggest we should, as opposed to doing so in the opposite manner. 317 I think that the intuitions of those who claim to find AU’s account of Two Hells (what I’ll call the “Pure Hell” case) incorrect might be wrongfully influenced by AU-friendly intuitions in cases that are slight variations of the “Pure Hell” case. As such, I will argue that people’s non-AU intuitions in the Two Hells case should be held to have less force. First of all, it’s important to notice that even if one subscribes to AU, there are reasons why one might find the total quantity of happiness (or, said differently, the number of individuals in a population) to be relevant to one’s assessment of a state of affairs—albeit indirectly. An important characteristic of a state of affairs for a population is how it is susceptible to change. By change, I don’t mean change by means of a change in welfare for any existing individuals, but a change in value brought about by the addition of individuals with high or low welfare levels. For example, if we know that one hundred individuals will be brought into existence, and thus added to a population, the overall value of the new larger population, all else equal, will remain closer to the value of the original population, the greater the size of the original population. The smaller the original population is, the more the larger population’s value will be influenced by the happiness of the new additions. Thus, while populations of two different sizes that share the same average happiness will be equally good, if one has reason to think that there will be new additions that are, on average, better off than those who currently exist, it will be better for the current population to be smaller. If there is reason to think that the additional individuals will be worse off, then it will be better if the current population is larger. This applies to the Two Hells case, because if there are additional (happy) individuals that would eventually become a part of the population, then this would give us reason to prefer the less populous Hell 1 to the more populous Hell 2. If so, then the prescriptions of AU and TU would converge, as the following table indicates (with the preferred results in bold, and with the 318 additional individuals in italics): Pure Hell AU TU Hell Variation AU TU Hell 2: -10, -10, -10, -10 -10 -40 Hell 2: 10, -10, -10, -10, -10 -6 -30 Hell 1: -11 -11 -11 Hell 1: 10, -11 -0.5 -1 Of course an analogous result would obtain if we compared populations with positive happiness and introduced an additional individual who experienced only suffering. Further, I think that this reasoning also might be what lies behind the beliefs of those who argue that, in cases of populations with positive happiness, there should be a sliding scale in terms of the value of maximizing a population’s total happiness and its per capita average happiness: Consider various populations of different sizes, each of which includes all of a population A’s members, but each of which also includes additional non-shared members. The population with the highest average happiness is not necessarily the one where the average happiness or the total happiness of the non- shared members is highest. Thus, perhaps the assumption of the existence of a background population of additional individuals is also what lies behind these sliding scale views. However, in what follows, I will keep the focus on the implications for our intuitions regarding the hell cases. I think that there are a few reasons why our intuitions in Hell Variation might be seeping into our assessments of Pure Hell. First of all, as a matter of fact, the actual cases we are confronted with in our world—cases of mass cruelty and genocide—are part of a much larger state of affairs in which there is plenty of good. Simply put, these cases which confront us are examples of Hell Variation, not Pure Hell. Therefore, the more people that are treated cruelly, the worse the effect is for the population on the whole. Thus, not only do individuals in agony not comprise the whole population, but the additional individuals tend to have positive happiness, meaning that the better state of affairs is the one in which there are fewer individuals in agony. I think these intuitions seep into people’s assessment of Pure Hell. Similarly, even if we consider a population of only individuals in agony, I think people are 319 still influenced by the fact that in our world, the majority of people are not in agony. Thoughts might turn to additional individuals that might come into existence, and we might think that these individuals are likely to be better off than the original individuals. Thus, it seems as though Pure Hell is not a case that we encounter, whereas we do encounter Hell Variation. As such, intuitions regarding the cases which we are familiar are likely to seep into our analysis of Pure Hell. 2.6. Conclusion to the Two Hells Chapter The Two Hells objection is arguably the strongest objection that has been raised against AU. Notwithstanding this, I’ve argued in this chapter that the Two Hells objection fails to cast doubt on AU. This was for four reasons. First, I argued that my arguments in providing the prima facie case for AU and in responding to the bottom up and top down objections to AU are broad enough to apply to cases of suffering as well, and thus we need not say anything further specifically in response to the Two Hells objection. Second, I argued that zero points don’t exist and the non-existence of zero points prevents the Two Hells objection from even getting off of the ground. Next, I provided two arguments that respond specifically to the Two Hells objection and which show why the Two Hells objection, despite prima facie appearances to the contrary, does not cast any doubt on the plausibility of AU. In sum, in my view, the Two Hells objection does not cast doubt on the plausibility of AU. ====================================================================== Part III Chapter 3. The Top Down Objection to AU 3.1. Introduction This chapter addresses a group of similar objections to AU that I will, collectively, call the “top down objection.” As discussed above, in the introduction to Part III, the top down objection 320 characterizes the problem with AU as follows: 1. Something must be of intrinsic value according to your account. 2. Whatever is of intrinsic value, more of this is better than less of this, since then we have more value. 3. Seemingly happiness is of intrinsic value. 4. Thus, an instance of (or a moment of) happiness is of intrinsic value. 5. All else equal, however, every version of AU seeks to maximize something that is distinct from maximizing instances of happiness. 6. Thus AU theories seek to maximize something other than what has intrinsic value. 7. Thus AU theories are false. Or, less formally, and in prose: While happiness is intrinsically valuable, AU seeks to maximize something else that is not intrinsically valuable, and often at the expense of maximizing total happiness, which is intrinsically valuable. Various authors, including Parfit and McMahan offer forms of this top down objection. Huemer, too, raises this concern with AU, although he frames it slightly differently. 229 He seeks to show that if happiness is what is of intrinsic value, then it follows that average happiness is irrelevant, and maximizing average happiness does not necessarily lead to maximizing intrinsic value. Thus, Huemer takes a step-by-step approach to show why TU follows if happiness is of intrinsic value. I begin, below, in Section 3.2, by taking a step back and making a number of important preliminary clarifications: First, I make a few clarifications about the framework within which my whole dissertation is operating and within which this chapter is also operating: I make clarifications about my frameworks’ assumptions regarding the nature of aggregation and regarding the relationship between the intrinsic value of a whole and the intrinsic value of a whole’s parts. As I say, it is crucial to keep in mind throughout this chapter that the framework within which this chapter is operating is one in which the intrinsic value of a whole is a function of (and supervenes on) the intrinsic values of the whole’s parts, and that the goal of this chapter is to explore the nature of the intrinsic value of these parts. Or, more specifically, the purpose of this 229 Huemer (2008). 321 chapter is to explore the nature of the intrinsic value of the smallest building blocks of an entire state of affairs or of an entire life: the intrinsic value of conscious moments of experience. Second, I make clear that most of the discussions and debates in the literature about intrinsic value (of which there are many) are orthogonal to my inquiry in this chapter (despite ostensibly being about the same concept), and thus, when I use the term “intrinsic value” (or “goodness”), I do not intend to import any meanings or conceptual baggage from any of the various other debates and discussions in the literature. Third, I make a clarifying distinction regarding the meaning of “intrinsic value”: I explain the difference between what I call “intrinsic value” and what I call “contributory value,” and I explain how it is that, despite being different in key respects, the two notions do not appear to be distinguished by others. I then explain the importance of distinguishing between these two notions, and I explain the mistakes that can occur in one’s theorizing if one fails to distinguish between these notions—and I discuss an example of Broome making this mistake. Having made these preliminary clarifications, I then proceed to the meat of the chapter. I start by further explaining the top down objection and by providing a close look at Huemer’s version of it. Having done so, I will show that Huemer and others who offer top down arguments assume their conclusion, and thus that the top down argument is not a successful objection to AU. In short, either directly or indirectly, these arguments assume that instances of happiness have intrinsic value, and thus that the more of these instances that exist, the more intrinsic value there is. If this were the case, I think that there would be a divergence between what AU says is of value and what it seeks to maximize, and I think that this would be a problem. However, the alleged problem with AU as it stands—the divergence between what it seeks to maximize and what is valuable to us—is not due to a failure of AU, but rather it is due to a mistaken assessment of what an AU theorist’s position is (or should be) regarding what has intrinsic value. 322 After I show that the top down argument assumes its conclusion in assuming that instances of happiness are of intrinsic value, there are some readers who might think that I have at that point already successfully responded to the top down argument and that I can conclude the chapter without saying anything further. After all, I will have seemingly shown that the top down argument has a fatal flaw. In my view, however, it will not yet be the case that, at that point, that I have fully (or successfully) responded to the top down argument. In my view, it is a mere first step in responding to the top down objection that I show that the top down argument assumes its conclusion (by assuming that an instance of happiness is of intrinsic value). In order for AU to have a successful reply to the top down objection, it must be shown that proponents of the top down argument are not justified in assuming that instances of happiness have intrinsic value. In order to show this, the AU theorist must offer a plausible account of his own of what has intrinsic value—if not instances of happiness. I think that there are a few different accounts that an AU theorist can offer, and my next section thus will raise and discuss these possibilities. One possible move that an AU theorist could make would be to hold that the only thing that can have intrinsic value is an entire state of affairs, and that, strictly speaking, things such as lives or conscious moments do not have intrinsic value. 230 After addressing this possible account, and arguing that it fails, I next consider a possible account the AU theorist might espouse according to which happiness is only a conditional good—though I argue that this account fails as well. I then consider an account that I call the “happiness-per-life ratio” account, which I argue has potential to be a successful account for an ATLU theorist. I ultimately argue that it is not fully satisfactory either. Despite the fact that I argue that each of these accounts fails, I discuss them 230 One version of this approach would be to maintain that this is the case for the AU theory, but allow that the TU holds that moments or lives can have value. A second version of this approach, however, is for the AU theorist to argue that for a TU theorist, too, it is the entire state of affairs (and not moments or lives) that have intrinsic value. Both of these versions seem potentially plausible. 323 because I think that we gain insights from each of the failures, and these failures ultimately pave the way for finding an account that I do think is successful. In Section 3.5, I then finally describe and explain an account that I do think works: the “happiness-per-moment ratio” view (which I also call the “intensity of moments of happiness” view). After explaining the account, I provide a prima facie case for this account, and this shows that the account works on its own terms and fits the bill at least as a potential contender for an account of intrinsic value that (1) is plausible (and thus is a plausible alternative to an account on which an instance of happiness is of intrinsic value), and (2) is an account that could serve as a foundation for an AU view. At this point, some of the readers who were not satisfied, at the earlier juncture, that I had successfully warded off the top down argument, will now be satisfied that I have. In my view, however, I will not, even at this point, have warded of the top down objection yet. It could still be, at this point, that although the AU account of intrinsic value has been shown to be a plausible contender, that one might think that the TU account is a more plausible account (or, even, a much more plausible account). After all, as the reader will see, there might be some aspects of the “intensity of moments of happiness” view that might be prima facie counterintuitive. 231 Thus, in 231 A key feature of TU’s account of intrinsic value, according to which instances of happiness have intrinsic value, is that if more such instances exist, then there thus is more value because there are more instances of value—i.e. there are more tokens of value. The reader will see that the search for an AU account of intrinsic value will be seeking to articulate a unit which will bring it about that more of that unit will indeed lead to the result that there thus is more value—thus trying to maintain the intuition behind the TU theorist’s position that if something is of value, then more of it should result in there being more value. The investigation that I carry out in this chapter is challenging, because it seems prima facie as though the AU theorist is committed to saying that more instances of a unit of value do not lead to there being more value, and the crux of the top down argument is to make this point and to suggest that it renders AU implausible. The account that I offer on behalf of AU (in Section 3.5) is one that I think works, and I think that it works largely because it is able to get the results that (1) having more of the thing of intrinsic value does indeed lead to more value, and (2) the account is still able to get results consistent with AU. (If the unit of intrinsic value were the instance of happiness, this would successfully enable the AU theorist to satisfy the first requirement, but it would fail to satisfy the second requirement, because the unit would entail the truth of TU and the falsity of AU.) Despite my belief that my AU account works, there might still be some readers who at this juncture will find it to be less plausible than TU’s account, and this might either be because they think, despite my claims to the contrary, it fails to maintain the intuition that more of the unit leads to more value, or, secondly, they might think that there is some other counterintuitive and implausible aspect of the unit that I espouse. For these reasons, it is not necessarily the case, at this juncture, that I can claim to have warded off the top down objection. I will need to provide reasons for the reader 324 my view, I will need to do additional work, even at this point, to ward off the top down objection. To successfully ward off the top down objection, I will need to show that the AU account of intrinsic value is at least as plausible as the TU account of intrinsic value. It is this that I will set out to do next. Thus, in Section 3.6, I set out to assess the relative plausibility of TU’s account of intrinsic value and the account of intrinsic value that I have offered on behalf of the AU theorist. There are five outcomes that could result from this inquiry: 1. TU’s account is much more plausible than AU’s account Here I fail to even weaken the top down argument against AU 2. TU’s account is somewhat more plausible than AU’s account Here I weaken the top down argument against AU 3. AU’s account is similarly plausible to TU’s account Here I succeed in defraying the top down argument against AU 4. AU’s account is somewhat more plausible than TU’s account Here I provide a positive top down argument of my own against TU that has some force. 5. AU’s account is much more plausible than TU’s account Here I provide a strong positive top down argument of my own against TU. Having already taken myself, in Section 3.5, to have offered a plausible prima facie case for my AU account of intrinsic value, option 1 (in the above list) will not actually be a possible result of the inquiry in Section 3.6. Thus, Section 3.6 will determine which of options 2 through 5 is the option that constitutes the result of my inquiry. Options 2 through 5, depending on how, exactly, they are articulated, constitute a successful accomplishment of at least what was accomplished in the options listed above it (since the options listed above any given option x constitute lesser included accomplishments of option x). Ultimately, this chapter will provide an argument for option number 5 (a strong positive top down argument against TU). Thus, this chapter will not only serve as a defensive chapter (attempting to defray the top down argument against AU), but it will also constitute a positive argument against TU. However, if I indeed successfully argue for option number 5, this will also constitute a successful achievement of options numbers 2 through 4 as well, and thus I will have to really believe that the AU account of intrinsic value I offer, and eventually defend, is indeed at least equally as plausible as is TU’s account of intrinsic value. 325 also successfully defrayed the top down argument against AU. It is possible, of course, that the reader will not feel that I have successfully established option number 5. If so, however he might still think that I have successfully accomplished one or more of the other options in the list—and if he thinks that I have accomplished option 3 or 4, he will think that I have successfully (and fully) defrayed the top down objection against AU. 3.2. Introductory discussion about intrinsic value Before delving into the meat of this chapter, I think that there are some preliminary points that will be helpful to make in order to clarify and carve out what it is that I take myself to be exploring in this chapter. All of these clarifications pertain to the question of what I have in mind when I refer to “intrinsic value.” I begin by providing clarifications regarding (1) the nature of aggregation of value, (2) the relationship between the intrinsic value of a whole and the intrinsic value of the whole’s parts, and (3) a disclaimer regarding certain debates that I will not be engaging with. Next, I make a distinction between what I call “intrinsic value” and what I call “contributory value,” and I then also discuss the relationship between these two concepts and the ways in which they can converge or diverge. 3.2.A. Clarifications regarding the nature of aggregation of value and regarding the relation ship between the intrinsic value of a whole and the intrinsic value of a whole’s parts The task at hand in this dissertation is one of aggregation, and it’s important to be clear about what aggregation is and what it involves. As Larry Temkin explains at the beginning of his endeavor to offer a plausible account of aggregation in Rethinking the Good, Aggregation is the process or method of combining different parts into a whole. In ethics, there are principles of aggregation for assessing individual lives. Basically, these tell us how the different elements or components of individual lives combine to affect the goodness, comparative quality, or desirability of those lives. Likewise, there are principles of aggregation for assessing outcomes. Basically, these tell us how the different elements or components of outcomes combine to affect the 326 goodness, comparative quality, or desirability of those outcomes. 232 This type of a framework of aggregation, described by Temkin, is indeed the type of framework that my dissertation is operating within. Thus, what I’m seeking to do in the dissertation, on the whole, is to articulate a plausible account of welfare-based goods (“Theory X”), and, the key task that coming up with such an account involves is coming up with an account of aggregation that explains, among other things, (1) how to determine the value of a whole (e.g., an entire state of affairs or a life) given the values of its parts, (2) how to determine what the relevant parts are, and (3) how to determine the values of the parts. In this chapter, my goal is to carry out a similar task. Rather than attempting, here, to articulate a plausible Theory X, the goal here is to carry explore the accounts of intrinsic value that theories that are candidates for Theory X employ. The key features of aggregation, discussed above, however, will also be features of the framework in which I explore intrinsic value. Thus, the inquiry in this chapter will be the possible accounts of intrinsic value that underlie the framework of aggregation, discussed above, that underlies my entire dissertation. To be more specific, however, and to fix ideas, I now briefly elaborate on some of the assumptions regarding intrinsic value and its aggregation that I will rely upon in this chapter: I will be assuming here that states of affairs can have intrinsic value, and that their intrinsic values are a function of the intrinsic values of their parts. I will be assuming that a whole is a conjunction and that the parts are the conjuncts. Thus, if a whole has some intrinsic value, then there is an explanation of that intrinsic value in terms of carving it into parts, and some of the features of those parts. And how does the whole state of affairs relate to facts of the conjuncts? There will be a straightforward function taking one from the values of the parts to the value of the 232 Temkin, p24. 327 whole. Although one can explain the values of wholes in terms of the parts, we ultimately will get to a sort of bedrock, and this “atom,” or smallest possible part that can have value, will be the intuitive idea of the “building block”—which, on my account, and as argued for in-depth in Part II Chapter 1 is the “conscious moment of experience.” Thus, what I will be exploring in this chapter are what the theories that the candidates for Theory X could articulate as their possible accounts of intrinsic value at the level of the building block (i.e. accounts of intrinsic value at the level of the conscious moment). In other words, I’m seeking to articulate and explore accounts of intrinsic value of the building blocks, which would then constitute the foundation for a broader account of aggregation that would explain the intrinsic value of wholes. I’m seeking an explanatory account of the intrinsic value of the whole and of the parts (building blocks)—where the intrinsic value of the whole supervenes on the intrinsic values of the parts. Without such an account of intrinsic value of the whole and of the parts (building blocks)—and a plausible such account—, it seems that (at least on the framework within which this dissertation is operating) an account of Theory X would not be a plausible one. (Thus, the exploration in this chapter is of key importance for the broader goals of the dissertation of articulating a plausible account of Theory X.) Given that the framework within which I’m working (utilitarianism) is one in which happiness is seemingly what is of intrinsic value, the above goals can be articulated in another way as well: It turns out that, in articulating an account of intrinsic value of a part (i.e. the building blocks, or the conscious moment), we will need to be more precise than merely saying that happiness is what has intrinsic value (which as the reader will see, is the vague articulation that Huemer and various other authors have made and assumed is necessarily what is of intrinsic value for any utilitarian view). Thus, in this chapter, I will be exploring and determining more precisely 328 in what sense it is true (for, say AU and TU) that happiness is what has intrinsic value—at the level of the building block (i.e., at the level of the conscious moment). It is this determination of in what precise sense happiness has value, at the level of the building block, that will constitute the foundation of the theories’ (such as AU’s and TU’s) accounts of intrinsic value more broadly and on the whole (including, also, their accounts of intrinsic value of wholes such as entire states of affairs), and it is these accounts of intrinsic value on the whole that these theories adopt that will ultimately serve as the foundation for their theories more broadly. Thus, the plausibility or non- plausibility of the theories’ (AU’s or TU’s) underlying accounts of intrinsic value at the level of the building block (i.e., the conscious moment) will ultimately be of crucial importance and relevance in assessing the overall plausibility of the theories (AU or TU) on the whole. 3.2.B. Disclaimer regarding intrinsic value and goodness and other debates about these topics In the passage about aggregation from Temkin in the previous sub-section (and in the title of the book from which it is taken, “Rethinking the Good”), Temkin refers to the property of “goodness,” as it pertains to components, lives, and outcomes. Similarly, throughout the introduction to this chapter, and elsewhere in this dissertation, I have referred to and employed both the notion of “goodness” and the notion of “intrinsic value”—and I have often employed these notions interchangeably. What, though, is goodness? And what is intrinsic value? Much has been written about the nature of goodness and of intrinsic value, and various accounts can be given, and the nature of goodness and of intrinsic value can constitute—and, in fact, have constituted—the subject of entire books. For our purposes in this chapter, however, I think that we can sidestep many of the debates because they are, for the most, orthogonal to the debates that I am interested in. To be clear, in this chapter and elsewhere, I will use the terms “goodness” and “intrinsic 329 value” interchangeably, and I do not mean to import any particular assumptions by doing so. Hereafter, I will aim to use the term intrinsic value, but at times I might revert to using the term “goodness” to express this notion. 3.2.C. Intrinsic value versus contributory value I now will discuss a distinction between what I call “intrinsic value” and what I call “contributory value.” This distinction between intrinsic and contributory value does not constitute a debate in the literature—in fact, I cannot say that I have even found this distinction made or discussed in the literature—, but it will help set the foundation for some of the discussion that lies ahead. In light of the fact that this chapter will be concerned with intrinsic value, it will be important that the reader distinguish this notion from the notion of contributory value—and this is particularly the case because, as I will show, the two notions are frequently confused with one another. 3.2.C.i. The distinction between intrinsic value and contributory value As I have argued (and also assumed) at various points throughout the dissertation, while a person’s life or an entire state of affairs can have a certain amount of intrinsic value attributed to it, this value is a function of its component building blocks and their respective intrinsic value. Thus, if the unit relevant for ethical inquiry 233 is the conscious moment, then both AMU and TU will provide accounts of what precisely brings about or constitutes intrinsic value in a conscious moment. It’s important to note, though, that while these respective accounts of value, as attributed to conscious moments, form the building blocks of AU and TU theories, this is not the only way in 233 See Part II Chapter 1 for a discussion of “the unit relevant for ethical inquiry.” 330 which we can understand the value of a conscious moment. In addition to attributing a quantity of intrinsic value to a conscious moment, we can also attribute a certain amount of what I call “contributory value” to the conscious moment. A conscious moment’s contributory value is the amount of value that it adds or subtracts to the value of a certain grouping or collection of conscious moments. In other words, a conscious moment’s contributory value is the difference between the amount of value of: first, a particular grouping of conscious moments in a state of affairs where that grouping would include the conscious moment in question if it does exist, and the conscious moment does in fact exist; and, second, a particular grouping of conscious moments in a state of affairs where that grouping would include the conscious moment in question if it does exist, but where the conscious moment does not in fact exist. For example, it could be that there is a conscious moment, call it “c,” of intensity level five 234 that both an AMU theorist and a TU theorist would say has an intrinsic value of five. However, if we are to assess the value of some grouping of conscious moments, call the grouping “person p,” then there is a question that can be asked about what the contributory value of c would be if it were to be included in the group of moments called person p. To figure out the contributory value of c, we would compare the value of p, where c is included, to the value of p, where p is the same but where c is not included (and, perhaps, where c does not even exist). Similarly, the contributory value of conscious moment c could be assessed for its contribution to the population of a city, a nation, or to a total spatio-temporal state of affairs. Furthermore, just as a conscious moment can have a contributory value (to any particular group), so, too, can a group have a contributory value to any particular larger group. For example, regardless of what one’s theory says about how to calculate the intrinsic value of a person’s life (and regardless of what this value is determined to be), this person’s life would have a 234 I will loosely refer here to “intensity level five,” but I will explore this notion in greater depth later in this chapter. 331 contributory value for its contribution to larger groups—such as the population of a city, a nation, or a total spatio-temporal state of affairs. Of course, what this contributory value is determined to be will depend both on how much intrinsic value is attributed to the person’s life, and also on what function one’s theory uses to determine the value of the broader group from the intrinsic values of its components, sub-components, or whatever combination of this data is deemed relevant (including, perhaps, the length of each specific individual’s life). 235 3.2.C.ii. Whether and how there can be a divergence between an object’s intrinsic and contributory value The question that might arise next, though, is whether there can be any difference between an object’s intrinsic value and its contributory value. The answer is that there can be, and whether or not there is a difference is a function, largely, of one’s aggregation mechanism of choice—be it TU-type aggregation or AU-type aggregation. 236 Not surprisingly, for a TU theorist, the intrinsic value of conscious moment c (namely five) will be equal to the contributory value of c to any group of which it might be a part—be it a 235 To clarify what I have said in this paragraph: my claim that a group (such as a person) can itself have contributory value to a larger group does not rely on the claim that the unit relevant for ethical inquiry is the person. Even someone who (like myself) thinks that the unit relevant for ethical inquiry is the conscious moment, can think that a person can have both a certain amount of intrinsic value and also various contributory values (for the groups to which it contributes). Of course, those who hold different units to be relevant to ethical inquiry might also have different methods for determining the intrinsic value of a person’s life, but, by the same token, even those who agree about what is the unit relevant for ethical inquiry—be it the conscious moment or be it the person—can have different methods for determining the intrinsic value of a person’s life. 236 Before proceeding here, however, a clarificatory prefatory note is in order: It should be clarified that, strictly speaking, AU theorists and TU theorists are only committed to holding, respectively, that AU-type and TU-type aggregation methods are the methods for determining agent-neutral goodness—thus for aggregating value from the unit relevant for ethical inquiry to the value of a total state of affairs. I am thus making two simplifying assumptions in this discussion. First, I am assuming here—though I argued for this claim in Part II Chapter 1—that the unit relevant for ethical inquiry is the conscious moment. Second, although TU theorists are likely to employ TU-aggregation for assessing the goodness of groups of conscious moments that are a proper part of a total state of affairs (e.g. a person’s life), and though AU theorists might be inclined to employ AU-type aggregation for assessing groups of this sort, these are not views that must be held by TU theorists and AU theorists respectively. AU and TU are agent-neutral theories about the unit relevant for ethical inquiry’s relation to the total state of affairs, and they do not indicate what the mechanism is for assessing the goodness of groups of intermediate levels of abstraction. However, in order to simplify this discussion, I will assume that AU theorists use AU for the aggregation mechanism from conscious moments to any group that is a proper part of the total state of affairs, and, conversely, that TU theorists use TU as the aggregation mechanism from conscious moments to any group that is a proper part of the total state of affairs. 332 person, the population of a city, a nation, or a total spatio-temporal state of affairs. In each of these cases, the contributory value of c will be five. Similarly, for the TU theorist, the intrinsic value of a group with an intermediate level of abstraction, such as a person’s life, will be the same as its contributory value to larger groups such as a total state of affairs. For an AU theorist, however, these claims do not necessarily hold true. For an AU theorist, while the intrinsic value of c will be five, the contributory value of c to various groups might be five, but it could also be any other sum—be it positive, negative, or zero. Furthermore, the contributory value for c may differ for the different groups for which its contribution value is assessed. Similarly, for an AU theorist, the intrinsic value of a group with an intermediate level of abstraction, such as a person’s life, will not necessarily be the same as its contributory value to larger groups, such as a total state of affairs. Thus, what this means is that a TU theorist might refer to the value of a conscious moment, and without further specification, and depending on the context, this could refer to the intrinsic value of the conscious moment or it could refer to the contributory value it has for one of various larger groups. Regardless of which is meant, though, the numerical value will be the same. Thus, while this might make things simpler in one way, it leaves open the possibility for confusion and equivocation. Furthermore, this equivocation and or confusion is particularly likely to be troublesome precisely because of the fact, on an AU-type account, the intrinsic value of a moment and its contributory value to various larger groups will only be the same in rare cases. While, for TU theorists, the distinctions between intrinsic value and contributory value to various larger groups are smoothed over (because they all have the same numerical value), this is not the case for AU. Thus, while this imprecision and equivocation might be harmless if one were only addressing TU theorists, it causes confusion if the discussion is broad enough in scope to include AU theorists. Furthermore, this is precisely the case with our topic of discussion, because the question of what has intrinsic value is a question that is meant to be an important question regardless of 333 which aggregation mechanism one espouses. Thus, absent a decision about which aggregation mechanism to espouse—and, specifically, absent a decision in favor of TU-type aggregation—, it is necessary to distinguish between intrinsic and contributory value. The foregoing confusion can wreak havoc with one’s theorizing in various ways (and one example of this confusion wreaking havoc with one’s theorizing can be found in one of Broome’s discussions in Ethics Out of Economics 237 ), so it’s important, going forward, to be clear about the 237 While the confusion, just described, can and does occur in various contexts and in various ways, I will now briefly import (from Part IV Chapter 3 of this dissertation) the discussion of one example of this confusion in Broome’s work: (See the discussion there for further explanation of the relevant concepts and terms.) It seems to me that this confusion, just described (and its instantiations that I will describe in this section) which results from the TU theorist not needing to distinguish between intrinsic and contributory value, is precisely what is underlying the lack of precision in Broome’s discussion about the reducibility of goodness to betterness. Because he implicitly assumes the plausibility (or truth) of a TU-type aggregation framework, and because he thus is working within a TU-type aggregation framework, intrinsic value always equals contributory value, and thus the “higher order reducibility claim” and the “same order reducibility claim” appear to be interchangeable. As discussed, however, in and of themselves they are not interchangeable, because, while they might appear interchangeable in the TU-type aggregation framework, the fact that they are not, in and of themselves, interchangeable is clear when one sees that they can potentially diverge if one’s aggregation mechanism is AU-type aggregation. Broome’s confusion and lack of precision is apparent when he goes back and forth between using the same order reductionist position and the higher order reductionist position without distinguishing between them. As discussed, this can cause confusion if we are not necessarily exclusively in a TU framework, because it might not be clear in a particular instance whether referring to the value of, say, life x, is referring to its intrinsic value or its contributory value (to, say, a total state of affairs), and if we are not in a TU framework, these values might differ. Further, there are specific examples in Broome’s discussion of where this treatment of the higher order and same order versions of the reductionist claim could be problematic: Consider, for example, Broome’s term “goodness of a life,” or “goodness of a person being alive,” or what we could also, perhaps, call a “life worth living.” It seems as though all of these terms could be used to refer to contributory goodness to a total state of affairs, but these terms could also be used to refer to intrinsic goodness of a particular life. Thus, since the same order reducibility claim matches up with intrinsic goodness and since the higher order reducibility claim matches up with contributory goodness, Broome’s assumption of a TU framework is what leads to the higher order version and the same order version of the reductionist claim seem interchangeable. Interestingly, the foregoing confusion is also closely related to another one. Instead of there being two notions that are being smoothed over and both referred to with one term (such as “goodness of a life”) by Broome or others who assume a TU framework, there actually are three. Discussed in the previous paragraph were, on the one hand, the notion of intrinsic value of a life, and, on the other hand, the notion of contributory value of a life. Both of these notions of value, however, were being understood in an agent-neutral sense. However, the third notion that also gets described with the same term (“goodness of a life”) is the agent-relative value of a life—the goodness of the life for the person who lives it. In the TU framework, this agent-relative value would be the same as both the agent-neutral intrinsic value of the life and also the agent-neutral contributory value of the life. Thus, in light of the fact that, on TU, agent-relative value will be quantified with the same sum as will both intrinsic agent-neutral value and contributory agent-neutral value, there also becomes an assumption that agent-neutral and agent-relative values of lives would be the same. On an AU framework, however, not only need not intrinsic agent-neutral value and contributory agent-neutral value be equal, but agent-relative value also need not equal either intrinsic agent-neutral value or contributory agent- neutral value. Of course, this also depends upon what one’s account is of how agent-relative goodness is calculated. However, if AU is the agent-neutral aggregation mechanism, various agent-relative aggregation mechanisms (certainly TU-type aggregation, and perhaps others as well, depending on the specifics of the account) could result in cases where one might think that it is good (in an agent-relative sense) for person x to exist, yet bad (in an agent- 334 distinction between intrinsic value and contributory value and to be clear about the ways in which they relate to each other. 3.2.D. Summary regarding the introductory discussion about intrinsic value At this point, we are now hopefully better equipped not only to recognize the differences between the various views that Broome and others tangle up, but hopefully with this knowledge and with knowledge of what leads to the tangling, we will be better able to keep the various views and notions separate. Not only is it important to keep separate discussions of agent-relative and agent-neutral values, but (and most importantly for our purposes) it is crucial to recognize that intrinsic value and contributory value are different and thus address them separately. Recognizing this and addressing intrinsic and contributory value separately will provide us with much greater clarity as we proceed. Additionally, not only is it important, going forward, to keep in mind the intrinsic value versus contributory value clarification, but it is also important to keep in mind the other clarifications I made at the beginning of this section: First, it’s important to be clear about the fact that, even though this chapter is about intrinsic value, in this chapter I will be bracketing a lot of debates and discussions about intrinsic value and goodness that occur in the literature because, despite being ostensibly about the same topic (i.e., “intrinsic value”), many of these debates are actually orthogonal to my project and inquiry in this chapter. Thus, not only do I not engage with these various debates and discussions, but I also do not intend to import any meanings or other conceptual baggage regarding the terms “intrinsic value” or “goodness” from these debates and discussions when I use these terms. neutral sense) for person x to exist. For further, and in depth, analysis both about various accounts of agent-relative goodness and about how these agent-relative accounts might relate to accounts of agent-neutral goodness, see Part II Chapter 3 of this dissertation. 335 Second, it’s also important to keep in mind the preliminary points I made about the nature of aggregation and about the relationship between the intrinsic value of a whole and the intrinsic value of a whole’s parts. It is crucial to keep in mind the fact that the framework within which this chapter is operating is one in which the intrinsic value of a whole is a function of the intrinsic values of the whole’s parts, and that the goal of this chapter is to explore the nature of the intrinsic value of these parts. Or, more specifically, the purpose of this chapter is to explore the nature of the intrinsic value of the smallest building blocks of an entire state of affairs or of an entire life: the intrinsic value of conscious moments of experience. 3.3. The top down argument against AU 3.3.A. The top down argument against ATLU 3.3.A.1. Introduction As is the case for the bottom up argument against AU (discussed in Part III Chapter 1), although the top down argument is an argument broad enough to be used against any version of AU, the versions of it found in the literature are all versions that are offered as objections to one specific form of AU—namely ATLU. 238 The reason for this is fairly straightforward: As discussed in Part 1 Chapter 2, aside from a few examples here and there, the vast majority of authors simply assume that ATLU is the only version of AU that exists. Furthermore, among those who recognize that other alternatives exist, it is either explicitly or implicitly assumed that ATLU is the most plausible—or least implausible—version of AU, and certainly the only one meriting consideration. In light of this, I will begin by describing the top down argument against ATLU in particular, since this is the most common version of the top down argument. The top down argument against ATLU comes in different sub-types as well, and Parfit, 238 Each of these (including ATLU) could, in turn, be further precisified by stipulating which temporal population is relevant. For an in depth discussion of these different possible sub-views, see chapter 6. 336 McMahan, and Huemer all offer slightly different versions of the top down argument against ATLU. In what follows, however, I will describe and discuss Huemer’s version of the top down argument against ATLU. Section 3.3.A.2. Huemer’s top down argument The general structure of the top down argument against ATLU is to argue as follows: 1. Assuming that we accept a utilitarian framework, happiness is the only thing that is intrinsically valuable. 2. Thus, average per capita happiness is not of intrinsic value. (from 1) 3. Maximizing average per capita happiness does not always maximize happiness. 4. Thus, ATLU, which seeks to maximize average per capita happiness, prescribes maximizing something other than that which is intrinsically valuable. (from 2 and 3) 5. Thus, ATLU is false. (from 4) Huemer’s argument follows this form as well, but he (correctly) recognizes that the truth of steps 1 and 2 are the only controversial steps, and he thus focuses his efforts on the truth of two different claims. First, he (fairly briefly) strives to show that we do and should believe step 1. Next, he spends the bulk of his discussion arguing for the claim that “if step 1 is true, then step 2 is true,” or, more specifically, “if happiness is the only thing that is intrinsically valuable, then average per capita happiness is not of intrinsic value.” Huemer believes that, if he successfully argues for his two claims, then he will be able to show that average happiness is irrelevant, because maximizing average happiness does not necessarily maximize intrinsic value, and, thus, the falsity of ATLU (and the truth of TU) will follow. As for his claim that step 1 is true, while Huemer doesn’t say much in defense of this claim, he does offer two points. First, he says, it seems undisputed by all utilitarians that happiness is what is of intrinsic value. Second, he says it is both commonly thought and also plausible that “it is good, all else equal, to maximize the sum of happiness in a single life.” This, he says, provides us with evidence that we think that happiness is what is of intrinsic value. Thus, in light 337 of these considerations, Huemer takes it as self-evident that happiness—and happiness alone—is what is of intrinsic value (for the utilitarian). As for Huemer’s claim that it is true that “If happiness is the only thing that is intrinsically valuable, then it follows that average happiness is not of intrinsic value,” the bulk of his argument is devoted to establishing the truth of this conditional. In order to do so, Huemer offers what he calls the “Equivalence Argument.” This argument aims to show that if one believes that it is good, all else equal, to maximize the sum of happiness in a single life, then one is committed to believing that we should maximize total happiness in a population, regardless of how low the average happiness in the population is. I laid out and discussed Huemer’s Equivalence Argument, at length, in Part II Chapter 1, during the course of making my argument for the claim that the unit relevant for ethical inquiry is the conscious moment, and, thus, rather than describing and analyzing (and critiquing) Huemer’s Equivalence Argument again here, I instead will simply direct the reader to my text in Part II Chapter 1 Section 1.4.A.iii. See also Section 1.4.A.ii (in Part II Chapter 1) for my background discussion regarding Huemer’s framework of “equivalences” and “dimensions.” Thus, in sum, and in language I introduced in Part II Chapter 1 of this dissertation, Huemer attempts to do two things in his argument: His second claim (that if step 1 is true, then step 2 is true) amounts to the claim that the conscious moment is the unit relevant for ethical inquiry—and thus that the unit relevant for ethical inquiry is not the person. His first claim is that a moment of happiness is what is of intrinsic value. Putting these two claims together, we get the conclusion that TU is true and that ATLU is false. 3.3.B. Assessing Huemer’s top down argument against ATLU Despite the fact that both of Huemer’s claims might seemingly be plausible, in my view, 338 both claims fail. 3.3.B.1. A first response to Huemer’s argument: The problem with Huemer’s second claim: Huemer assumes his conclusion As for Huemer’s second claim (that “if happiness is the only thing that is intrinsically valuable, then average per capita happiness is not of intrinsic value”), I lay out my argument for why Huemer’s claim (via his Equivalence Argument) fails, in Part II Chapter 1. Rather than carry out that analysis again here, I direct the reader to Part II Chapter 1 Section 1.4.A.v. In short, however, I argue that Huemer’s Equivalence Argument (which is essentially an argument for the position that conscious moment as the unit relevant for ethical inquiry) relies on his (explicit) assumption that “the intrinsic value of an event supervenes on its intrinsic, non-evaluative properties.” Although I, too, believe that the unit relevant for ethical inquiry is the conscious moment and I am thus sympathetic to the conclusion of the Equivalence Argument, Huemer’s argument against ATLU is not persuasive, and this is because it assumes its conclusion—or, slightly more precisely, the target of his argument, the ATLU theorist, will not accept this assumption that he makes. And, importantly, his argument doesn’t work without this assumption. Or, to boil his Equivalence Argument down to what I believe it boils down to, and to put it in nakedly question-begging terms: We can understand Huemer to be saying the following (in terminology I have coined—e.g., the “unit relevant for ethical inquiry”): “If the unit relevant for ethical inquiry is the conscious moment, then ATLU must be false.” As should be clear, Huemer’s argument assumes its conclusion: The assumption that is required for the argument to follow is one that his opponent (the ATLU theorist) will not accept. Huemer assumes that the unit relevant for ethical inquiry is the conscious moment, and it is the truth of this claim that he is arguing 339 for. 239 Despite the fact that the ATLU theorist can offer the response I just discussed in response to Huemer’s argument, I will move forward assuming that Huemer is right to assume that “if step 1 is true, then step 2 is true,” or, more descriptively, “if happiness is the only thing that is intrinsically valuable, then average per capita happiness is not of intrinsic value,” and this is for three reasons: First of all, as I said, I ultimately think that Huemer’s position on the matter is correct, even if he doesn’t offer a sufficient argument to support his position. Thus, I don’t think that this avenue is a particularly plausible line of response for the ATLU theorist who is attempting to ward off the top down objection. Second of all, I don’t think that the ATLU theorist needs to take this avenue of response, because there is a different avenue of response that is a significantly more fruitful one to pursue. Third of all, while the current avenue of response is a response that the ATLU theorist can make to Huemer’s top down objection, as I will explain below, this avenue is more of a response to a specific aspect of Huemer’s top down objection against ATLU, as opposed to being a response to what I will explain to be the core aspects of a more general top down objection to AU. Thus, I will focus my efforts more on articulating a response to the more general version of the top down argument. Thus, without further ado, I will turn to the assessment of Huemer’s first claim (that step one is true, or, more descriptively, that happiness is what is of intrinsic value for a utilitarian). It is the denial of this step that I think is the more fruitful avenue of response for the ATLU theorist. Additionally, I believe that denying this step will end up being an avenue of response that can and should be given against the top down argument against AU, more generally, and not merely 239 As I made clear in Part II Chapter 1, however, I do think that the goodness of an event is a function solely of its intrinsic features, and thus I agree with Huemer’s claim that the unit relevant for ethical inquiry is the conscious moment. Huemer, however, simply assumes his conclusion, and he does not provide any reason at all for the ATLU theorist to be persuaded that he should accept that the conscious moment is the unit relevant for ethical inquiry. In Part II Chapter 1, I provide arguments for why one should think that the conscious moment is the unit relevant for ethical inquiry, and I think that the reasons that I offer there would provide an ATLU theorist strong reason to question his position about the unit relevant for ethical inquiry. 340 against the top down argument against ATLU. 3.3.B.2. A second response to Huemer’s argument: The problem with Huemer’s first claim: Huemer assumes his conclusion again. Huemer claims that happiness is what is of intrinsic value for a utilitarian, and he supports this point by claiming (1) that it is undisputed by all utilitarians that instances of happiness are of intrinsic value, and (2) that the fact that it is both commonly thought and also plausible that “it is good, all else equal, to maximize the sum of happiness in a single life” is evidence that we think that instances of happiness are of intrinsic value. In light of these considerations, Huemer takes it as self-evident that instances of happiness—and this alone—are what is of intrinsic value (for the utilitarian). A second possible route for the ATLU theorist to respond to Huemer’s argument is to deny that a moment of happiness itself is of intrinsic value. If Huemer were right that a moment of happiness had intrinsic value, then it would seem that he would be right to conclude that the more instances of intrinsic value that exist, the more intrinsic value there is. If this were the case, I think that there would be a divergence between what ATLU says is of value and what it seeks to maximize, and this would be a fatal problem for the theory. However, the alleged problem with ATLU as it stands—the divergence between what it seeks to maximize and what is of intrinsic value for the ATLU theorist (or, more generally, for us)—might not be due to a failure of ATLU, but, rather, it might be due to our mistaken assessment of what has intrinsic value for the ATLU theorist (or, more generally, for us). A new account of intrinsic value could potentially eliminate this divergence, and thus it could prevent Huemer’s argument from getting off the ground, because it would provide grounds to reject step one of his argument. So far so good, but the question then becomes what an ATLU theorist might espouse as an alternative account of what has intrinsic value. In light of the fact that it does seem to be 341 commonly thought (by utilitarians) that instances of happiness are intrinsically valuable, what reason does the AU theorist have for rejecting the assumption that instances of happiness are intrinsically valuable? I will explore this question in depth in Section 3.4, but one possibility might be that the happiness-per-life ratio is what is of intrinsic value, and another possibility might be that happiness only has conditional intrinsic value. How to understand these concepts and whether they will turn out to be plausible accounts will be explored, but suffice it to say that a seemingly promising route for the ATLU theorist to respond to the top down argument is to deny step one, and to thus claim that something other than a moment of happiness is what is of intrinsic value. 3.3.B.3. It thus seems to me that there are two points that the ATLU theorist can (and must) make in response to the top down argument against ATLU. He will reject the claim that the conscious moment is the unit relevant for ethical inquiry (which Huemer argued for in his Equivalence Argument), and he will also reject the claim that a moment of happiness is of intrinsic value. The ATLU theorist will claim that the proponent of the top down argument assumes his conclusion in making both claims. As I will now show, however, the claim about what constitutes the unit relevant for ethical inquiry is not a claim that is part of the top down argument against AMU, or, thus, what I’ll call “the core top down argument against AU.” Thus, for the most part, for the purposes of this chapter, I will assume that ATLU has a defensible claim regarding the unit relevant for ethical inquiry—even though I myself do not believe that ATLU has selected the correct unit relevant for ethical inquiry. Notwithstanding my view on this matter, I will strive to defer to ATLU on this matter, because, as will become clear, what I want to call the “core top down argument against AU” does not address the specific claims that specific versions of AU make about what constitutes 342 the unit relevant for ethical inquiry. Thus, when I refer to whether or not ATLU has a successful response to the top down objection, this will refer to whether ATLU is correct to say that the proponent of the top down argument assumes his conclusion in asserting Premise 1—that a moment of happiness is of intrinsic value. While I ultimately do reject ATLU and I do not find it plausible, it will become clear in Sections 3.4 and 3.5 that this is not because I think it is unable to response to the top down argument. I will argue that ATLU is able to respond successfully to the top down objection. Rather, as I argue, its failures stem from its incorrect view about what constitutes the unit relevant for ethical inquiry. 240 3.3.C. The top down argument against AMU As I have shown, Huemer offers a top down argument against ATLU. As I stated in the beginning of this chapter, however, what I call the “top down argument” is actually an argument that not only can be made against ATLU, but it is an argument broad enough to be made against all versions of AU. In light of the fact that Huemer’s top down argument was an argument made against ATLU, however, it might not be completely obvious which portions of that argument are parts of a general top down argument against AU and which portions of that argument are parts that are specifically tailored to ATLU—and thus not part of what I will call the core of the top down argument against AU. In order to make headway on this front, I now proceed to explore what a top down argument would look like if it were tailored to and made against AMU. As the reader should recall, AMU and TU have some things in common. Many commonalities exist—such as the fact that both are moral theories and that both are impersonal moral theories—but the commonality I wish to focus on here is the following: Unlike ATLU, both AMU and TU hold that the unit relevant for ethical inquiry is the conscious moment (as opposed 240 And, of course, the reasons that I do think are good reasons for rejecting ATLU (i.e., reasons other than the top down argument) are the subject of Part II of the dissertation. 343 to the person). Thus, the top down argument that Huemer would raise against the AMU theorist would not need to include the Equivalence Argument, which aimed to show that the unit relevant for ethical inquiry should be the conscious moment. Both AMU and TU would agree that the Equivalence Argument shows that worlds G and H are equally good in light of the fact that they are each composed of two components and all four components are equally good. The difference between the two positions (AMU and TU) arises, however, when calculating (i.e. aggregating) how much value is in the worlds. Huemer ultimately concludes that the value of world G is VG1 + VG2, and that the value of world H is VH1 + VH2. An AMU theorist, however, would conclude that the value of world G is the average of VG1 and VG2 and that the value of world H is the average of VH1 and VH2. Huemer arrives at this conclusion because of his belief that Premise 1 from his top down argument against ATLU is true: More descriptively, he arrives at this conclusion because he thinks that, assuming that we accept a utilitarian framework, a moment of happiness is what is of intrinsic value. Thus, to go back to what Huemer’s top down argument would look like if it were made against AMU (instead of ATLU), the only substantive 241 difference would be that Huemer would not need to include Premise 2, and, thus, he would not need to include the Equivalence Argument, which would (if successful) get him from Premise 1 to Premise 2 in the top down argument against ATLU. Other than this one substantive difference, however, the top down argument against AMU would be substantively the same as the top down argument against ATLU, and the same points 241 There would also, however, be a few “non-substantive” differences. These “non-substantive” differences would include making the appropriate changes to the argument so that it would be geared to AMU and not to ATLU. These changes would changing “per capita happiness” to “per conscious moment happiness” in Step 3, changing “ATLU” to “AMU” and “capita” to “conscious moment” in Step 4, and changing “ATLU” to “AMU” in Step 5 (i.e. the conclusion). 344 would apply in this argument. Huemer would say that it seems that the AMU theorist would hold that a moment of happiness is of intrinsic value, but that if this is the case, then more (happy) conscious moments, all else equal, should bring about more value than would fewer such moments. In response to this argument, AMU is faced with essentially the same situation as the ATLU theorist faces when confronted with the top down argument against ATLU. 242 However, assuming that the ATLU theorist chooses to object to Huemer’s Step 1 (and not to Huemer’s Step 2), AMU and ATLU are not only in essentially the same situation with respect to the top down argument, but they are in the identical situation with respect to the top down argument. Both theories will have to deny Huemer’s claim that instances of happiness are of intrinsic value. The problem that both of these AU theories confront, however, is that it does seem—both prima facie and after even some thought—that instances of happiness are of intrinsic value to all versions of utilitarianism! After all, what else could the utilitarian think is of intrinsic value? 3.3.D. What the top down argument against ATLU and the top down argument against AMU have in common, and thus a general and intuitive explanation of the core of the down objection against AU The commonalities between the top down argument against ATLU and the top down argument against AMU have now been alluded to, but how can we best articulate the core point of the top down argument against AU? The proponent of the core top down argument against AU will say the following: All AU views do not necessarily take the same unit as the unit relevant for ethical inquiry. 242 Unlike the ATLU theorist, however, AMU will not object to Huemer’s move from Step 1 to Step 2 (which is supported with the Equivalence Argument). After all, the top down argument against AMU does not include Step 2, and thus it does not appeal to the Equivalence Argument. If Step 2 (and thus the Equivalence Argument) had been included, however, the AMU theorist would not have objected to it. 345 On the one hand, ATLU holds that the unit relevant for ethical inquiry is the person, and it attributes a certain value to a person’s life (by using TU-type aggregation to sum the total amount of happiness across all of a person’s time-slices). On the other hand, AMU holds that the unit relevant for ethical inquiry is the conscious moment, and it attributes a certain value to a conscious moment. Despite the differences among AU theories with regard to what they take as the unit relevant for ethical inquiry and or how they determine the value of this unit, they all share the following feature: Despite articulating that this unit has a particular value, the goodness of a state of affairs is determined by averaging the goodness of each unit, instead of following TU’s approach of aggregating the value of the components summatively. Thus, although AU attributes a particular amount of value to one instance of a unit relevant for ethical inquiry, AU seemingly mysteriously (and certainly implausibly) does not think that there will be more value, all else equal, if more such units exist. (The labels of “mysterious” and “implausible” are according to the proponent of the core top down argument against AU.) How, though, could this be, the proponent of the top down argument against AU says, if the unit in question (be it a person’s life or be it a conscious moment) is the bearer of a certain amount of (positive) intrinsic value? According to the proponent of the top down argument against AU, if an instance of the unit in question bears a certain amount of (positive) intrinsic value, the existence of more such instances should increase the amount of value that exists. AU, however, seemingly mysteriously disagrees. Not only are the various versions of the top down argument against AU the same in the above-described ways, but it also seems, at least prima facie, that the top down argument has both the same type of force and also the same extent of force in each context. There is another way of illustrating the above points, and it is by referring to the chart of dimensions and equivalences, which I provided in Part II Chapter 1 in the context of my 346 discussion of Huemer’s Equivalence Argument, and which I reproduce here: Chart of dimensions and equivalences: 243 Intensity Duration Widespreadness Equivalences Non-Equivalences AMU: Matters Doesn’t matter Doesn’t matter D-W I-D, I-W ATLU: Matters Matters Doesn’t matter I-D I-W, D-W TU: Matters Matters Matters I-D, I-W, D-W None What AU views have in common is that they maintain that the dimensions within the unit relevant for ethical inquiry do matter, but that the dimensions outside of the unit relevant for ethical inquiry do not matter. For example, for AMU, according to which the conscious moment is the unit relevant for ethical inquiry, intensity does matter (because it is within the unit relevant for ethical inquiry), but duration and widespreadness do not matter (because they are outside of the unit relevant for ethical inquiry). Similarly, for ATLU, according to which the person’s life is the unit relevant for ethical inquiry, intensity and duration do matter (because they are within the unit relevant for ethical inquiry), but widespreadness does not matter (because it is outside of the unit relevant for ethical inquiry). Said differently, and as can also be seen in the above chart, AU theories hold that there are equivalences between dimensions that are on the same side of the divide of “within” and “outside of” the unit relevant for ethical inquiry, but that there are non- equivalences between dimensions that are on opposite sides of this divide. For TU, on the other hand, all of the dimensions matter, and this is so regardless of whether they are within the unit relevant for ethical inquiry (intensity) or outside of the unit relevant for ethical inquiry (duration and widespreadness). Having now used these terms to make the foregoing descriptive statements about the difference between AU and TU, the proponent of the top down argument against AU can also couch the top down argument against AU in these terms as well. It states: 243 For definitions and thorough explanations of the various terms in the chart, see the text in Part II Chapter 1 where the chart is introduced for the first time. 347 To be consistent, all dimensions should matter, and this is the case regardless of whether they are within or outside of the unit relevant for ethical inquiry. More instances of a unit relevant for ethical inquiry (with positive intrinsic value) means that there must be more intrinsic value. 3.3.E. Conclusion of the first section on the top down argument against AU—a section which has (1) described the argument, (2) shown the problems that the argument faces and preliminarily assessed the argument, and (3) shown what the AU theorist’s task must be in refuting the argument Up until now, I hope to have described the top down argument against ATLU and the top down argument against AMU, and I hope to have identified what the common thread between these two arguments is that makes them versions of what I call the “core” top down argument against AU. It is this “core” top down argument that I seek to explore in the remainder of this chapter. As should be clear by now, I have identified one potential flaw with the top down argument against AU. Since this flaw is currently only a potential flaw, I can call it the “weak spot” of the argument—or the step of the argument that the AU theorist should concentrate his efforts to put pressure on and reject. This weak spot of the argument is the assertion of Premise 1: The claim that an instance of (or a moment of) happiness is of intrinsic value. The AU theorist should deny this premise and maintain that the proponent of the top down argument is assuming his conclusion. 244 244 While I say that AU theorist should deny Premise 1 and maintain that the proponent of the top down argument is assuming his conclusion, there are actually two different ways in which my claim can be understood, and how one understands my claim depends on whether or not one thinks that the claim that “happiness is of intrinsic value” implies the claim “a moment of happiness is of intrinsic value.” On the one hand, if one thinks that the claim “happiness is of intrinsic value” implies the claim “a moment of happiness is of intrinsic value,” then my response to the top down argument is that it is assuming its conclusion by saying “happiness is of intrinsic value.” On the other hand, if one thinks that the claim “happiness is of intrinsic value” does not imply the claim “a moment of happiness is of intrinsic value,” then my response to the top down argument is not that it assumes its conclusion when it states that “happiness is of intrinsic value,” but, rather, that it then proceeds to assume its conclusion when it assumes that because “happiness is of intrinsic value,” it is the case that “a moment of happiness is of intrinsic value.” 348 As I say, however, my claim that the top down argument assumes its conclusion in stating that a moment is of intrinsic value will only be plausible if I, on the behalf of the AU theorist, can offer a plausible alternative account of intrinsic value. After all, prima facie, it does seem as though a utilitarian will and should say that an instance of happiness is of intrinsic value. Thus, absent my showing a plausible alternative, what I condemn as the top down argument assuming its conclusion would simply amount to it making a reasonable assumption. 245 Thus, at this point, I have not found a flaw with the top down argument. I have merely identified its weakest link. What remains to be discovered is whether this “weakest link” does in fact amount to a flaw. Thus, at this point in the dialectic, it becomes the onus of the AU theorist to provide a plausible alternative account of intrinsic value. One further note on the following: What’s relevant (for the most part) in this chapter is not a debate between ATLU and AMU, but whether AU as a general theory is able to respond successfully to the top down argument. So, for the most part, one can just proceed through this chapter with either ATLU or AMU in mind when I refer to AU (and one’s choice about which one to keep primarily in one’s mind should be a function of whether one believes that the more plausible candidate for the unit relevant for ethical inquiry is the conscious moment or the person). Thus, the first of these two options amounts to stating that “happiness is of intrinsic value” is not an ambiguous claim, and that it assumes its conclusion. The second of the two options amounts to stating that “happiness is of intrinsic value” does not assume its conclusion, but that it is an ambiguous statement, and thus the assuming of the proponent of the argument’s conclusion comes one step downstream from this uncontroversial first premise. From hereon out, however, there is no need to distinguish between these two readings of the argument, and, similarly, of the two readings of my response to the argument. The differences between the two readings are merely definitional, and I will refer to them interchangeably in the remainder of this chapter. 245 As I stated in the introduction to this chapter, there might be some readers who disagree with me about this and who would thus think that I have already, even at this point in the chapter, sufficiently warded off the top down without need for any further discussion at all. In my view, however, this is not true, and I must indeed make further arguments in order to ward off the top down argument. One view that a reader, who thinks I can end at this point, might make is that the AU theorist should just maintain that what is of value is an entire state of affairs (and not its parts), and thus, he might say, nothing further needs to be shown, and we can conclude that the top down argument assumes its conclusion. Even this response, however, would require an additional showing beyond where we are at this point: The proponent of this response would need to provide this very response (that the entire state of affairs is what has intrinsic value), but we have not yet gotten to the AU theorist’s opportunity to provide his own account of what has intrinsic value. In fact, this very view (that only an entire state of affairs can have intrinsic value) is one of the views I consider in Section 3.4. So even the proponent of this view would have to concede that, at this juncture here, the AU theorist has not done enough yet to ward off the top down argument—even if he later will be able to. 349 There will be some portions of the chapter, however, that explicitly address either ATLU or AMU, and, further, in Sections 3.4.C and 3.5, I will specifically compare and contrast the relative strength and plausibility of the top down arguments against ATLU and AMU, respectively, and of the responses that can be offered by ATLU and AMU, respectively. Absent these explicit specifications, however, the remainder of the discussions of this chapter are meant to apply to the core top down argument—which is the general argument that afflicts both ATLU and AMU. 3.4. Possible positive AU accounts of intrinsic value The first task for the AU theorist who attempts to respond to the top down argument against AU is to show that the top down argument assumes its conclusion. So far so good. This task was accomplished in the previous section. The next step for the AU theorist, however, is to show that proponents of the top down argument are not justified in assuming that instances of happiness have intrinsic value. After all, it seems to be commonly assumed among utilitarians that this is a reasonable assumption. In order to show that this is an unjustified assumption, the AU theorist must offer a plausible alternative account of his own of what has intrinsic value—an account that posits something that is of intrinsic value other than instances of happiness. Absent an alternative account of this sort, it seems that the top down argument would be a fatal objection to AU theories. If, however, the AU theorist can articulate a theory of intrinsic value that both is reasonably plausible in and of itself and that can serve as the foundation for AU theories, then the AU theorist will have at least weakened the top down argument, and, if the AU theorist can articulate a theory of intrinsic value that is at least equally as plausible as TU’s account, and that can serve as the foundation for AU theories, then he will have successfully (fully) warded off the mighty top down objection. As I have said, the AU theorist will have warded off the top down objection to AU if he is 350 able to provide a viable alternative account of what has intrinsic value. It’s important to note that this task does not require the AU theorist to show that the TU theorist’s account of intrinsic value is implausible—or even that the TU theorist’s account of intrinsic value is less plausible than the AU theorist’s account of intrinsic value. The AU theorist will have mounted a successful reply to the top down argument even if he shows just that there is a viable AU account of intrinsic value that exists side by side with a similarly plausible TU account of intrinsic value. Furthermore, the AU theorist will arguably have provided a partially successful reply to the top down argument even if the alternative account he offers, while reasonably plausible, seems to be somewhat less plausible than the TU theorist’s account of intrinsic value. There is no question that, at the very least, such a state of affairs would result in a weakening of the top down objection—with the extent of its weakening being a function of the relative plausibility of the AU theorist’s account of intrinsic value. Furthermore, in light of the fact that the top down argument is viewed as a powerful objection, and in light of the fact that it has not been successfully weakened by others, even a mere weakening of the objection would constitute substantial progress. The foregoing points addressed what would be required to bring about a full warding off of the top down objection, and they also addressed the fact that there are upsides for the AU theorist even if the AU theorist falls somewhat short of fully warding off the top down argument. These two states of affairs are not the only two possible states of affairs that might result at the end of my discussion, however. Of course, there is the possibility that my arguments will fail to even weaken the top down objection, but that is not the possible state of affairs to which I am referring. It could be that my discussion not only will show that the AU theorist fully wards off the top down objection by offering an account of intrinsic value that is equally as plausible as the TU theorist’s account of intrinsic value, but it could also be that the discussion shows that while the AU theorist’s account of intrinsic value is plausible, the TU theorist’s account of intrinsic value— 351 contrary to popular belief—itself is not viable or plausible. If this eventuality obtains, then the attempt to ward off the top down objection against AU will result not only in a successful warding off of the objection, but it will result in a positive top down objection against TU. The strength of a top down argument against TU, could, like the top down argument against AU, admit of different strengths. On one end, it could show that TU’s account of intrinsic value is utterly implausible, and thus constitute a fatal objection. On the other hand, it could merely show that TU’s account of intrinsic value is slightly less plausible than AU’s account, and if so, it would not be fatal, but it would still be a positive argument to be reckoned with. Throughout the chapter, the goal will be merely to offer as strong of a reply as possible to the top down objection against AU (be it a weakening or a full warding off), and if it turns out that a positive top down argument against TU exists, this will be an added bonus—with the extent of the bonus being proportional to the strength of the potential top down argument against TU. * * * * I think that there are a few different accounts of intrinsic value that an AU theorist can offer, and this section will raise a few such accounts, and it will consider which if any of them are viable accounts for the AU theorist. Sections 3.4.A, 3.4.B, and 3.4.C will each offer a different potential AU account of intrinsic value, but, despite their seeming promise, I will ultimately argue that these accounts should be rejected. Finally, armed with the lessons learned from these failed attempts (which indeed will be fruitful to consider because of the insights that these failed attempts will provide), I will offer a new account, in Section 3.5, that I believe is successful in warding off the top down argument against AU. 352 3.4.A. Attempt number one: Only entire states of affairs—and not partial states of affairs or events—are bearers of intrinsic value 3.4.A.1. The “entire state of affairs as the only bearer of intrinsic value account” One possible move that an AU theorist could make would be to hold that the only thing that can have intrinsic value is an entire state of affairs, and that, strictly speaking, things such as lives or conscious moments do not have intrinsic value. If this were the case, this would seemingly provide the AU theorist with a plausible response to the top down objection. After all, the top down argument against AU essentially argues as follows: 1. According to TU, an instant of happiness is of intrinsic value, and thus the intrinsic value of a state of affairs is greater, all else equal, the more such moments of happiness it includes, because more instances of intrinsic value means that there is more intrinsic value than if there were fewer such instances. 2. Since maximizing average happiness does not necessarily maximize the number of instances of happiness, average value cannot be the correct theory unless something other than instances of happiness is of intrinsic value. But it seems that nothing other than instances of happiness are of intrinsic value. By claiming that the only bearer of intrinsic value is an entire state of affairs, however, the AU theorist seemingly provides a plausible response to the top down argument. If entire states of affairs, and not persons or conscious moments, are what have intrinsic value, then the top down argument cannot even get off the ground, because it is premised on the claim that conscious moments or persons (or something that is a mere proper part of an entire state of affairs) are bearers of intrinsic value, and this is what the AU theorist’s response denies. So far so good. 246 246 It’s worth noting that there are two versions of this response that the AU theorist might offer. One version of this response would be to concede that that TU as a theory might hold that conscious moments or persons can be bearers of intrinsic value, but still maintain that AU holds that only states of affairs can be bearers of intrinsic value. A second version of this response, however, would be for the AU theorist to maintain that, for both AU and TU, only the entire state of affairs (and not moments or lives) can be bearers of intrinsic value. While neither of these two versions is obviously wrong, it does seem to me that the more plausible of the two versions is the second version I mention—namely, the version according to which the proper understanding of both AU and TU is to have the entire state of affairs be what is the bearer of intrinsic value. My reasons for this can be traced to more lengthy discussions in both chapter 3, chapter 5, and chapter 6, but, in short, it seems that the debate between AU and TU is a debate merely between two different aggregating mechanisms. Both are impersonal moralities and both share various features, and it seems likely that everything other than the choice of aggregation mechanism is and should be a question that is orthogonal and independent of the question of AU versus TU. Thus, 353 3.4.A.2. The problem with the “entire state of affairs as the only bearer of intrinsic value account” 3.4.A.2.i. Even if there is something to be said for the AU theorist’s potential “entire state of affairs as the bearer of intrinsic value account” response to the top down objection against AU, I do not think that it is a particularly plausible or satisfying response, and I do not think that it satisfactorily disarms the top down objection. The reason for this is the following: While it seems as though it is plausible that an entire state of affairs can be the bearer of intrinsic value, it seems as though an entire state of affairs is composed of its component parts—be these parts persons or conscious moments—and thus it seems that the intrinsic value of an entire state of affairs should be a function of the values of its component parts. After all, what else could the intrinsic value of the entire state of affairs be a function of? To make things more tangible, consider the following (run-of-the-mill and arbitrary) example: Consider one possible world, w1, in which there are three individuals, each with a happiness level of five. Now consider another possible world, w2, in which these three individuals exist just as they did in w1, but in which there also exists a fourth individual who has a happiness level of ten. According to both AU and TU, w2 would be better than w1. It seems, however that there is still a question that can be asked about why w2 is better. What is it that explains the fact that w2 is better than w1? What is it about the fourth individual that makes the world it inhabits better than the one in which he does not exist? It seems as though the value of the entire state of affairs must, at least in some sense, be reducible to the values of—or at least explicable in terms of although this surely could be disputed, it seems to me that that the AU theorist who offers this “state of affairs as the bearer of intrinsic value account” should argue that even the TU theorist is mistaken about what his theory should say about what is the bearer of intrinsic value. Going forward, however, we need not concern ourselves with these details, however, because I will argue that the AU theorist’s general claim that the state of affairs is the bearer of intrinsic value is not plausible and that it does not provide a satisfactory response to the top down objection. 354 some features of—its building blocks. I just said “at least in some sense.” What do I mean by that? What are the different senses in which the intrinsic value of an entire state of affairs could be reducible to the values of its building blocks? The most straightforward possibility is that the building blocks of the entire state of affairs—be they persons’ lives or conscious moments—are themselves bearers of intrinsic value. This is the most straightforward account, because it certainly seems as though persons’ lives can be intrinsically valuable or disvaluable, and it also seems as though certain moments or events in a person’s life can be intrinsically valuable or disvaluable. If so, then the AU theorist’s claim that only entire states of affairs are bearers of intrinsic value would be false. There is another account that the AU theorist could defend, however. Perhaps, he might say, although these building blocks—be they persons’ lives or conscious moments—are not bearers of intrinsic value, this does not mean that the intrinsic value of the state of affairs is not reducible to and a function of these components. These building blocks, the AU theorist might say, despite not having intrinsic value, could still be described as having a different kind of value that we might call “goodness.” Let’s suppose the term “goodness” can be used loosely to refer to a property of all sorts of objects—be they entire states of affairs, persons’ lives, or conscious moments. These objects with the property of “goodness” need not be capable of being bearers of intrinsic value, but, at the very least, the fact that they can have “goodness” means that they can bear some relation to intrinsic value. Thus, we can refer loosely to this property of theirs as goodness. It seems as though the AU theorist would have to concede that the building blocks of states of affairs are capable of being bearers of goodness. If so, though, there are two problems that the AU theorist is confronted with. First, it seems that calling the property of the building blocks “goodness” rather than “intrinsic value” seems to be a merely terminological distinction, and the 355 debate about which of the two properties the building blocks are capable of having would be merely a definitional issue. It’s not clear what would hang on this. Secondly, and relatedly, even if we did concede that these building blocks are capable of having goodness but not intrinsic value, it seems that top down objection against AU could then keep its teeth if it were tweaked just slightly: Instead of stating the top down objection in terms of intrinsic value it could instead be stated in terms of “goodness.” The revised top down objection against AU would then essentially say the following: 1. TU has an account of what has the property of goodness and how more goodness in a state of affairs results in the state of affairs having more intrinsic value. 2. AU, however, does not have such an account. What, according to AU, has the property of goodness, and thus what thing is it that if there is more of it, the resulting state of affairs will have more intrinsic value? It seems to me that this new version of the top down argument is no less powerful than the original version that is couched in terms of intrinsic value. 3.4.A.2.ii. While the foregoing discussion in this section should be sufficient to show why the “entire state of affairs as the only bearer of intrinsic value account” is not a satisfactory reply to the top down argument, there is an additional way of articulating and thinking about the foregoing points: by pointing to some seemingly bizarre consequences of such an account: Consider, for example, a state of affairs “s” where the net total happiness (happiness minus unhappiness) in a population is zero. According to both TU and AU, such a state of affairs would thus have a value of zero. Thus, according to the “state of affairs as the only bearer of intrinsic value account,” there is nothing that has either positive or negative value. Suppose, however, that there are two versions of s: s1 and s2. In s1, the entire population lives lives of value zero. In s2, on the other hand, the total (and average) happiness is zero because half of the population lives 356 fantastic lives, whereas the other half live terrible lives. According to the “state of affairs as the only bearer of intrinsic value account,” neither s1 nor s2 includes any intrinsic value or intrinsic disvalue, and they are indistinguishable in this regard. This, however, strikes me (and likely will strike the reader) as implausible. S2 does seem to include a lot that is good and include a lot that is bad. 247248 3.4.A.2.iii. The foregoing considerations should suffice to show why the “entire state of affairs as the only bearer of intrinsic value account” is not a satisfactory reply to the top down argument, but it’s worth noting that other authors (including, among others, John Broome and Larry Temkin) share my view that the goodness of an entire state of affairs is a function of the goodness of its components. As Larry Temkin explains at the beginning of his endeavor to offer a plausible account of aggregation in Rethinking the Good, Aggregation is the process or method of combining different parts into a whole. In ethics, there are principles of aggregation for assessing individual lives. Basically, these tell us how the different elements or components of individual lives combine to affect the goodness, comparative quality, or desirability of those lives. Likewise, there are principles of aggregation for assessing outcomes. Basically, these tell us how the different elements or components of outcomes combine to affect the goodness, comparative quality, or desirability of those outcomes. 249 Thus, it should be clear that there are two key processes involved in articulating a plausible account of aggregation. First, we must be able to identify the elements or components that constitute the relevant parts of the whole in question. Second, we must be able to articulate a 247 We will see that the type of “bizarre consequence” articulated throughout this sub-section will also afflict the “happiness-per-life ratio” account, which I consider, below, in section 3.C. 248 Furthermore, a similar counterintuitive consequence follows regarding intrinsic value for persons: The analogue of s1 would be a life where every moment is of value zero. The analogue of s2 would be a life where half of the moments are fantastic and half of the moments are terrible. The “entire state of affairs as the only bearer of intrinsic value account” implies that there is nothing that has either negative or positive intrinsic value for either of these two people’s lives, and that the two lives are indistinguishable in this regard. This, however, seems implausible. 249 Temkin, p24. 357 function that explains how to plausibly get from the goodness of the parts to the goodness of the whole. The fact that aggregation is composed of these two different processes should not come as a surprise, and all theorists who offer theories of aggregation at least implicitly recognize this. Further, with respect to the first of the two key processes just mentioned, as Temkin says (and Broome also echoes these sentiments elsewhere), it is almost always claimed that the components relevant to the goodness of a life are the “different moments of a life.” 250 Thus, for a utilitarian, the elements relevant for the goodness of a life would be the “‘local’ utilities possessed by that individual over the course of her life, day by day, or perhaps moment by moment.” 251 Similarly, it is almost always claimed that the components relevant to the goodness of an outcome are the individual lives of sentient beings and their respective goodness or desirability. (As I argue in Part II Chapter 1 of this dissertation, however, I myself believe that, while the goodness of a life is a function of the moments of that life, the goodness of an outcome is a function not of the goodness of the lives in the outcome, but rather, of the goodness of the moments in that outcome. In other words, I argue that the unit relevant for ethical inquiry is the conscious moment.) Regardless of what one thinks the unit relevant for ethical inquiry is, however, the view that I and other authors share is that there is a unit that is relevant for ethical inquiry—or, said differently (and, perhaps, more precisely), there is a unit relevant for ethical inquiry and the unit relevant for ethical inquiry is not an entire state of affairs, but, rather, the unit relevant for ethical inquiry is a proper part of an entire state of affairs. Thus, the agreement is that the goodness of a state of affairs is a function of the goodness of its components—whatever these components might be—, and the goodness of a state of affairs is thus determined by using the goodness values for the components and whatever aggregation mechanism one believes is appropriate to transform these values of the components into the value of the entire state of affairs. 250 Ibid. 251 Ibid. 358 In light of the foregoing references, it seems fair to say that I am not alone in thinking that the goodness of an entire state of affairs is a function of the goodness of its components, and, thus, that the “entire state of affairs as the only bearer of intrinsic value account” is not a viable account. Thus, this is further reason to think that the “entire state of affairs as the only bearer of intrinsic value account” is not a satisfactory reply to the top down argument. 3.4.A.2.iv. For all of the foregoing reasons (the main reasons that I offer, the “bizarre consequence” reason that I offer, and, lastly, the appeal to authors holding similar views), it seems that, despite the fact that the “entire state of affairs as the only bearer of intrinsic value account” seemed to have some prima facie plausibility as a response to the top down objection against AU, I think that, on further reflection, this account does not provide a satisfactory reply to the top do