Morality with Humean foundations: a proposal for a solution to the accommodation problem

PDF

Download

Open document

Transcript (if available)

Content MORALITY WITH HUMEAN FOUNDATIONS:
A PROPOSAL FOR A SOLUTION TO THE ACCOMMODATION PROBLEM

by

Keith E. Erwin

________________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHILOSOPHY)

December 2008

Copyright 2008 Keith E. Erwin

ii

Table of Contents

List of Figures v
Abstract vi
Chapter I: The Problem and the Project 1
§I.1 Introduction 1
§I.2 The problem 5
§I.3 Project summary 10
§I.3.1 Ontological accommodation 11
§I.3.2 Semantic accommodation 17
§I.3.3 Normative accommodation 24
§I.3.4 A proposal for a solution 26

PART ONE: Interpreting Hume’s Moral Theory 28
Chapter II: Hume’s Moral Ontology 29
§II.1 An overview 29
§II.2 The distinction between moral sentiment & moral judgment 31
§II.3 Debunking non-descriptivist interpretations of Hume 37
§II.4 Two kinds of moral ideas 39
§II.5 Complex ideas of impression-dependent moral properties 43
§II.6 Bolstering the impression-dependency interpretation 57
§II.7 An objection to the proffered interpretation & a reply 69
§II.8 A second objection to the proffered interpretation & a reply 72

Chapter III: Impressions, Ideas, and Their Relations 85
§III.1 Problems with Hume’s account 85
§III.2 The distinction between original & secondary impressions 91
§III.3 Moral impressions & the passions 96
§III.4 Some concluding remarks 110

Chapter IV: The Foundations of Hume’s Moral Ontology 114
§IV.1 The general principle of morals 114
§IV.2 Sympathy: the non-moral basis of morality 119
§IV.3 An objection to Hume’s account of sympathy 125
§IV.4 Hume’s reply & the transition from non-moral to moral motives 130
§IV.5 Shared perspectives, objectivity, & the proffered account 139
§IV.6 Sympathy, the principles of the moral view, & disinterestedness 141
§IV.7 A second objection to Hume’s account & Hume’s reply 143
§IV.8 Two final points about the general principles of the moral view 147

iii
Chapter V: Sympathy, Moral Motivation, and Causal Reasoning 150
§V.1 A problem for moral motivation: an objection to the account 150
§V.2 The solution to the objection 161
§V.3 A reply to Capaldi 169
§V.4 Mercer’s objection to Hume’s account of sympathy & a reply 176
§V.5 Another objection to Hume’s account of sympathy & a reply 183
§V.6 The relationship of sympathy, reason, utility, & moral judgment 185
§V.7 Two final remarks 195
§V.8 Coda 196

Chapter VI: The Disinterested Moral Point of View 199
§VI.1 General rules & Hume’s conception of the moral view 199
§VI.2 The influences of general rules 204
§VI.3 General rules & the regulation of causal reasoning 207
§VI.3.1 Eight general rules for regulating causal reasoning 212
§VI.4 General rules, sympathy’s variability, & comparison 216
§VI.4.1 Corrective general rules: a parallel approach 222

Chapter VII: The disinterested moral point of view (continued) 227
§VII.1 General rules & the regulation of sympathy and comparison 227
§VII.1.1 A problem with Hume’s restriction to the narrow circle 230
§VII.2 Abramson’s interpretation 235
§VII.3 The proffered interpretation 245
§VII.3.1 Disinterestedly sympathizing from the agent’s position 266
§VII.3.2 Impartiality & the artificial virtue 273
§VII.3.3 Equanimity 276
§VII.4 The MPV: a summary 277
§VII.5 The conclusion of part one 281

PART TWO: Reconstructing Hume’s Moral Theory 286

Chapter VIII: Hume’s Moral Ontology 287
§VIII.1 Introduction 287
§VIII.2 Empirical evidence in support of Hume’s account 292
§VIII.2.1 The origins & workings of empathy 296
§VIII.2.2 Empathy’s variability 302
§VIII.2.3 Regulating empathy 305
§VIII.2.4 The foundations of moral cognition 314
§VIII.2.5 Event-feature-emotion complex framework 321
§VIII.3 A Humean moral ontology: Whither now? 322

Chapter IX: Semantic Accommodation 340
§IX.1 Introduction 340

iv
§IX.2 An attributive analysis of ‘good’ 344
§IX.3 Troubleshooting the proffered attributive analysis of ‘good’ 346
§IX.4 The attributive analysis of ‘good’ & Hume’s moral ontology 354
§IX.5 Differentiating moral properties & values 357
§IX.6 Advantages over the traditional semantic approach 367
§IX.7 Congruity with Hume’s account 370
§IX.8 Summation 374

Chapter X: Normative Accommodation 376
§X.1 Moral normativity 376
§X.2 An analysis of ‘ought’ 378
§X.2.1 Differentiating non-normative & normative ought-statements 383
§X.3 Addressing a discrepancy in the proposed analysis of ‘ought’ 388
§X.4 Instrumental ought-statements & moral ought-statements 390
§X.5 A Humean account of moral normativity 397
§X.6 Congruity with Hume’s account 404

Chapter XI: Ontological Accommodation 406
Accommodating the Fact of Moral Disagreement
§XI.1 Summary of part two 406
§XI.2 A proposal for accommodating moral disagreement 410

Bibliography 417

v

List of Figures
Figure (a) 315
Figure (b) 316

vi

Abstract
The general goal of this dissertation project is to accommodate the peculiar
relationship between moral judgment and human conduct. To this end, this project aims
to sketch out an account of how the basic foundations of Hume’s moral theory have
significant potential to both explain and justify the somewhat peculiar connection moral
judgment appears to have with motivation, action, and normativity. Hume explains the
relationship between moral judgment and human conduct by appealing to human
psychology, in particular, a moral “sense” or sensibility: particular attitudinal states, with
the aid of sympathy and causal reasoning, are the source of moral distinctions and the
motivational efficacy of those distinctions.
This project proposes and defends a particular interpretation of Hume’s moral
theory according to which Hume has a descriptivist account of moral judgments and a
sentiment-dependent/dispositional account of the moral properties ascribed by those
judgments. The project goes on to mount a contemporary defense of this Humean moral
ontology in part by considering recent evidence in the fields of cognitive neuroscience
and social psychology that support Hume’s contention that morality is founded upon our
emotional responses engendered by a psychological process of empathy, similar to what
Hume denominates “sympathy”.
The project also expands its defense by drawing an important distinction between
moral properties, which are feeling-dependent properties of objects, and values per se,
which are not feeling-dependent. Values per se are conceived as relational properties that

vii
hold between objects and ends. Moral qualities are interpersonal, dispositional, emotion-
dependent properties that account for the connection between particular ends, as well as
the objects that fulfill or foil those ends, and particular conative feelings.
This ontological distinction enables this project to offer a sentimentalist,
dispositional account of moral properties that avoids many of the pitfalls typically
associated with such conceptions of morality. In light of the proffered ontology of
morals and values, the project develops a novel semantic analysis of moral and evaluative
terms and expressions, and then sketches out how this analysis can be deployed to resolve
various traditional metaethical problems such as those associated with the issues of
factuality, moral objectivity, and normativity.

1

Chapter I: The Problem and the Project

The hypothesis which we embrace is plain. It maintains that morality is
determined by sentiment. (David Hume, Enquiry Concerning the
Principles of Morals [henceforth EPM], Appendix I)

§I.1 Introduction
David Hume distinguishes between the “speculative and practical”, and claims
that “as morality is always comprehended under the latter division, ‘tis supposed to
influence our passions and actions, and to go beyond the claim and indolent judgments of
the understanding” (A Treatise of Human Nature [henceforth T], p 457). Hume stresses
in this passage and elsewhere that morality is essentially a practical matter rather than a
purely theoretical one. Moral thought is, after all, thought about how to act and what to
pursue.
According to Hume, any noteworthy philosophical analysis of morality must
explain this fundamental and intimate relationship between morality and action.
Consequently, in his writings on morality, Hume is primarily concerned with
understanding this connection. This too is our primary concern. Generally speaking, the
goal of the present project is to explain the relationship between moral thought/judgment
and human conduct.
Significantly, this means providing an explanation that accommodates the fact
that moral thought/judgment can be practical in its issue to just the extent that it is.
1
For

1
Compare this with Michael Smith’s claim that the challenge is “to explain how deliberation on the basis
of our values can be practical in its issue to just the extent that it is” (1992: 330, his emphasis). This project

2
instance, among other things, utterance of a moral thought/judgment typically expresses
the speaker’s intention and motivation to behave in a particular way. It also often
motivates others to act similarly. Perhaps more importantly, sincere moral evaluations
make a claim about what one ought or ought not to do or pursue. Hence they also make a
claim about how one ought or ought not to be motivated to act. In short, the content of
moral thought/judgment is supposed to be action-guiding and hence normative.
The commonplace view is that the normative requirements of morality apply to
persons independently of their particular motivations and intentions. Having motivations,
intentions, or beliefs in opposition to a moral judgment does not release one from, nor
diminish, its normative authority or force. Moral evaluations, as such, function as
practical reasons with a special—some say “queer”—kind of normativity: they make a
claim of obligation upon persons by giving them reasons—viz., normative or ‘ought’
reasons—to conduct themselves in a certain manner irrespective of their own attitudes,
desires, ambitions, intentions, etc. We expect persons to accept and be motivated to act
in accordance with those reasons, no matter what his or her antecedent motivations.
In order to stress this putative attitude-independence, moral evaluations, along
with their peculiar normative status (or reason-giving force), are sometimes described as
categorical. Moral evaluations also consequently have a kind of universality built into
them insofar as they make a claim upon anyone who finds him or herself in the relevant

contends that morality, or moral value, has a more intimate or special connection with practicality than
value per se. (See the following discussion for further explanation of this claim.)

3
circumstances to behave (or to refrain from behaving) in a specific manner, regardless of
that person’s desires, ambitions, intentions, etc.
2

This project aims to demonstrate how a Humean moral theory can accommodate
the extent to which moral evaluation is, by all appearances, practical. In other words, its
goal is to show that the basic foundations of Hume’s moral theory have significant
potential to explain and justify the somewhat peculiar connection moral thought appears
to have with motivation, action, and normativity.
3
However, this project does not give a
full justification. It only offers a sketch. Moreover, as we shall see, the Humean
foundations to which this work appeals rest on substantial empirical claims. While the
scientific evidence in support of those claims continues to be amassed, the evidence is
still incomplete. Nevertheless, we shall examine some of the mounting scientific
evidence because it turns out to be important: not only does it lend credence to Hume’s
account; it also enables this project to improve upon that account.
Hume’s view is that morality must have an inherent connection with sentiment,
i.e., certain motivational feelings.
4
He argues that any plausible analysis of moral
evaluations must ultimately make reference to particular conative attitudinal states,

2
I am not at this point trying to defend or even fully explain moral normativity; rather I am merely trying to
quickly capture what I take to be the commonplace view of morality’s normative character.

3
Discussion of the commonplace view of moral evaluation’s normativity (given above) is not explicitly
discussed in Hume’s work. Nevertheless, I shall demonstrate that a defense of the commonplace view can
be constructed out of the account of moral normativity Hume does offer.

4
Hume arrives at this conclusion in large part because he thinks that reason alone can never motivate us:
Since morals, therefore, have an influence on the actions and affections, it follows, that
they cannot be deriv’d [solely] from reason… Morals excite passions, and produce or
prevent actions. Reason of itself is utterly impotent in this particular. The rules of
morality, therefore, are not conclusions of our reason [alone]. (T 457)

4
namely, attitudes of “approbation” or approval and “disapprobation” or disapproval.
According to Hume, our moral feelings are a product of our capacity for sympathy. In
fact, sympathy is not only a necessary precondition for our moral sentiments, for Hume,
it is also the non-moral, empirical basis of morality. By holding that these inherently
motivating feelings—viz., our sympathetic and moral feelings—are the foundation of
morality, Hume thereby explains why morality is essentially a practical matter.
Extinguish all the warm feelings and prepossessions in favor of virtue, and
all disgust and aversion to vice; render men totally indifferent towards
these distinctions; and morality is no longer a practical study, nor has any
tendency to regulate our lives and actions. (EPM, section I)

Nevertheless, Hume immediately goes on to state:
I am apt to suspect…that reason and sentiment concur in almost all moral
determinations and conclusions. The final sentence, it is probable, which
pronounces characters amiable or odious, praise-worthy or blameable; that
which stamps on them the mark of honour or infamy, approbation or
censure; that which renders morality an active principle and constitutes
virtue and happiness, and vice our misery: it is probable, I say, that this
final sentence depends on some internal sense of feeling, which nature has
made universal in the whole species. For what else can have an influence
of this nature? But in order to pave the way for such a sentiment, and give
proper discernment of its object, it is often necessary, we find, that much
reasoning should precede, that nice distinctions be made, just conclusions
drawn, distant comparisons formed, complicated relations examined, and
general facts fixed and ascertained. (Ibid, Hume’s emphasis)

So Hume is sensitive to the fact that reason too plays a vital role in the discovery
of moral distinctions by clearing the way for sentiment. In his view, moral sentiment is
responsive to a particular kind of reasoned evaluation, namely, our assessments of
objects’ tendency to bring about or balk various ends.
5
From this assessment, which is a

5
§II.6 and §V.6 discuss this account in more detail.

5
product of causal reasoning, we can determine whether an object is “useful and
beneficial”. Hence causal reasoning is often a necessary precondition for the production
of our moral feelings.
In sum, Hume explains the relationship between moral thought and human
conduct by appealing to human psychology, in particular, a moral “sense” or sensibility:
particular attitudinal states, with the aid of sympathy and causal reasoning, are the source
of moral distinctions and the motivational efficacy of those distinctions.
§I.2 The problem
While I shall argue that Hume’s contention—viz., that moral thought/judgment
has an essential connection with attitudes of approval and disapproval—is correct and of
great import, explaining sufficiently the connection between moral thought/judgment and
our conduct remains a formidable task nonetheless; for the problem of accommodating
the practicality of moral thought/judgment—i.e., the extent to which it is practical—
extends in important respects beyond Hume’s contention. There is, for instance, an
obvious prima facie tension between Hume’s moral theory, which holds that morality is
in some sense dependent upon certain attitudes, and the commonplace assumption that
morality’s normativity is independent of individuals’ attitudes. This tension must be
resolved.
There are also other commonsense assumptions about the linguistic and
nonlinguistic phenomena associated with moral thought/judgment that are interrelated
with its practicality in philosophically significant ways. They too are therefore in need of
accommodation. Among the commonsense assumptions to be accommodated are the

6
following: (1) moral evaluation/thought is descriptive—i.e., moral evaluations/thoughts
are straightforward beliefs; hence (2) our moral judgments/evaluations ascribe moral
properties to objects; (3) at least some moral judgments/beliefs are true; therefore, (4)
there are moral properties and moral facts; (5) both error and moral disagreement are not
only real possibilities but plausibly often do occur—i.e., we can be mistaken and
genuinely disagree about the moral facts, as well as the truth-value of a given moral
judgment/belief. To this list we can of course also add: the practicality of moral
thought/judgment, which not only includes (6) a direct link with motivation and action
but also (7) a special kind of normative authority (or reason-giving force), namely, one
that applies to individuals independently of their particular attitudes.
Taken together, these assumptions entail that the moral domain enjoys a certain
kind of objective status. First, the putative fact that we can be mistaken and disagree
about the moral facts and the truth-value of individual moral judgments/beliefs accounts
for one important aspect of the claim to moral objectivity. The second key aspect of
morality’s objectivity is captured by the categorical nature of moral normativity; i.e., the
commonplace view that a moral judgment’s normative authority or reason-giving force
applies to individuals independently of their particular attitudes.
Some, such as Alan Gibbard, disparagingly refer to the above assumptions as the
“objective pretensions” of morality, but they are more than mere pretensions, or
aspirations, to objectivity. Those commonsense assumptions about moral thought and
judgment serve as our starting-points. In other words, moral thought and practice appear

7
to presuppose them. Consequently, I suggest that these assumptions are more aptly
viewed as morality’s objective presuppositions or presumptions.
Each of these commonsense presuppositions (1-7) involves positive ontological
commitment. For instance, presuppositions 1-3 entail positive commitments concerning
the existence and nature of moral expressions’ putative truth-value; presuppositions 4 and
5 involve positive commitments with respect to the existence and nature of the moral
objects—e.g., properties, facts, and/or reasons—putatively denoted by moral expressions;
6 involves positive commitment regarding the existence and nature of the connection
between those putative moral objects and motivation; and the last presupposition (7)
entails positive commitments concerning the existence and nature of moral expressions’
putative normative character (or reason-giving force).
6

However, these positive commitments raise certain burdens of explanation. The
problem is that meeting these explanatory burdens reputedly subverts at least some of
those positive commitments and thereby threatens the commonplace view of morality.
For instance, one significant burden is to explain how an epistemologically accessible
moral property can possess the sort of practicality—i.e., the rather peculiar connection
with normativity and motivation—that the commonplace view of morality suggests.
Many are convinced that an adequate explanation cannot be given.
In short, we face a kind of expository struggle. On the one hand, accommodation
of the commonplace view of morality necessitates that we take on certain positive
ontological commitments. On the other hand, accommodation of the explanatory burdens

6
Whether or not such positive ontological commitment amounts to a commitment to, and presumption of,
moral realism remains to be seen.

8
raised by these commitments is thought to entail that we must give up at least some of
those very commitments. Following Mark Timmons (1999), we shall henceforth refer to
this expository struggle—i.e., the tug upon our commitments in opposing directions—as
the accommodation problem.
7

There are three broad strategies for responding to the accommodation problem.
One strategy is to challenge the explanatory burdens themselves by arguing that they are,
in effect, misplaced or misguided. According to this strategy, the fundamental nature of
morality is ultimately rather inexplicable. It is inexplicable because, for instance, moral
objects are irreducibly moral or normative and hence sui generis such that we can have

7
Timmons describes the problem as one of internal and external accommodation. (See also Stephen
Finlay (2007) who follows Timmons in noting the tension between these two forms of accommodation.)
Internal accommodation is a matter of capturing the various features suggested by the commonplace view
of morality—e.g., the assumptions (1-7) above—whereas external accommodation is a matter of fitting
these features within our current understanding of the world “external” to morality. However, the
distinction between internal and external accommodation may be somewhat misleading. Stating the
accommodation problem this way appears to presuppose that there is a kind of continuity between the
moral domain and the non-moral domain; that is, the task, as Timmons and Finlay describe it, is to
somehow fit the moral domain within the non-moral. But it seems at least possible, even if unlikely, that
the moral domain is wholly sui generis. It might therefore lack any continuity with our independent
understanding of the world “external” to morality; it simply might not fit. No doubt such a view will raise
important questions such as how we could have epistemological access to such a sui generis domain, but
moral epistemology might itself be entirely sui generis. Hence it seems that the explanatory burdens raised
by such a view might occur from within the moral domain itself, without any need to appeal to a world
“external” to it. If so, then the distinction between that which is internal and external to morality would not
be very helpful.
Having said that, I do think that a theory that fits the putative features of morality within a
systematic view of our independent understanding of the world, mind, etc. is preferable to one that does
not—just as a theory that explains is preferable to one that does not. Moreover, the moral domain certainly
seems prima facie to be at least somewhat continuous with the non-moral. For instance, moral language
appears to be descriptive precisely because it’s hard to see a semantic difference between our moral
judgments and our non-moral judgments. In fact, even contemporary non-descriptivists assume that there
is no fundamental semantic difference between moral and non-moral judgment: on their view, all
judgments express thought. But the point I’ve been stressing in this footnote is simply that we should not
set up the problem in a way that precludes the possibility that the moral domain is discontinuous with our
independent understanding of the world, mind, etc, no matter how improbable or problematic we may take
that possibility to be. For this reason, I have chosen to state the accommodation problem in terms of the
accommodation of morality’s objective presuppositions along with the sort of explanatory burdens this
places on those who take up the challenge of accommodation. As we shall see, the answer I provide is
perfectly consistent with a naturalistic conception of the world (both “internal” and “external” to morality).

9
knowledge of them but we cannot explain their rather peculiar nature. The second
strategy accepts the explanatory burdens but sees them as too great to discharge.
Consequently, this strategy accommodates these burdens by explaining away some or all
of morality’s objective presumptions as mere “pretensions” or appearances.
8
In other
words, it fails to accommodate all of the commonplace moral presuppositions. The final
strategy is to discharge the explanatory burdens and defend the commitments entailed by
morality’s objective presuppositions. This strategy attempts to demonstrate how the
subject matter of morality really is everything it is presumed to be.
This last strategy is the general strategy of this project. This project proposes a
particular Humean account of the content of moral thought/belief that is compatible with
the commonsense objective presuppositions of morality, or so I shall argue. Equally
important, the account offers a general explanation of the semantic, ontological, and
normative commitments these presuppositions entail, and thereby an explanation of the
intimate and somewhat peculiar connection between our moral beliefs/judgments, moral
motivation, and action. However, this project does not claim to offer a complete
justification of those commitments. As it stands, it provides nothing more than a
proposal for a solution to the accommodation problem, rather than a definitive solution.
Furthermore, as mentioned, the account rests on particular empirical claims. While
current scientific research does lend some support to the proposed Humean view and its
key contention that morality is founded upon particular psychological processes, the

8
This approach is therefore typically associated with moral antirealism (or irrealism).

10
evidence admittedly does not conclusively demonstrate that the proposed view is correct.
So, there is much more philosophical and scientific work to be done in the future.
§I.3 Project summary
Given that this project aims to defend all of the commonplace objective
presuppositions (1-7) by appealing to the foundations Hume lays out in his moral theory,
its proposal for a solution to the accommodation problem will involve both interpretation
and reconstruction. As we shall see, the account is a reconstruction of Hume’s moral
theory since it differs from his theory in some important respects, and it tackles issues
that Hume simply does not address. The account also provides an interpretation of
Hume’s moral theory since the proposed reconstruction is a reconstruction of Hume.
Indeed, it claims to share the very same foundations as Hume’s moral theory.
So before we consider the proffered reconstruction of Hume’s moral theory, we
will need to explain and defend the particular interpretation of his moral theory and its
foundations, upon which the reconstruction builds. Consequently, the project is divided
into two main parts. Part one (chapters II-VII) discusses the proposed interpretation of
Hume. Part two (chapters VIII-XI) gives a reconstruction based on that interpretation of
Hume and simultaneously offers a proposal for a solution to the accommodation problem.
For purposes of clarification and exposition, this project in effect divides the
accommodation problem—i.e., the task of accommodating all of morality’s objective
presuppositions—into three distinct but interrelated steps: ontological accommodation,
semantic accommodation, and normative accommodation. The first step in resolving the
accommodation problem is to provide ontological accommodation, namely: to offer an

11
account of the nature of moral properties and facts that can accommodate the various
features of our moral practice and discourse. The two remaining steps are to demonstrate
how this moral ontology can indeed accommodate the semantic and normative features of
moral discourse that our moral practice appears to presuppose. Semantic
accommodation, as such, is basically a matter of explaining how moral thought and
language express moral properties and facts. Normative accommodation primarily
involves explaining how these moral properties and facts can have a normative authority
or reason-giving force that is independent of individuals’ actual motivations.
With this in mind, what follows is a brief exposition of the project in light of the
challenge of providing ontological, semantic, and normative accommodation.
§I.3.1 Ontological accommodation
Hume’s moral ontology has significant potential to accommodate the various
features of our moral practice and discourse. However, Hume’s moral ontology is the
subject of significant interpretive dispute. For instance, many have suggested that
Hume’s moral ontology consists of nothing more than our feelings of approval and
disapproval. I disagree. Part one, the interpretive part of the project, argues that Hume
offers a more robust moral ontology than this. Hume claims that moral qualities are “a
relation, which nature has placed between the form [of an object] and the sentiment.”
9

Moral qualities, as such, are by no means identical to our sentiments, though they are

9
Of the Standard of Taste [henceforth SOT], paragraph 10.

12
sentiment-dependent properties.
10
As I read Hume, he holds the following general
characterization of the moral domain: an object X belongs to the moral domain if and
only if observation or contemplation of the typical effects of X elicits particular
feelings—viz., feelings of approbation or (as the case may be) disapprobation toward
X—within those who satisfy certain conditions. In short, the domain of moral properties
is determined by reference to particular affective attitudinal states, namely, moral
sentiments. Thus, Hume draws an important distinction between moral sentiments on the
one hand and moral properties on the other.
As we shall discuss in chapters II and III, our moral judgments ascribe these
sentiment-dependent moral properties to objects. Consequently, in Hume’s view, moral
judgments are descriptive, property ascribing, truth-evaluable judgments. Moreover,
Hume holds that many of our moral judgments are true. There are moral properties and
facts, and both moral error and genuine disagreement, including disagreement about the
truth-value of a given moral judgment, are not only possible but do actually occur.
11

Chapters IV and V explain how moral sentiments depend upon our capacity for
sympathy. Significantly, our sympathetic capacity is not only a necessary precondition

10
As far as I am aware, Mackie (1980) was the first to defend this sort of interpretation of Hume’s moral
ontology, although it is noteworthy that, on Mackie’s reading, Hume holds that there are no such
properties. D. F. Norton (1982) appears to be the first to defend the sort of interpretation I offer, according
to which (many) objects do possess response-dependent moral properties.

11
Chapter II defends the claim that Hume offers a straightforward descriptivist account of moral judgment
according to which there are (non-minimal) moral truths and facts. This contention is disputed however.
For instance, as we shall discuss in chapter II, several Humean scholars have offered an emotive or
expressivist interpretation of Hume’s theory of moral judgment. John Mackie (1980) initially suggests an
emotive or expressivist interpretation, but ultimately defends as the most plausible interpretation what he
calls the “objectification theory”, which is a kind of error theory. In his view, Hume holds that moral
judgments ascribe properties but there are no such properties. Hence all moral judgments are false.

13
for our moral feelings, for Hume, it is also the non-moral empirical basis of morality. It
is, of course, not at all difficult to see how such inherently motivating psychological
states—viz., our sympathetic and moral feelings—would indeed be intimately connected
with human conduct.
However, as we shall also discuss, our feelings of moral approval and disapproval
do not always accurately track moral distinctions or our judgments of them. This
certainly calls for explanation. To his credit, Hume was well aware of the problem. He
is, consequently, careful to appeal not merely to our feelings of approval and disapproval
but equally to the conditions under which these feelings occur—hence the real possibility
of error (e.g., mistaken sentiment leading to a false belief/judgment) and thus potential
for genuine moral disagreement, including disputes about the truth of a particular
judgment. These “standard conditions” act as a corrective measure to ensure uniformity
of belief/judgment. In other words, they serve as standardizing conditions. Chapters VI
and VII discuss the role of these standardizing conditions in Hume’s moral theory.
To help explain his account, Hume more than once explicitly draws comparison
between moral qualities and other secondary-qualities such as color, as well as between
moral judgments and our color judgments. For instance, the domain of color qualities is
arguably determined by reference to particular visual states we experience when we
satisfy certain conditions. Consequently, when judging the color of an object, people can
(and sometimes do) make mistakes because they can fail to meet the requisite conditions.
In other words, our color perceptions can diverge from actual color distinctions. When
this divergence occurs, it may lead to mistaken (i.e., false) color judgments/beliefs.

14
If we want to know the true color of an object and avoid such mistakes, we would
do best to appeal to those who are qualified, namely, those operating under the optimal or
at least “standard” conditions for making such a judgment—e.g., where the object is in
plain sight, subject to normal lighting conditions, and those observing are normally
sighted. According to Hume, in the case of all of our perceptions, we operate with such
standardizing conditions in mind in order to ensure uniformity of judgment.
12
So, for
instance, although the color an object appears to me seems to change as a result of
variations in my particular perspective and/or lighting conditions, I do not say that the
actual color of the object also thereby changes. For example, I do not say that an apple
ceases to be red when the lights are turned off; nor do I say that the apple’s redness
diminishes as the lights are dimmed. I recognize that, though the object’s color may
appear to vary, its color nonetheless remains the same.
Somewhat similarly, in the moral case, Hume observes that as a matter of
empirical fact I will presumably not “feel the same lively pleasure from the virtues of a
person, who liv’d in Greece two thousand years ago, that I feel from the virtues of a
familiar friend and acquaintance. Yet I do not say, that I esteem one more than the
other…” (T 581). In this case, I recognize that, although the moral value (or virtue) of
these character traits appears or feels different to me, their moral value remains the same
nonetheless.
Thus, much like the case of color judgments, if we want to know the moral value
of things, we must always consider the conditions and perspective under which our moral

12
Hume (T 582) explicitly claims that all forms of human perception are subject to correction to
secure uniformity of judgment, not only our perception of secondary-qualities.

15
observations take place. In other words, our feelings of approbation and disapprobation
are indicative of genuine moral distinctions only if we are operating/judging under the
right sort of conditions, that is, only if we, as moral observers, are deemed qualified.
Moral observers or judges are qualified, according to Hume, in virtue of their “strong
sense, united to delicate sentiment, improved by practice, perfected by comparison,
cleared of all prejudice.” Their “joint verdict is the true standard” because they judge
“most coolly” from the “steady and general” moral point of view. In short, moral
properties are response-dependent properties, but our attitudinal responses are only
indicative of moral properties when we satisfy certain conditions.
Hence, on Hume’s account, legitimate moral judgments, like legitimate color
judgments, are intersubjective judgments of fact established from a privileged perspective
that imposes a particular set of conditions. In the moral case, these conditions involve
the activity of particular sentiments engaged from within the “general [moral] view.”
Again, in much the way color distinctions arguably ultimately rest upon universal or
nearly universal color experience (under standard conditions), Hume argues that moral
distinctions ultimately rest upon universal or nearly universal moral experience (under
standard conditions). (See again the two passages quoted in §I.1.) The Humean view
proposed in this work defends the plausibility of this contention. This is auspicious since
this contention plays a vital role in accommodating our commonsense presuppositions
about the objective standing of morality.
However, as we shall discuss, Hume’s account of sympathy, as well as his notion
of a universal moral point of view and its standard conditions, are subject to various

16
criticisms. For instance, the extent of moral disagreement prima facie speaks against a
universal moral perspective, and, in any case, Hume’s account of the standard conditions
of the moral view is arguably too abstract to be a reliable corrective measure for moral
judgment and our corresponding feelings. Furthermore, in those cases in which the judge
finds herself in non-standard conditions, she may fail to have sympathetic feelings. If so,
then no moral sentiment will be engendered. The possibility—indeed the likelihood—of
such cases appears to raise a problem for moral motivation and judgment: prima facie, a
moral sentiment cannot be generated to motivate us to act morally in these instances, and
if not, then moral thought and judgment consequently seem to lose their intimate
connection with motivation and action. In short, there appears to be sundry reasons for
reservation about Hume’s moral ontology.
Since the main concern of this work is to demonstrate the prospect for success of
solving the accommodation problem by developing an account of moral
thought/judgment along Humean lines, it must uphold the foundations of Hume’s moral
ontology against various criticisms leveled against them. Part one argues that Hume has
the materials needed to stave off many of the aforementioned attacks, but it ultimately
concludes that Hume’s moral theory is in need of various alterations and additions in
order to adequately defend his moral ontology and its foundations, namely: (i) the notion
of a privileged universal moral perspective that imposes particular conditions that
produce interpersonal agreement, and (ii) the claim that morality rests upon human
psychology, in particular, sympathy and the moral feelings it produces.

17
It is at this point (viz., the end of chapter VII) in which the project transforms
from an interpretive enterprise into a reconstructive one, and thus transitions from part
one to part two of the project. Part two represents this project’s attempt to improve upon
what I will have argued in part one is Hume’s account. If the interpretation of Hume’s
moral theory put forth in part one turns out to be incorrect, then part two will not be quite
as congenial to Hume’s moral theory as I suggest. But this in no way detracts from the
fact that Hume’s work has greatly inspired this project.
Part two argues that Hume’s moral ontology, or at least a moral ontology heavily
inspired by Hume, is not only congruous with the empirical evidence being gathered in
the physical and social sciences; it also has considerable potential to account for the
intimate connection between moral thought and human conduct, and thereby resolve the
accommodation problem. The remainder of the project defends this suggestion by
indicating how.
§I.3.2 Semantic accommodation
Perhaps the strongest objections against Hume’s moral ontology question its
ability to provide both semantic and normative accommodation. Hume scarcely
addresses either semantic or normative issues. Among other things, part two offers a
semantic analysis of moral evaluative terms and expressions consistent with Hume’s
moral ontology. At least a general analysis of the meaning of moral expressions will be
of major importance since some of the objective presumptions in need of accommodation
pertain to moral language itself. As I read Hume, all moral judgments ascribe sentiment-
dependent moral properties. Hence, in Hume’s view (or at least what I take to be Hume’s

18
view), an analysis of moral judgments and the moral properties they ascribe must
ultimately make reference to the moral sentiments. This project agrees.
However, Hume does not offer any explicit linguistic analysis. In particular, he
does not explain how any moral evaluative term or expression—e.g., “x is morally
good”—makes reference to a moral property. I suggest this is due to the fact that Hume,
influenced by many of his predecessors and contemporaries, was concerned with giving
an account of moral psychology and moral experience itself rather than an account of
moral language.
13
So, instead of reading Hume as offering explicit analysis of the literal
meanings of moral evaluative terms and expressions, I suggest he is more plausibly read
as providing a set of general application conditions for moral ascriptions. In other words,
Hume’s theory only provides conditions, in particular psychological conditions, for
identifying (i) when an object (e.g., an act or trait) is moral or immoral, and consequently,
(ii) when it is generally apposite to use certain types of moral evaluative expressions, for
example, ‘x is a virtue, ‘x is morally good’, ‘x is morally right’, or ‘one morally ought to
do x’. His theory does not give the semantic content of these expressions however. For
instance, it neither tells us whether there is any difference in meaning between these
expressions, nor does it fully specify their truth conditions.
Because Hume fails to provide explicit analysis of moral evaluative language, a
significant problem looms large: it is not clear how discrete moral expressions refer to

13
Recalling my opening remarks in §I.1, the reader might wonder how I can hold both that (i) Hume was
primarily concerned with providing an account of the intimate connection between moral judgment and
human conduct, and that (ii) Hume was not concerned with giving an account of moral language but rather
moral psychology. The answer is that Hume conceives of judgment as first and foremost a mental entity
rather than a linguistic one. Hence moral judgment, as Hume conceives it, is wholly an issue of
psychology. (It is noteworthy that I reject this sort of psychologism about judgment. See Chapter III for
further discussion.)

19
moral properties. To illustrate why this is a serious problem consider the following
example. As discussed, moral properties are determined by reference to one of only two
kinds of moral sentiment, namely: approval or disapproval. Hence, according to Hume’s
account and the account proffered by this project, all moral expressions must refer to one
of only two kinds of moral property. Hence the following moral expressions all appear to
somehow refer to the same moral property: ‘moral virtue’, ‘morally good’, ‘morally
right’, and ‘morally ought’. Yet these expressions presumably still wouldn’t be
semantically equivalent since the evaluative terms ‘virtue’, ‘good’, ‘right’, and ‘ought’ do
not appear to be synonymous. This demands explanation, but unfortunately no
explanation is found in Hume’s work. Thus, in the end, Hume leaves us with the vexing
problem of identifying the analysis of moral evaluative terms and expressions his moral
theory implicates, or at least the problem of identifying the analysis that best fits with a
moral theory that embraces Hume’s moral foundations, along with certain semantic facts.
Part two tackles this pressing problem. But more generally its semantic account
of moral and evaluative language is aimed at resolving the accommodation problem. As
we shall see, this undertaking will take us well beyond Hume, for the proffered analysis
of moral and evaluative language amounts to nothing less than going back to the
proverbial drawing board and constructing an account of evaluative language and value
properties that is absent from, though remarkably compatible with, Hume’s moral theory.
The semantic analysis to be proposed supplements Hume’s moral theory by drawing a
crucially important distinction between moral properties and value properties. Moral

20
value is thus taken to be a conjunction of these distinct, but interrelated, properties.
14

According to the proposed account, all moral evaluations refer to both of these
properties.
This project contends that values are relational properties, generally characterized
as a matter of an object—e.g., an act, person, or character trait—fulfilling, furthering,
foiling, or frustrating certain ends or goals.
15
Moral value judgments—e.g., that an object
o is morally good, wrong, etc—can, on this view, be broadly construed as assessments of
whether (or not, in the case of negated judgments) certain acts, traits, persons, etc. fulfill,
further, foil, or frustrate moral ends or goals.
16
Similarly, prudential value judgments are
broadly construed as assessments of whether an object fulfills, furthers, foils, or frustrates
prudential ends or goals.

14
Observe that evaluative language such as ‘good’, ‘bad’, ‘right’, ‘wrong’, ‘ought’, ‘should not’, and even
‘virtue’ and ‘vice’ is legitimately used in non-moral evaluations, arguably without any change in the
meaning of these terms. The suggestion put forth in part two of this project is that this is because these
evaluative terms per se do not ascribe any moral property; rather, they ascribe value properties, which are
distinct from moral properties.

15
This general theory of value, which holds that value is a relational property between objects (to which a
value property is ascribed) and ends or goals, or something similar such as interests, needs, or wants, traces
at least back to Socrates who, in Xenophon’s Memoirs, Book III, chapter 8, states: “Everything else that we
use is considered to be fine and good in accordance with the same standard—namely, the end for which it
is serviceable” (Conversations of Socrates, translated by H. Tredennick and R. Waterfield [New York,
1990]). This general theory of value can also be found in the work of George Santayana, Winds of
Doctrine (London: J. M. Dent & Sons, 1913); D.W. Prall, A Study in the Theory of Value (University of
California Publications in Philosophy, Vol. 3, No. 2, 1921; 215-227); Ralph Barton Perry, General Theory
of Value (New York: Longmass, Green & Com., 1926); Peter Geach, “Good and Evil” (Analysis 17, 1956:
33-42); Paul Ziff, Semantic Analysis (Ithica: Cornell University Press, 1960); John Mackie, Ethics:
Inventing Right and Wrong (Penguin Books: England, 1977: 58-62); and more recently, Stephen Finlay,
“The Conversational Practicality of Value Judgment” (Journal of Ethics 8, 2004: 205-23) and “Value and
Implicature” (Philosophers Imprint, Vol. 5, No. 4, 2005).

16
Here and throughout, I use frustrate and foil in a somewhat technical sense, namely, in contradistinction
to the difference marked out by fulfill and further: one can perform an act that furthers an end or goal
without fulfilling it. Thus, as I use the terms, one can perform an act that frustrates an end or goal without
foiling it.

21
This account may initially seem to conflict with Hume’s theory since, as stated,
the analysis does not appear to make any reference to human sentiment; but it will be
argued that the proposal is remarkably compatible with, and in fact complementary to,
Hume’s moral theory. Indeed, in my view, perhaps the most significant contribution of
this project lies in its explanation of how these two seemingly opposing theories—viz.,
Hume’s (or a Humean) moral theory and the end-relational theory of value—actually fit
quite nicely together to provide a promising solution to the accommodation problem.
According to both Hume’s and the proposed account, an end belongs to the moral
domain in virtue of its ties to human sentiment. More precisely but still generalizing
somewhat, an end is moral or (as the case may be) immoral if and only if observation or
contemplation of the fulfilling or furthering of that end or goal elicits in qualified
observers—that is, observers subject to certain conditions—feelings of approval or (as
the case may be) disapproval. Thus, the domain of moral properties is determined by
reference to particular motivational attitudes or sentiments that occur within persons
subject to certain conditions. In the characteristic way just described, moral properties
and value properties conjoin to produce moral value. Moral thought/judgment is
therefore about/refers to both of these properties.
In other words, this project argues that moral thought is, in part, evaluative
thought. And what makes a thought evaluative is the fact that it is an assessment of
whether an object—e.g., an act—fulfills or foils a given set of ends. This account has
much in common with Hume’s view since it too refers to ends. In Hume’s view, by
fulfilling the relevant ends, objects are useful or agreeable to others and/or to oneself. On

22
the proposed account, what makes moral thought distinctly moral is quite different from,
though intimately related to, what makes it evaluative. What makes a thought moral has
to do with our particular attitudinal responses to the fulfilling or foiling of the ends or
goals at issue. These intersubjective attitudinal responses (with certain conditions in
place) determine the moral domain, including the moral properties of ends and thereby of
value-states, as well as the object(s) that have such value, namely, moral value.
Thus, just as Hume argues, (i) moral distinctions depend upon human sentiment
(subject to standard conditions), and therefore (ii) sentiment plays a crucial role in both
the identification and evaluation of moral qualities. However, by explicitly drawing the
ontological distinction between values and moral properties, this project marks a
significant improvement over Hume. First, it offers a clear and concise account of both
kinds of property, which Hume does not do. Second, this ontology also enables the
project to provide an explicit semantic analysis of moral and evaluative language, which
Hume also does not do. This semantic analysis has significant explanatory power. For
instance, the distinction opens up an unexplored avenue for avoiding weighty objections
to other Humean-type sensibility theories that do not employ the distinction.
With respect to this latter claim, I have in mind the sensibility theories put forth
by David Wiggins (1987) and John McDowell (1987). Our Humean theories agree that
there is an essential tie between our motivational experiences and moral properties. More
precisely, we agree that there is a necessary connection between moral properties and
certain affective sensibility-states (viz., feelings of approval and disapproval). Just as
color terms and the intersubjective properties to which they refer seem to have a kind of

23
“to-be-seenness” conceptually built into them, so too do moral terms and properties
appear to have a kind of “to-be-pursuedness”, or action-guidingness, conceptually built
into them. In this much we agree.
However, neither Wiggins nor McDowell draw the distinction between moral
properties and values. In other words, they offer their sensibility theories as a theory of
value such as, for example, goodness per se.
17
The theory advanced here does not.
Value properties are taken to be entirely distinct from moral properties. Unlike moral
properties, values per se—e.g., the goodness of an act—do not exhibit a necessary a
priori tie to conative states. Values are entirely attitude-independent. As we shall
discuss, this distinction plays a crucial role in defending the sensibility theory proposed
here against various objections to which Wiggins and McDowell’s Humean-type
sensibility theories are arguably subject. For instance, their theories are hard pressed to
account for the lack of synonymy between evaluative terms such as ‘good’, ‘right’, and
‘virtue’ since each of terms are presumably tied to the same sensibility state, namely,
moral approbation.
18

This project also argues that the proposed distinction between moral and value
properties presents an account that is superior to the relational theory of value alone.

17
See also, for instance, G. Sayre-McCord (1988, 1991).

18
As we shall also discuss, the proposed view is able to explain why judgments of value simpliciter—e.g.
that some act X is a good—do appear prima facie to have direct ties with our emotional responses. In other
words, it is able to explain why a sensibility theory of value does appear to be prima facie plausible. This,
in turn, explains why so many philosophers have quite naturally read Hume as (inevitably) offering a
sensibility theory of value and evaluative terms. But I suggest that Hume’s theory is not forced into this
position because the distinction between moral and value properties offers an alternative that is remarkably
compatible with Hume’s account. I shall argue that this alternative is superior to other proffered options
for Humeans who, like Hume, embrace moral descriptivism. For example, it is superior to the option
pursued by Wiggins, McDowell, and Sayre-McCord.

24
Most, if not all, of those who have proffered the relational theory of value have tried to
accommodate the various features of moral discourse and practice by appealing to value
properties alone. This subsequently results in these theories construing moral properties
simply as particular instances of value properties. Unfortunately, moral properties, as
such, lose their inherent tie to motivational states because all values per se exhibit only a
contingent link with motivational states. One important consequence of this is the fact
that these theories cannot justify the objective normative status of morality that the
commonplace view presumes, or so I shall argue in part two.
§I.3.3 Normative accommodation
The proposed distinction between value properties and moral properties plays a
key role in accounting for morality’s normativity, that is, morality’s action-guiding
nature. The standardizing conditions of the moral perspective function as norms
governing evaluative thought on the one hand and our capacity for sympathy on the other.
These two sets of norms are put in place to ensure as far as is possible both that (1) we
accurately assess the value of objects—i.e., that the object in question really does fulfill,
further, foil, or frustrate the relevant ends—and that (2) our capacity for sympathy
functions properly so as to generate the inherently motivational feelings by which we
apprehend the moral properties of objects. In short, together, these norms are both
reason-guiding (regarding our assessment of objects in terms of fulfilling, furthering,
foiling, or frustrating ends) and motivation-guiding (regarding the ends of action), and
hence action-guiding.

25
These norms play an important role in explaining how moral thought/judgment
can be practical in its issue to just the extent that it is. Part two of this project also argues
that the Humean-style intersubjective and inherently motivational nature of moral
properties plays a vital role in accounting for the commonplace presumption that
individuals are subject to morality’s normative authority (i.e., morality’s normative
requirements or reasons) independently of their desires, intentions, and all other
motivational or affective states. As stated, this commonplace presumption is correct. It
is not a contingent matter whether the moral facts themselves, concerning what one
morally ought or ought not to do or pursue in a given situation, provide a person with
normative reasons.

According to the proposed Humean account, even if one is not
motivated by, or concerned with, moral considerations, one is nonetheless committed to
the belief that one morally ought to be so motivated and concerned. In other words,
moral considerations have a kind of normative force or authority to which individual
agents—in particular, their beliefs and commitments—are subject, irrespective of their
actual motivations, intentions, etc.
19

This project claims to thereby secure a kind of objectivity regarding the normative
status of morality, namely, the very kind of objectivity presupposed by the commonplace
view of morality. It does so in part by appealing to the intersubjective, or interpersonal,
nature of moral properties and to the standardizing conditions necessitated by that very
nature. Hence it also does so in part by appealing to human nature, in particular, our

19
I shall also demonstrate how this view—viz., this Humean defense of the commonplace view of moral
normativity—is consistent with the account of moral obligation that Hume offers.

26
psychological disposition to be concerned about and moved by moral ends, especially
when certain conditions are met.
It is therefore important to distinguish the above claim to moral objectivity from
the somewhat stronger claim that moral properties are wholly attitude-independent. The
latter claim should be rejected insofar as it is incompatible with the fact that moral
properties are, in an important sense, dependent upon our capacity for sympathy, as well
as other dispositions to experience particular attitudinal states under certain conditions—
similar to the way in which an account of the nature of color properties is dependent upon
dispositions to experience particular visual states under certain conditions. In both cases,
the relevant states determine the domain of the property in question. Thus, although this
project does claim to explain and at least partially justify the commonplace
presupposition regarding the normative standing of morality, it nonetheless resists the
temptation to over objectify moral properties. The domain of moral properties cannot be
determined without appealing to certain attitudinal states.
§I.3.4 A proposal for a solution
In sum, this project contends that moral judgment has precisely the foundations
Hume claims. Moral judgments ascribe (non-minimal) moral properties, and these
properties depend upon and are explained by a shared psychology, in particular, a shared
capacity for sympathy (or empathy) that results in shared moral feelings when certain
conditions are satisfied. In the final assessment, this project claims to vindicate what is
arguably the key contention of Hume’s moral theory: a plausible analysis of moral
thought/judgment must ultimately appeal to these shared moral feelings. An analysis of

27
moral thought/judgment must ultimately appeal to these particular sentiments because,
according to the proposed account, the moral properties ascribed by our moral
evaluations are sentiment-dependent properties; i.e., the domain of moral properties is
determined by those sentiments (with certain conditions in place). Indeed, the main point
of the analogy with color and other secondary qualities is to draw out this important
necessary relationship between the property and the corresponding sensibility-state.
This Humean moral ontology, in conjunction with the indispensable distinction
between moral properties and value properties and the proffered semantic analysis, serve
as the bedrock of this project’s proposal for a solution to the accommodation problem.
They are its bedrock because they play a fundamental role in providing an account of the
content of moral thought and judgment that (i) explains and at least partly justifies our
commonplace objective presuppositions and (ii) answers the explanatory burdens raised
by these presuppositions’ positive commitments without being pulled away from those
commitments.
Finally, a crucial issue in the moral realism/antirealism debates is whether moral
realism is best conceived as a commitment to wholly attitude-independent properties, that
is, whether response-dependent properties fall within the domain of realism. How one
answers this important question will determine whether the proposed Humean theory is
rightly considered a brand of moral realism. But the upshot of this project is that it is
possible to secure moral objectivity irrespective of whether one answers that question
positively or negatively. Morality’s objective presuppositions can be accommodated in
either case.

28

PART ONE: Interpreting Hume’s Moral Theory

29

Chapter II: Hume’s Moral Ontology
§II.1 An overview
Our main goal in part one (chapters II-VII) of this project is to explain Hume’s
account of the nature of moral properties. This project interprets Hume as holding the
view that our moral evaluations are descriptive, property ascribing, truth-evaluable
judgments, some of which are true. There are moral properties and facts, and both moral
error and disagreement, including legitimate disputes over a given moral
judgment/belief’s truth-value, are not only possible but do actually occur. As Hume
conceives of them, moral properties are sentiment-dependent properties. He holds the
following general characterization to be true: an object x has a moral property if and only
if observation or contemplation of the (typical) effects of x elicits particular
feelings/sentiments in those who satisfy certain conditions.
20

Our first task in part one of this project is to introduce and defend the above
interpretation of Hume’s moral ontology. This task is taken up in this chapter. Our
second task (chapters III-VII) will be to provide further insight into the sentiment-
dependent nature of moral properties by explaining more precisely how this moral
ontology fits within Hume’s larger “system of morality”.
To this end, part one will argue that moral judgments, according to Hume’s
theory, are assessments of whether an object (e.g., an act or trait) fulfills or foils
particular ends, namely, moral ends. And an end belongs to the moral domain if and only

20
This characterization will serve for present purposes, but it will be refined later on.

30
if observation or contemplation of that end being fulfilled or foiled elicits particular
feelings/sentiments within those who satisfy certain conditions, namely, the conditions of
the “moral view”.
Crucial to this explanation is an account of the “true standard” for the moral
sentiments, namely, the “moral view” and its corrective conditions. However, as we shall
see, Hume’s notion of the “common” moral view can only be adequately explained by
first appreciating the role and relation of the “principles of humanity and sympathy”
referred to in the quotation below.
It appears…that, in our general approbation of characters and manners, the
useful tendency of the social virtues moves us not by any regards to self-
interest, but has an influence much more universal and extensive… And it
appears, as an additional confirmation, that these principles of humanity
and sympathy enter so deeply into all our sentiments, and have so
powerful an influence, as may enable them to excite the strongest censure
and applause. The present theory is the simple result of all these
inferences, each of which seems founded on uniform experience and
observation. (EPM, section V, part II)

As I read Hume, his moral ontology rests upon the principle, or capacity, of
sympathy and thus, in part, upon human nature.
21
In my view, the real import of Hume’s
moral ontology is that it takes the nature of morality to be rooted in psychology.
Chapters IV and V focus on the psychological capacity for sympathy and its role within
Hume’s moral theory, while chapters VI and VII focus on the regulative principles that
constitute the moral point of view. As we shall see, among other things, these principles
regulate sympathy.

21
I shall argue in chapter IV that the principle of “humanity” is quite closely related to the principle of
sympathy. It is a principle of benevolence, which depends upon and falls out of sympathy.

31
Along the way, we will consider and answer various objections to the proposed
interpretation of Hume’s moral theory. Upon responding to these objections, part one
concludes with the suggestion that Hume provides, or at least points the way to, a
promising moral ontology for resolving what I have called the accommodation problem.
Part two of the project defends this suggestion by indicating how it can do so. It is at this
point that the project transitions from an interpretive enterprise (part one) into a
reconstructive enterprise (part two). Part two is, in effect, an attempt to reinvigorate
Hume’s moral ontology in part by arguing that is congruous with much of the empirical
evidence being gathered in primatology, social psychology, and cognitive neuroscience.
But it does so primarily by demonstrating the considerable potential of Hume’s moral
ontology to resolve the accommodation problem and account for the intimate and rather
peculiar connection between moral judgment and human conduct.
§II.2 The distinction between moral sentiment & moral judgment
In order to defend the proposed interpretation of Hume, we need to distinguish
between moral sentiment and moral judgment. Following Hume’s lead, I shall use the
analogy with color as an expository device throughout our discussion. There are many
interesting similarities between the relation of our moral sensibilities, moral properties
and our judgments about them, and the relation between our color sensibilities, color
properties and color judgments.
22
For instance, Hume compares these two kinds of
sensibilities, properties, and judgments and observes the following similarity:

22
As we shall discuss later in this chapter, there are also some important dissimilarities as well.

32
…[W]hen you pronounce any action or character to be vicious, you mean
nothing more, but that from the constitution of your nature you have a
feeling or sentiment of blame from the contemplation of it. Vice and
Virtue, therefore, may be compar’d to sounds, colours, heat and cold,
which, according to modern philosophy, are not qualities in objects, but
perceptions in the mind: And this discovery in morals, like that other in
physics, is to be regarded as a considerable advancement of the
speculative sciences; tho’, like that too, it has little or no influence on
practice. (T 469)

I suggest that we should read this passage not as denying that objects have moral
properties but rather as denying that we can account for these properties without
ultimately appealing to our sentiments. In other words, Hume is claiming (albeit
somewhat crudely) that moral properties are impression-dependent properties, or what
has also been called secondary qualities. Because these properties have an essential tie to
our impressions, there is no saying what these properties are without appealing to our
sensibilities/impressions.
There is therefore an important sense in which physical objects do not possess
impression-dependent qualities such as color. To illustrate this sense, compare Hume’s
claim to what I think is the modern version of it, namely, that colors are not part of the
fundamental causal/explanatory structure of the world.
23
From this more objective
scientific viewpoint—viz., one that does not appeal to our sensibilities—objects are, in

23
Of course, unlike colors, science now has a conception of sound and heat that is part of the fundamental
causal/explanatory structure of the world, thus rejecting a necessary link between these properties and the
respective impressions of sound and heat that each property produces. But there is a conception of heat
that does lend itself to an impression-dependent account. This is our notion of heat that is related to the
sensation of heat as opposed to feeling cold. Of course, objects feel hot or cold, if they do, because their
mean kinetic molecular energy falls within a range that elicits that particular sensation in us. Thus, feeling
hot is a phenomenal quality of heat. Sound too has its own phenomenal qualities.

33
effect, colorless.
24
Or as Hume would put it, this physical description of the world

reveals that colors are not “qualities in objects”.
25

In order to defend this account as Hume’s, it is crucial that we understand and
employ the distinction between Hume’s conception of moral sentiment and his
conception of moral judgment. As we shall see, failure to do so, results in considerable
potential for misinterpreting many of Hume’s claims. To understand this distinction, we
will first need to consider Hume’s more basic distinction between impressions and ideas.
And the commonsense differences between an experience and a description of an
experience will help illustrate Hume’s distinction between impressions and ideas.
26

Consider, for instance, some differences between an experience of seeing red and
a description of an experience of seeing red. First, a description of such an experience is
referential in nature—being a description of a particular experience it must somehow
refer to that experience—whereas the experience itself is non-referential in nature.
Second, related to this is the fact that a description of an experience can be assigned a
truth-value, whereas the experience itself is truth-valueless. Third, there appears to be a
qualitative phenomenal difference in our immediate experience—e.g., our experience of
seeing a red object—and a description of such an experience. In other words, the

24
This point is also made by Darwall, Gibbard, and Railton (1997: 19).

25
Curiously, elsewhere Hume also claims that qualities such as color are “in” objects; see, e.g., T 2. This
chapter will offer an explanation of this apparent contradiction.

26
Capaldi (1989: 101ff.) suggests this comparison. As far as I am aware, Capaldi is one the first scholars
to carefully and systematically lay out Hume’s distinction between moral sentiment and moral judgment.
However, as we shall discuss in a moment, I disagree with Capaldi’s account of the moral ideas contained
in our moral judgments.

34
description arguably fails to capture vital phenomenal aspects of the immediate
experience.
These intuitive differences between our immediate experiences and descriptions
of our experiences help illustrate the differences found in Hume’s distinction between the
two modes of “perception”, namely, impressions and ideas.
27
For instance, in the
opening lines of the Treatise, Hume marks the distinction between ideas and impressions
by observing the difference with respect to their phenomenal “force”, or what he
elsewhere refers to as their “vivacity”.
All the perceptions of the human mind resolve themselves into two
distinct kinds, which I shall call IMPRESSIONS and IDEAS. The
difference betwixt these consists in the degree of force and liveliness, with
which they strike the mind, and make their way into our thought or
consciousness. (T 1)

This, however, is not the central distinguishing characteristic between ideas and
impressions, for some impressions can be “indolent” and weak, similar to the
phenomenal qualities of ideas, and as we shall discuss later, some ideas can be quite vivid
and vivacious, similar to the phenomenal qualities of impressions. Instead, the
distinguishing characteristic is that, on the one hand, our impressions—e.g., all of our
sensations and emotions— being “original existences” are non-representational and non-
referential in nature.
28
Our ideas on the other hand—e.g., our ideas of particular colors—
are representational and referential. They are “images” or mental reflections of our
original impressions. In some cases, these reflections “may very nearly approach” our

27
Ibid.

28
Hume says that his distinction between impressions and ideas doesn’t need much explanation since it is
just the intuitive one between “feeling and thinking” (T 2).

35
impressions in their likeness—e.g., our dreams or vivid memories (T 2).
29
When they do
so, our ideas will be so nearly as vivid and vivacious as our original impressions that they
will be almost indistinguishable from them.
Because of these referential differences in the nature of our impressions and ideas,
our impressions are never truth-evaluable. As Hume puts it:
A passion [which is a kind of impression] is an original existence, or, if
you will, modification of existence, and contains not any representative
quality, which renders it a copy of any other existence… ‘Tis impossible,
therefore, that this passion can be opposed by, or be contradictory to truth
and reason; since this contradiction consists in the disagreement of ideas,
considered as copies, with those objects which they represent. (T 415,
Hume’s emphasis)

On the other hand, because our ideas are referential/representational, Hume thinks
they can also be truth-evaluable. He writes:
Truth or falshood [sic] consists in an agreement or disagreement either to
the real relations of ideas, or to real existence and matter of fact.
Whatever, therefore, is not susceptible of this agreement or disagreement,
is incapable of being true or false, and can never be an object of our
reason. Now ‘tis evident our passions, volitions, and actions, are not
susceptible of any such agreement or disagreement; being original facts
and realities, compleat in themselves, and implying no reference to other
passions, volitions, and actions. ‘Tis impossible, therefore, they can be
either pronounced true or false, and be either contrary or conformable to
reason. (T 458)

So Hume conceives of ideas as representational mental entities. And as such,
according to Hume, all thoughts/judgments are constituted by one or more ideas.
30

29
Hume suggests that the contents of our dreams are ideas, but perhaps they can also include impressions.

30
In my view, Hume’s apparent psychologism about judgments is seriously problematic for reasons that
cannot be thoroughly explored here. Although we will also briefly touch on the problem in §III.1, it should
be said here that Hume is unaware of the importance of analyzing any judgment as first and foremost a
propositional entity rather than a mental entity. In opposition to Hume’s view, mental content—i.e., a
belief—is analyzed in terms of its propositional content. It is a matter of dispute whether Hume’s theory of

36
Hence all judgments are representational. They are either thoughts/judgments about the
relations of ideas or judgments of “matters of fact”—i.e., judgments of mere existence or
causes and effects.
31
(See §II.2.2). If they agree with the “real” relations of ideas or “real”
matters of fact—i.e., if things are as the judgments say they are—then they are true;
otherwise, they are false.
32
Our immediate experiences or impressions, on the other hand,
are “original existences”, which are neither representational nor referential. Hence they
are therefore truth-valueless.
Having briefly discussed the distinction found in Hume between impressions and
ideas, we are now in a better position to grasp the significance of the division between
moral sentiment and moral judgment. Moral sentiments are feelings of approval or
disapproval. In other words, they are a particular kind of impression.
33
Hence, on
Hume’s view, moral sentiments are non-representational, non-referential, and truth-
valueless. Moral judgments, on the other hand, are composed of ideas, at least one of
which is a moral idea. And since all ideas are representational, moral judgments are

judgment allows for propositional content. But even if it does, Hume would view a judgment’s
propositional content as first and foremost mental content. After all, Hume’s theory is based upon the
construction of complex ideas from simple ones. This, in my view, is a mistake. Fortunately, these issues
will have no significant impact on the present discussion. (However, see §III.1)

31
Existence is not a distinct idea from the idea of the object according to Hume. In other words, existence
is not a distinct property. Hence predications of existence are not representational; they aren’t genuine
predications. Hume’s example is God is (T 96-7n). (See also the following footnote.)

32
The issue of the truth-value of mere existence claims is more complicated since Hume treats existence as
merely a manner of conception rather than a property. Although existence is sometimes predicated of
objects, in ordinary logic existence is treated as a quantifier rather than a property. It could be argued that
Hume’s treatment anticipates the analysis that existence claims receive in modern logic. (Thanks to John
Dreher for help on this point.) In any case, I shall gloss over this issue since moral judgments are clearly
not mere existence claims.

33
As we shall discuss in §III.3, they are what Hume refers to as impressions of reflection.

37
representational. They are complete representational thoughts that are more than mere
claims of existence. Hence, in Hume’s view, moral judgments are truth-evaluable
descriptions.
§II.3 Debunking non-descriptivist interpretations of Hume
Some have questioned whether Hume offers an analysis of moral judgment as
truth-evaluable descriptions however. They suggest that Hume may well offer a non-
descriptive/emotive or expressivist theory.
34
But we now know that if there are moral
ideas, then we can use these moral ideas to form truth-evaluable moral judgments, in
which case Hume would clearly not be a non-descriptivist/emotivist. So the question is:
Does Hume hold that there moral ideas? Indeed he does. Consider the following:
The rule here holds without any exception, and that every simple idea has
a simple impression, which resembles it; and every simple impression a
correspondent idea. That idea of red, which we form in the dark, and that
impression, which strikes our eye in sunshine, differ only in degree, not in
nature… the case is the same with all our simple impressions and ideas…
But if anyone should deny this universal resemblance, I know no way of
convincing him, but by desiring him to show a simple impression, that has
not a correspondent idea. (T 3-4)

Hume claims that “all our ideas are deriv’d from correspondent impressions” (T
105). He classifies all ideas and impressions as either simple or complex. Hume states
that “simple perceptions or impressions and ideas are such as admit of no distinction nor

34
See, for instance, I. Hedenius (1937), J.N. Findlay (1944), his student A. N. Prior (1949), D.D. Raphael
(1947), P.H. Nowell-Smith (1954) who cites Raphael, Paul Edwards (1955) who cites both Raphael and
Prior, R.B. Brandt (1957), M.J. Scott-Taggart (1961), S. Blackburn (1984, 1993), A. MacIntyre (1988), and
F. Snare (1991). J.L. Mackie (1980) also initially suggests an emotive interpretation of Hume’s theory of
moral judgment, but ultimately settles for what he calls the “objectification theory”, which is a descriptivist
interpretation of Hume but also a kind of error theory. (See Capaldi [1989: 138-50] for a detailed account
of the historical development of emotive interpretations of Hume.) Almost all of the authors above
interpret Hume as inconsistent and thus confused in his claims about moral judgment. Of course the
principles of charitable exegesis suggest that we look for plausible alternative interpretations in which
Hume is neither confused nor inconsistent. I shall argue for just such an interpretation in what follows.

38
separation,” whereas complex impressions and ideas “are the contrary to these, and may
be distinguished into parts” (T 2). In other words, complex impressions are composed of
simple impressions and/or other complex impressions that can ultimately be broken down
into simple impressions; and complex ideas are composed of simple ideas and/or other
complex ideas that can ultimately be broken down into simple ideas.
According to Hume, every simple idea resembles a correspondent simple
impression. In the passage above, Hume suggests that our color impressions are simple.
They are simple because they “admit of no distinction nor separation.” Since simple ideas
resemble simple impressions, our ideas of our color impressions must also be simple.
Given the comparison Hume draws between colors and morals, the above passage
alone strongly suggests Hume holds that there are indeed simple moral ideas that “copy”
our simple moral impressions. Moreover, if our moral ideas of our moral impressions
were complex, then they could be broken down into the simple ideas that compose them.
But neither in the Treatise nor elsewhere does Hume ever discuss or even intimate at
what these simple ideas and corresponding impressions might be.
Nevertheless, this circumstantial evidence admittedly still leaves room for doubt.
Fortunately, Hume makes it clear that there are moral ideas when, toward the beginning
of the Enquiry Concerning Human Understanding, he explicitly uses a moral idea to
explain the relation between impressions and ideas. Hume writes:
When we think of a golden mountain, we only join two consistent ideas,
gold, and mountain… A virtuous horse we can conceive [i.e., have an idea
of]; because, from our own feeling, we can conceive [i.e., have the idea of]
virtue; and this [idea] we may unite to the [ideas of] figure and shape of a

39
horse, which is an animal familiar to us… all our ideas or more feeble
perceptions are copies of our impressions or more lively ones. (EHU 19)

So, whether simple or complex, there are moral ideas. Since all of our ideas are
derived from correspondent impressions, moral ideas are presumably derived (at least in
part) from moral impressions.
35
These ideas, like all ideas, are representational. They are
necessary on Hume’s account for thinking and arguing about, as well as “pronouncing”
or ascribing moral qualities and distinctions. In other words, moral ideas are constituent
parts of our moral judgments, which are complete representational thoughts and therefore
truth-evaluable. Thus, despite what some have claimed, Hume clearly offers a
descriptivist account of moral judgment rather than a non-descriptivist/emotivist account.
In sum, when reading Hume, we need to carefully distinguish between moral
sentiment/impressions and moral judgment/ideas. It is also important to distinguish
between Hume’s claims concerning the apprehension of moral distinctions via moral
sentiment and his claims concerning the thought/pronouncement of moral distinctions
(and their apprehension). Failure to do so will likely result in significant confusion.
§II.4 Two kinds of moral ideas
We have explained the distinction between sentiment and judgment by appealing
to Hume’s distinction between moral impressions and ideas. According to Hume, moral
impressions are “original existences” and our moral ideas are derived from them. All

35
I suggest that our moral ideas of our moral impressions are simple because our moral impressions are
simple; they cannot be broken down into simpler impressions. However, I shall argue that our moral ideas
of moral properties are complex ideas; they can be broken down into other ideas, including our simple
moral ideas of our simple moral impressions. Thus, we find in Hume a distinction between moral ideas of
our moral impressions and moral ideas of moral properties. The former are simple ideas whereas the latter
are complex. This account will be discussed and defended in this and the following sections.

40
moral judgments contain at least one moral idea. Hence all moral judgments are derived
in part from moral impressions. But we need to be careful at this step in the discussion
because, in Hume’s view, moral judgments are not simply judgments about the speaker’s
own immediate moral impressions, any more than color judgments are simply judgments
about the speaker’s own immediate color impressions. Hume’s account of moral
judgment is not a form of idiosyncratic, or speaker, subjectivism! Nevertheless, Hume’s
account of moral judgment has been interpreted as a form of idiosyncratic subjectivism in
which our moral ideas refer to our own moral impressions.
36

This is not Hume’s view. Although the moral ideas typically contained in our
moral judgments are derived from our moral ideas of moral impressions, I shall argue
Hume holds that they are, in fact, distinct. They are moral ideas of moral properties.
And our moral judgments attribute these properties to objects (e.g., traits or acts).
As I read Hume, an object has a moral property if and only if that object is such as
to elicit impressions of approval or (as the case may be) disapproval in anyone who
satisfies certain conditions.
37
So this account posits an essential link between moral
properties and moral impressions. In fact, there is no saying what moral properties are
without ultimately appealing to moral sentiments. Moral qualities and judgments are, in
this respect, analogous to color qualities and judgments. Our color judgments are clearly
not simply descriptions of, or judgments about, our own visual impressions. I suggest

36
See, for instance, C. L. Stevenson (1937), Geoffrey Hunter (1962), Philippa Foot (1963), Pall Ardal
(1966), and Jonathan Harrison (1976). This reading is also commonly found in brief synopses of Hume.

37
David Wiggins (1987) suggests something similar to this characterization as a possible, but unlikely,
interpretation of Hume. As we shall discuss later, Wiggins’ own proposal differs in one important respect
from the one offered by this project. He proposes a necessary link between value properties and our moral
sensibilities, whereas this project does not. (See chapter V for further discussion.)

41
that they are, in effect, judgments about the particular type of visual impression anyone
would experience were they to satisfy certain conditions. If this is correct, then there is
no saying what colors are without appealing to the relevant visual impressions (because
of the inherent connection between colors and our color impressions).
This impression-dependency analysis of colors accounts for the obvious fact that
colors are apprehended by our color sensibilities and not by reason. Similarly, moral
qualities are apprehended by our moral sensibilities rather than by reason alone, just as
Hume claims (T 455ff). This is why, when Hume asks at the beginning of Book III of the
Treatise “whether ‘tis by means of our ideas or impressions we distinguish betwixt vice
and virtue, and pronounce an action blameable or praiseworthy?” (T 456), he rightly
answers that moral distinctions are discovered—i.e., initially apprehended—by our
impressions and not by our ideas/reason.
It is noteworthy that, in answering this question, Hume does not deny that there
are moral ideas. We need moral ideas to think and argue about moral distinctions and the
nature of moral qualities. Analogously, our ideas of color enable us to think and argue
about color distinctions and the nature of color qualities. Of course we can also describe
and think about our own and others’ color experiences using the ideas that “copy” or
resemble our color impressions, just as we can describe and think about our own and
others’ moral impressions using moral ideas of moral impressions.
Thus, if my reading of Hume is correct, then a distinction between our moral
ideas of moral impressions and our moral ideas of moral properties is latent within
Hume’s moral theory. Although these two kinds of ideas are integrally related (because

42
moral properties and their corresponding impressions are integrally related), they are also
distinct. It is therefore important to distinguish between these two kinds of moral ideas in
order to fully understand Hume’s distinction between moral sentiment and moral
judgment—i.e., judgments that ascribe moral properties.
Hume admittedly does not explicitly draw the distinction between our moral ideas
of moral impressions and our moral ideas of moral properties. Consequently, it is easily
overlooked, especially since these two kinds of moral ideas are so closely related to each
other.
38
For instance, Nicholas Capaldi, who is one of the few scholars to carefully and
systematically discuss Hume’s distinction between moral sentiment and moral judgment,
neglects the distinction. Capaldi (1989: 113-4) suggests that Hume uses the first set of
expressions below to refer moral sentiments, whereas he uses the second set of
expressions to refer to moral judgments:
a. “distinguish moral good and evil” (T 456); “distinguish betwixt vice and virtue”
(ibid); the apprehension of “vice and virtue” (T 470); having “the sense of virtue”
(T 471); the “sentiment of blame or approbation” (EPM 291).

b. “pronounce an action blameable or praiseworthy” (T 456); “pronounce an action or
character to be vicious” (T 469); “pronounce the action criminal or virtuous” (EPM
291); “ascribe virtue” (T 472); “apply the terms expressive of our like or dislike” (T
582); “express that this man possesses qualities whose tendency is pernicious to
society” (EPM 272); “The very nature of language guides us almost infallibly in
forming a judgment of this nature; and as every tongue possesses one set of words
which are taken in a good sense, and another in the opposite” (EPM 174).

While Capaldi is correct in drawing a distinction between sentiment and judgment
regarding Hume’s claims, Capaldi’s assertion that the first set of expressions refers
directly to moral sentiment is mistaken, or so I shall argue. It noteworthy, however, that

38
As we shall discuss, the way in which Hume often presents his account often fosters confusion.

43
his mistake is not exactly the same as those who proffer an idiosyncratic interpretation of
Hume. Capaldi, like many others, holds that such expressions refer not just to one’s own
moral sentiments but moral sentiments that are shared or universal.
39

Regardless, I shall now argue that this account too is not Hume’s view. Rather,
Hume holds that the first set of expressions above refer to intersubjective, impression-
dependent moral properties, which are distinct from universal moral sentiments.
40

Whereas the non-descriptivist/emotivist misinterpretation of Hume is apparently due to a
failure to employ the distinction between moral sentiment and moral judgment, I argue
that the sort of misinterpretations of Hume just mentioned—viz., the view that moral
ideas contained in our moral judgments refer to either idiosyncratic or universal moral
sentiments—is the result of a failure to fully understand that distinction.
§II.5 Complex ideas of impression-dependent moral properties
Hume maintains that all moral judgments contain at least one moral idea, and all
moral ideas are derived from moral impressions. Hence all moral judgments are derived,
at least in part, from moral sentiments. It is noteworthy that when Hume speaks of moral

39
See, for instance, R. Firth (1952), J. Harrison (1976), B. Stroud (1977), S. Darwall (1983), R. Fogelin
(1985), C. Brown (1988), A. Baier (1991), E. Radcliffe (1994), C. Korsgaard (1996), K. Abramson (1999),
and J. Rawls (2000); both J. McDowell (1987a and b) and D. Wiggins (1987) also seem to suggest this sort
of reading. Note that there is an important distinction to draw within this sort of view: some hold that
moral expressions refer to our actual universal moral sentiments, while others hold that they refer to the
universal moral sentiments one would have were he or she ideally situated.

40
C.D. Broad’s (1944) discussion of Hume also suggests an impression-dependent account. As far as I am
aware, J. L. Mackie (1980) was the first in this century to explicitly defend this sort of reading, although he
claims that Hume also denies that such properties exist. D. F. Norton (1982) appears to be the first to argue
that, in Hume’s view, such response-dependent moral properties do exist. See also Norton (1984 and
1993), A. E. Pitson (1982, 1985), and N. Sturgeon (2001). G. Sayre-McCord (1994, 1996) also seems to
concur with this reading of Hume. Curiously, Capaldi (1989) explicitly says that moral distinctions are
secondary-qualities; yet he also repeatedly states that moral judgments refer to moral sentiments.

44
judgment, he does not conceive of it as primarily a linguistic entity since ideas are not
words. As discussed, Hume conceives of moral judgments as complete representational
thoughts. Moral judgments are strings of ideas, i.e., mental entities. Hume does claim,
however, that “we use words for ideas” (T 62), so the following passages indicate that the
moral sentiments associated with moral judgments are not simply the speaker’s own
immediate impressions of approval or disapproval.
41

The distinction, therefore, between these species of the sentiment being so
great and evident, language must soon be moulded upon it, and must
invent a peculiar set of terms, in order to express those universal
sentiments of censure and approbation, which arise from humanity, or
from views of general usefulness and its contrary. (EPM 274)

But when he bestows on any man the epithets of vicious or odious or
depraved, he then speaks another language, and expresses sentiments, in
which he expects all his audience are to concur with him… If he mean,
therefore, to express that this man possesses qualities, whose tendency is
pernicious to society, he has chosen this common point of view, and has
touched the principle of humanity, in which every man, in some degree,
concurs. (EPM 272)

Hume explicitly states that our moral utterances express “universal sentiments”.
However, there doesn’t appear to be anything incompatible with using moral language to
express universally shared impressions and also to express, at the same time, one’s own
immediate impressions. While this is true, Hume makes it clear that he rejects an
idiosyncratic subjectivism. In Hume’s view, we can form moral judgments that do not
express our own immediate moral impressions. We originally acquire our moral ideas
from impressions, but after we have acquired these ideas, we can use the terms that refer

41
In next chapter, I challenge Hume’s account of the relation between language, ideas, and impressions.

45
to these ideas in order to express universal moral sentiments even when we do not
immediately experience those sentiments. Hume writes:
In general, all sentiments of blame or praise are variable, according to our
situation of nearness or remoteness, with regard to the person blam’d or
prais’d, and according to the present disposition of our mind. But these
variations we regard not in our general decisions, but still apply the terms
expressive of our liking or dislike, in the same manner, as if we remain’d
in one point of view. Experience soon teaches us this method of
correcting our sentiments, or at least, of correcting our language, where
the sentiments are more stubborn and inalterable. (T 582)

The intercourse of sentiments, therefore, in society and conversation,
makes us form some general inalterable standard, by which we may
approve and disapprove of characters and manners. And tho’ the heart
does not always take part with those general notions, or regulate its love
and hatred by them, yet are they sufficient for discourse, and serve all our
purposes in company, in the pulpit, on the theatre, and in the schools. (T
603)

So, while our moral judgments can and frequently do express our own immediate
impressions of approval or disapproval, sometimes they do not. Sometimes our feelings
conflict with, and may even be contrary to, the moral judgments we make and the moral
feelings they express. In both of the above passages, Hume is suggesting that the terms
we deploy need not express our actual feelings at the time of utterance. If we understand
moral properties to be powers or dispositions of objects to elicit particular feelings in
anyone who satisfies certain conditions—i.e., if moral properties engender “universal
sentiments”—then we can readily explain why it is that utterances of moral judgments
frequently do express our own sentiments and why they sometimes do not.
One key issue here is deciphering what exactly Hume means when he says that
moral judgments “express” moral sentiments. Does he mean that our moral judgments

46
simply refer to universal moral sentiments, or does he mean that they somehow convey
these sentiments to others? I suggest that the latter is the correct reading of Hume. Moral
judgments are expressive of universal moral sentiments insofar as they convey these
sentiments to others in virtue of the impression-dependent nature of the moral properties
to which moral judgments refer. But moral judgments do not refer to those sentiments.
Once again the analogy with color is probative. Typically when we judge that an
object is a particular color, say red, we are at the same time experiencing a corresponding
visual impression, namely, an impression of red. This is because our visual impressions
indicate an object’s color, at least when certain conditions are met. Colors are
impression-dependent properties, and so our color impressions naturally guide our
judgments. Consequently, we typically assume that individuals have color impressions
that correspond to the color judgments they utter, but, of course, they need not. In either
case, judgments of color are still expressive of color impressions insofar as the domain of
colors is determined by color impressions. But they do not refer to our color impressions.
Similarly, because moral properties are impression-dependent, our moral
impressions naturally guide our moral judgments. Consequently, we typically assume
that individuals who utter moral judgments are also experiencing the corresponding
moral impressions.
42
Thus, utterances of moral judgments usually do express, or convey,
the speaker’s own moral feelings without explicitly referring to them, and others naturally

42
As we shall discuss in chapters IV and V, we are much more likely to experience moral feelings when we
issue moral judgments than we are to experience color impressions when we issue color judgments because
we can experience moral feelings even in those instances in which our moral feelings do not guide our
judgments; i.e., sometimes our moral judgments guide our moral feelings. This is due to the fact that moral
feelings can be engendered from our sense of moral duty. (See chapter V for further discussion.)

47
assume that those utterances tend to express the speaker’s feelings too. However, this is
not always the case. We can issue moral (or color) judgments without the corresponding
impressions, and we can also experience moral (or color) impressions that are not
indicative of moral (or color) properties, namely, when the requisite conditions are not
satisfied. Indeed, we may currently be experiencing feelings that countervail the moral
feelings indicated by our moral judgments. We might even morally condemn ourselves
because of this fact. Of course, others may not know this, and so assume that our
judgment is expressive of our actual feelings.
In any case, as Hume observes in the second to last passage quoted above, even in
such instances the moral terms we utter can still be expressive of approval or (as the case
may be) disapproval. I suggest that this is due to the intimate link Hume observes
between moral properties and moral sentiments. Because of this link, those who hear our
moral utterances will typically take them to be expressive of those universal moral
sentiments. This is significant because, by expressing these attitudes, our moral
judgments can thereby serve to put psychological pressure on others to choose and act
accordingly.
43
For example, suppose that I judge that some particular act is in fact
morally wrong. According to the proposed moral ontology, this fact will presumably
create an expectation that others “see” it (i.e. feel it) as immoral too. That is, they should
have the appropriate feeling, namely, the one indicative of the moral property I have
ascribed to the object.

43
Expressivists are quick to stress this feature of moral judgments. For example, S. Blackburn (1993)
claims that the significance and function of moral judgments lies in “expressing attitude, endorsing
prescriptions, or, in general, putting pressure on choice and action.”

48
In sum, as I read Hume, he holds that moral terms refer to moral ideas, but these
moral ideas do not refer to one’s own impressions or to universal moral impressions.
Instead, they refer to moral properties. Hume conceives of moral properties as having a
necessary tie to our moral impressions, which amounts to the claim that the domain of
these properties must be determined by the moral sentiments of anyone who meets
certain conditions. In other words, moral judgments refer to moral properties with an
intersubjective or interpersonal grounding that is rooted in our—not just my own—
impressions (subject to certain conditions). As discussed, this account explains why
utterances of moral judgments are typically indicative, or expressive, of the speaker’s
actual moral feelings. It also explains why our moral judgments can be expressive of
moral attitudes even when those attitudes are not the actual attitudes of the speaker.
It follows from this account of moral properties that an analysis of both moral
judgments and the moral properties to which they refer must ultimately appeal to moral
impressions. As we shall see in a moment, this conclusion is entailed by Hume’s
arguments against the rationalists, as well as his claims about the relation between
impressions and ideas, or so I shall argue. However, at this point it’s still not entirely
clear how, in Hume’s view, we acquire an idea of such an impression-dependent property
in the first place. So we need a better grasp of the relationship between moral
impressions, our moral ideas of those impressions, and our ideas of moral properties.
When assessing Hume’s account of the relation between our impressions and our
ideas of objects and properties (and thus the relation between sentiments and judgments),
it is important to keep in mind that, according to Hume’s empiricist epistemology, our

49
ideas of objects and properties ultimately depend upon our impressions. In other words,
our knowledge of the objects and properties that we come into contact with, and
subsequently think about, is ultimately the product of our impressions. However,
although all of our ideas are, in a sense, afterthoughts of our impressions, at least some of
our ideas must represent, or refer to, objects and properties. Otherwise we could not
think or talk about them. In short, we do not discover or apprehend our impressions;
rather we discover objects and properties through our impressions. However, given that
all of our ideas are derived from impressions, it is not immediately obvious how one’s
ideas ever come to refer to anything besides one’s own impressions.
Unfortunately, it is beyond the scope of this project to consider this matter at
length but a cursory sketch should suffice for our purposes. According to Hume, we
naturally suppose that either external objects and properties ultimately cause our
impressions, and/or various internal physiological processes do. Hume accepts this
“vulgar” view (T 187ff).
44
In other words, for Hume, the question is not whether we
should accept as fact that objects and properties cause the impressions that our ideas
“copy”. Although he does not think we can prove this belief beyond all skepticism,
45

44
Those who insist on a phenomenalistic reading of Hume will miss this important point entirely. Hume,
however, is not a pure phenomenalist. Although Hume does distinguish the contents of our minds by their
phenomenal “vivacity” or strength, he nonetheless thinks that this has physiological underpinnings. To his
credit, Hume does not speculate about these physiological processes however (T 275-6)—to his credit
because of the extremely limited understanding of those processes in the 18
th
century. Robert Anderson
(1966) argues persuasively for the materialist foundations of Hume’s philosophy. For an account of the
origins of phenomenalistic readings of Hume, see Capaldi (1975, chapter 3), in which he suggests the rise
of positivism is largely to blame. Interestingly, emotive readings of Hume also surfaced during this time.

45
Hume writes: “As to those impressions, which arise from the senses, their ultimate cause is, in my
opinion, perfectly inexplicable by human reason, and ‘twill always be impossible to decide with certainty,
whether they arise immediately from the object, or are produc’d by the creative power of the mind, or are
deriv’d from the author of our being… [Nevertheless] We may draw inferences from the coherence of our

50
Hume nonetheless claims that it is “a point, which we must take for granted in all our
reasonings” (T 187). So for Hume, the question is an epistemological one, namely: How
did we arrive at this conclusion in the first place?
The temptation, says Hume, is to take the “philosophical” view, which attempts to
infer objects and properties, and causes and effects, directly from our ideas of our
impressions. Hume argues against this approach. He claims that the only idea “that can
be trac’d beyond our senses, and informs us of existences and objects, which we do not
see or feel, is causation” (T 74, Hume’s emphasis). So the only way we can escape
beyond the ideas of our own present and past impressions is through causal reasoning.
46

In order to see how causal reasoning can accomplish this, we need to understand how we
initially arrive at ideas (and beliefs) of cause and effect.
According to Hume, these ideas (and beliefs) are not produced by reason alone;
indeed he argues that they cannot be. Nor are they produced merely by our impressions.
Our ideas of causal connections are, and can only be, arrived at by our impressions in

perceptions, whether they be true or false; whether they represent nature justly, or be mere illusions of the
senses” (T 84).

46
Hume also holds that causal reasoning is the most important kind of reasoning related to human conduct.
This is because reasoning about what to do or pursue always involves reasoning about matters of fact, and
Hume claims that all such reasoning resolves into causal reasoning. He writes:
If you were to ask a man, why he believes any matter of fact, which is absent; for instance, that
his friend is in the country, or in France; he would give you a reason; and this reason would be
some other fact; as a letter received from him, or the knowledge of his former resolution and
promises. A man finding a watch or any other machine in a desert island, would conclude that
there had been men in that island…If we anatomize all the other reasonings of this nature, we
shall find that they are based on the relations of cause and effect… (ECHU, 26-7)

51
conjunction with other principles of human nature, namely, custom and habit, as well as
memory and imagination.
47
Consequently, Hume concludes that:
Custom, then, is the great guide of human life. It is a principle alone
which renders our experience useful to us, and makes us expect, for the
future, a similar train of events with those which have appeared in the
past. Without the influence of custom we should be entirely ignorant of
every matter of fact beyond what is immediately present to the memory
and senses. We should never know how to adjust means to ends, or to
employ our natural powers in the production of any effect. There would
be an end at once of all action, as well as of the chief part of speculation.
(Enquiry Concerning Human Understanding, 44-5)

Hume’s view is that through repeated and reproducible interaction of human
beings with their environment, we develop habits of mental association between ideas
that are “constantly conjoined” due to those past and present impressions—ideas that
Hume, in effect, argues have no conceptual connection, which is why reason cannot
produce them. Hume then argues that our belief in external objects and their properties is
due to these habits of mental association, including the association of our ideas of
relations of causation, contiguity, coherence, and constancy among our impressions.
48

Roughly, these principles of mental association give rise to a conception of persisting
objects and properties as the cause of some of our impressions.

In other words, we are
naturally disposed to attribute causes to our impressions, causes that in some cases must

47
Although custom and habit is, in effect, a non-rational human disposition, cause and effect is a
philosophical as well as a natural relation for Hume because of the operation of the imagination and the
role of reflection not only in producing our ideas of cause and effect but also in evaluating our causal
reasoning. Causal reasoning can be evaluated, according to Hume, using general rules by which to judge of
causes and effects. Hume lists 8 rules (T 173-4). These general rules and the role of imagination in
producing them are discussed in chapters VI and VII.

48
Hume argues (T 194-206, 265) that the source of these habits of mental association is the imagination.

52
have a continued identity even though they are not continually perceived.
49
We then
posit various other causal accounts to help explain their continued existence.
50
Although
we are justified in believing these accounts, Hume stresses that we cannot prove them.
51

In the case of moral (as well as color) properties, we acquire our ideas of these
properties based upon the similarity of not just one’s own but our motivational (or visual)
experiences under certain conditions both across time and compared to one another. We,
in effect, observe that the same external objects cause the same effects in all of those who
meet certain conditions (e.g., conditions such as having a conception of self and other).
52

In roughly this way, custom and habituation enables us to step outside of our own
personal experiences and form ideas of cause and effect, other selves, shared or
“universal” impressions, external objects, and qualities of those objects, including

49
We are disposed to attribute a continued identity to objects and properties in part because of the continual
“contradictions” that would otherwise arise (given our habits of mental association); see, e.g., T 190-209.

50
My aim here is not to defend this account but merely to demonstrate that Hume attempts to give a
framework within which we are justified in believing in objects and properties that cause impressions.

51
In fact, Hume appears to hold the stronger claim that we cannot consistently deny such causal accounts.
In my view, Hume’s position is that the skeptical arguments he presents in Book I of the Treatise illustrate
the limits of reason. These skeptical arguments depend upon premises that the skeptic herself, if consistent,
ultimately calls into question. Hume demonstrates this by describing how the skeptic’s own actions
undermine the conclusions for which she persuasively argues. Thus, Hume does not so much as refute the
skeptic as defend some of our commonsense beliefs (e.g., the belief in an external world) as a starting-
point, that is, “a point, which we [including the skeptic] must take for granted in all our reasonings.” On
this point, see T 218. G. E. Moore (1925) famously defends this sort of “common sense” position nearly
two centuries later. For further discussion of Hume’s skepticism, see also Capaldi (1989), Ainslie (2003),
and D. Owens, “Scepticism with regard to reason” (forthcoming).

52
This account sheds light on why moral (as well as color) properties presuppose a common perspective
that, in effect, imposes a particular set of conditions on all of us. But it also raises a host of other questions
such as why we would be motivated to take up the common perspective rather than say our own
idiosyncratic perspective, or why we would privilege that common perspective and its particular
conditions—rather than some other perspective and set of conditions—in such a way as to claim that it
reveals an object’s “true” or “real” color or moral status. These important issues are addressed later in part
one and again in part two.

53
qualities that have an interpersonal or intersubjective grounding inherently rooted in our
impressions—i.e., impression-dependent properties of objects.
53
Hume sums up and
illustrates this basic approach in the passage below.
‘Tis therefore by EXPERIENCE only, that we can infer the existence of
one object from that of another. The nature of experience is this. We
remember to have had frequent instances of the existence of one species of
objects; and also remember, that the individuals of another species of
objects have always attended them, and have existed in a regular order of
contiguity and succession with regard to them. Thus we remember to
have seen that species of object we call flame, and to have felt that species
of sensation we call heat. We likewise call to mind their constant
conjunction in all past instances. Without any farther ceremony, we call
the one cause and the other effect, and infer the existence of the one from
that of the other. In all those instances, from which we learn the
conjunction of particular causes and effects, both the causes and effects
have been perceiv’d by the senses, and are remember’d: But in all cases,
wherein we reason concerning them, there is only on perceiv’d or
remember’d, and the other is supply’d [by the imagination via habit and
custom] in conformity to our past experience. (T 87)

It is noteworthy that in this passage Hume distinguishes between the object,
namely, the flame, which is the cause, and the sensation, namely, heat, which is the
effect. This is noteworthy because it may initially appear to contradict the interpretation I
am propounding. That is, given Hume’s comparisons between morals and secondary
qualities such as heat, color, sound, etc., this passage seems to indicate that our moral
ideas are ideas of universal moral impressions rather than ideas of impression-dependent
moral properties. After all, Hume states that heat is the effect and a sensation, not the
cause of the sensation. So a crucial issue at this point is whether the proffered account of

53
Again, just as Hume thinks that one cannot prove the existence of external objects, he also thinks that it is
impossible to prove that objects have such causal properties. But Hume does not think that this fact
provides any reason to reject belief in their existence. Indeed he thinks that we cannot consistently deny
that there are external objects or that external objects can possess such causal properties.

54
moral properties is really Hume’s. Indeed, in the moral case, Hume explicitly speaks of
“gilding or staining all natural objects with the colours borrowed from internal
sentiment”, which prima facie suggests psychological projection rather than genuine
property attribution. If this is in fact the case, then it certainly seems that, contrary to my
claim, Hume does not have an impression-dependent conception of moral properties.
In defense of my interpretation, I shall argue that Hume is using the metaphor of
“gilding” and “staining” to stress moral properties’ impression-dependent nature and not
to deny that such qualities can genuinely be ascribed to objects. In other words, Hume’s
point is not that the domain of moral properties is identical to its corresponding
impressions, but rather that the domain of moral properties, like other impression-
dependent qualities, is determined by the requisite corresponding impressions. To
recapitulate: an object has a secondary quality if and only if that object is such as to cause
the relevant corresponding impression in anyone who satisfies certain conditions. For
instance, assuming that heat is an impression-dependent property, an object has the
property of being hot if and only if that object causes the sensation of heat in anyone who
satisfies certain conditions.
54

Given this account of heat as an impression-dependent quality of objects, the fact
that flames cause an impression of heat is no objection to reading Hume as also being
able to claim that the flame is hot—i.e., that the flame has the impression-dependent

54
In this case, the relevant conditions would include, for instance, that the observer be relatively close to
the object, have a properly functioning sensory apparatus (e.g., the observer couldn’t be paralyzed), etc.

55
property of being hot.
55
To the contrary, it is precisely because flames are such as to
cause the sensation of heat in us that we legitimately attribute the impression-dependent
property of heat to the flame.
56

The analogy with color is, once again, probative. Few doubt that our color
judgments genuinely ascribe colors to objects. Moreover, the colors we ascribe to objects
are not identical to our color impressions since we can accurately ascribe colors to objects
when we aren’t looking at them. Indeed colors cannot be identical to color impressions
since colored objects (or at least the vast majority of them) quite obviously do not have
color impressions. But the colors of objects are nonetheless delimited by reference to our
color impressions. It is precisely because various objects are such as to cause particular
color impressions in us when we satisfy certain conditions that we attribute those colors
to them. Hence, as long as we meet the requisite conditions, the impressions we
experience are indicative of the property.
Similarly, as long as the conditions of the moral point of view have been met, our
moral feelings are indicative of moral properties. We “discover” or “apprehend” moral
qualities by means of our impressions, just as Hume claims (T 470ff). This is because
moral properties, being impression-dependent properties, have an essential link with
moral impressions, even though they are clearly distinct from those impressions. We
attribute moral properties to objects precisely because those objects are such as to cause
particular impressions (viz., moral impressions) under certain conditions.

55
As previously mentioned, science has a conception of heat of which an impression-dependent analysis is
implausible. But we also have a notion of heat linked to a phenomenal quality of feeling (i.e., feeling hot).

56
Indeed Hume (T 194) distinguishes the heat that persists “in the fire” from the sensation the heat causes.

56
So as I read Hume, our moral ideas of our moral impressions (as well as our ideas
of color impressions) are simple because our moral impressions (and color impressions)
are simple. However, our moral ideas of moral properties (as well as our ideas of colors)
are complex ideas. These complex moral ideas are formed from repeated and
reproducible experiences according to which we recognize a constant conjunction
between certain objects (e.g., acts and traits) and our moral impressions—that is, not just
our own moral impressions but also the moral impressions of anyone who satisfies certain
conditions. From our ideas of this relation between these impressions and objects, we
acquire complex moral ideas of impression-dependent moral properties. These complex
moral ideas are constituted by our ideas of objects, our simple moral ideas of our own
moral impressions, and our moral ideas of others’ moral impressions.
We then go on to infer, through our imagination in conjunction with habit and
custom, a cause and effect relationship between various objects and our moral
impressions. Our moral ideas of moral properties are therefore distinct from, yet
interrelated with, our moral ideas of moral impressions. This is due to the
characterization of moral properties as impression-dependent qualities: an object has a
particular moral property (e.g., an act is immoral) if and only if that object causes the
relevant moral impression (e.g., the feeling of disapprobation) in anyone who satisfies
certain conditions.
57
This analysis of moral properties as impression-dependent
properties not only accounts for their intersubjective nature; more importantly for our
present purposes, it demonstrates that Hume can deploy the notion of causation in the

57
It follows from this account that I will have a certain motivational experience because an act is, e.g.,
immoral; just as the fact that an object is, e.g., red explains why I have a certain visual experience.

57
moral case to explain how we can get outside of our own impressions to form ideas of
moral properties as distinct from our ideas of our own moral impressions.
58

Thus far I have argued that Hume attempts to provide a theoretical framework
within which it is at least possible for us to form complex ideas of various qualities that
have an interpersonal or intersubjective grounding inherently rooted in “universal”
impressions. But is there stronger textual evidence to support my contention that Hume
actually conceives of the moral ideas typically contained in our moral judgments in this
way? Or, to put the question in a slightly different light: Is there further textual evidence
that Hume conceives of moral qualities as distinct from moral impressions? If there is,
then this will presumably also count as evidence that Hume is committed to a distinction
between our moral ideas of these moral qualities and our moral ideas of moral
impressions.
In what follows, I shall argue that the textual evidence does indeed suggest that
Hume is most plausibly read as conceiving of moral qualities as impression-dependent
properties, which are distinct from, though inherently connected to, moral impressions.
Hence there are two kinds of moral ideas.
§II.6 Bolstering the impression-dependency interpretation
A quick overview of some of Hume’s claims reveals why there has been no end to
the exegetical disputes concerning Hume’s conception of qualities such as moral and

58
In chapter V, I shall argue that Hume’s view can and does allow for the possibility that our moral beliefs
(viz., that a moral property obtains in a given situation) can play a vital role in the production of moral
impressions within us. (Note that this possibility marks one noteworthy way in which the analogy with
color breaks down.) Nonetheless, in Hume’s view, that possibility is ultimately explained by the fact that
we initially form moral beliefs (e.g., that an act is morally wrong) because we experience certain moral
feelings. (In this case, the analogy with color is informative.) See chapter V for further discussion.

58
aesthetic beauty. The debate has been ongoing precisely because Hume makes so many
seemingly incompatible and even contradictory statements about such qualities. For
instance, Hume states that “natural” and “moral” beauty and deformity are “not qualities
in objects”.
59
And a few paragraphs earlier, he refers to such qualities as color as a
“phantasm of the senses” (SOT, paragraph 12). These claims suggest that Hume
conceives of moral, aesthetic, and color qualities, not as impression-dependent properties,
but as identical to particular impressions. In other words, they suggest that we are merely
projecting our sentiments onto objects.
However, in the very same sentence in which he claims that color is a “phantasm
of the senses”, Hume also boldly speaks of objects’ “true and real colour”. Elsewhere
Hume contends that the quality of beauty “inheres” in objects—Hume’s example is a
beautiful house (T 279).
60
And later he makes the equally strong claim that “beauty is an
order and constitution of parts... fitted to give a pleasure and satisfaction to the soul” (T
299). These claims suggest that Hume conceives of moral and aesthetic beauty as being
constituted by the intrinsic qualities of objects.
Curiously, only a sentence later Hume concludes: “Pleasure and pain, therefore,
are not only necessary attendants of beauty and deformity, but constitute their very
essence” (T 299). And a bit further in the same paragraph, he states: “we may conclude,
that beauty is nothing but a form, which produces pleasure, as deformity is a structure of
parts, which conveys pain; and since the power of producing pain and pleasure make in

59
Of the Standard of Taste [henceforth SOT], paragraph 16.

60
See also T 2 in which Hume refers to qualities such as color as being “in” the object.

59
this manner the essence of beauty and deformity, all the effects of these qualities must be
deriv’d from the sensation” (ibid). Hume also repeatedly speaks of moral qualities
“causing”, “producing”, “exciting”, “occasioning”, “conveying”, and giving “rise” to
moral impressions.
61
Finally, Hume refers to moral and aesthetic beauty as a “relation,
which nature has placed between the form and the sentiment” (SOT, paragraph 10).
It is prima facie perplexing that Hume makes such seemingly incompatible
claims, often within the span of a few paragraphs, sentences, and in some cases even
within the same sentence. An important issue among Humean scholars is how to
adjudicate among these claims. The proffered account, according to which Hume
distinguishes between two kinds of moral ideas, appears to have an exegetical advantage
over alternative readings such as the common interpretation presented earlier. This “two
moral ideas” account is able to make sense of the above claims in a manner that does not
render Hume as holding contradictory or competing positions.
According to that account, the last of Hume’s claims listed above most accurately
describes his position. He conceives of moral qualities (and other “secondary qualities”)
as, in effect, a causal relation between objects and particular sentiments. With this
conception in hand, we can explain the apparent tension in all of Hume’s claims listed
above as follows. In some claims, Hume only stresses the sentiment side of the relation,
which makes him sound as if he is suggesting that moral qualities are reducible to our
moral impressions. Other claims only stress the object side, which makes Hume sound as
if he is suggesting that moral qualities are reducible to the intrinsic qualities of objects.

61
See, for instance, T 294-303 and T 470-3.

60
Yet other claims stress both sides in turn. Consequently, because Hume is often not
entirely clear about his position, this can give rise to the appearance that Hume issues
various competing, or even contradictory, claims. I suggest that this apparent tension is
simply the result of Hume struggling to draw out the relation between objects and our
sentiments that is inherent in our ideas of moral qualities.
A significant part of the problem is that there is no way to pinpoint precisely
where moral qualities reside since, according to the relational account, they can neither
be placed fully in the object, nor fully in the sentiment. They are one example of Hume’s
maxim: “an object may exist, and yet be no where” (T 235). Nonetheless, Hume clearly
tends to stress the role of sentiment in this causal relationship, which explains why Hume
is frequently read as holding that moral qualities are identical to moral impressions.
While this is not Hume’s view, there is an important reason why he frequently and often
emphatically emphasizes the side of sentiment: Hume wants to warn against confusing
moral and aesthetic qualities with relations of ideas that are discovered by reason alone.
62

Hume argues that the “essence of morality” does not consist merely in the
relations of ideas or “matters of fact” discoverable solely by the understanding (T 468).
For instance, moral qualities are neither akin to mathematical relations, nor are they
identical with, or reducible to, objects’ intrinsic (or “absolute”) qualities. In short, moral
qualities are not merely objects of reason. Insofar as moral qualities consist of a relation
between objects’ intrinsic properties and particular sentiments, they are also objects of

62
See, for instance, T 455-70.

61
feeling. As such, there is an important sense in which moral qualities are sentiment-
dependent properties. Hume stresses this dependency in the following passage:
But can there be any difficulty in proving, that vice and virtue are not
matters of fact, whose existence we can infer by reason [alone]
63
? Take
any action allow’d to be vicious: Willful murder, for instance. Examine it
in all lights, and see if you can find that matter of fact, or real existence,
which you call vice... The vice entirely escapes you, as long as you
consider the object [alone]. You can never find it, till you turn your
reflexion into your own breast, and find a sentiment of disapprobation,
which arises in you, towards this action. Here is a matter of fact; but ‘tis
the object of feeling, not of reason. (T 468-9)

As Hume remarks, moral sentiments are directed toward an object, for instance, a
particular trait or intention. Whereas the sentiment is a “matter of fact” that is an object
of feeling, the object’s form is a “matter of fact” that is an object of reason.
64
Our ideas
of moral qualities include both kinds of “matter of fact”. “Thus”, says Hume, “the course
of the argument leads us to conclude, that since vice and virtue are not discoverable
merely by reason, or the comparison of ideas, it must be by means of some impression or
sentiment they occasion, that we are able to mark the difference betwixt them” (T 470).
Two points about Hume’s conclusion are worthy of emphasis. First, Hume in
effect claims that reason, or the comparison of ideas, does play a role in the apprehension
of moral qualities. If moral qualities were simply identical to our moral impressions, then
there would presumably be no role for reason in their discovery. Second, Hume speaks
of moral qualities “occasioning,” that is, causing, moral impressions. This too supports

63
It is important to understand the context in which this passage occurs: Hume is addressing the moral
rationalists’ position, namely, that moral qualities can be inferred, or apprehended, solely by reason. Hence
my addition of ‘alone’ is harmless since it is already implicit given the context. In support of this reading,
consider Hume’s conclusion (T 470) provided in the paragraph below.

64
See §V.6 for discussion of what I suggest is Hume’s conception of an object’s “form” in moral cases,
that is, what Hume conceives of as the form of a particular trait or act.

62
the impression-dependency interpretation. Hume appears to conceive of moral qualities
as a relation between an object and the particular kind of impression caused by the object.
According to this account, “morality, therefore, is more properly felt than judg’d
of”, just as Hume claims (T 470). For without these feelings, we could not apprehend
moral qualities—not to mention that, in Hume’s view, we would not have any moral
motivation to act in accordance with our moral judgments without experiencing moral
feelings. Moreover, and perhaps more importantly, without moral sentiment there would
not be a relation between the object and the sentiment; i.e., there would be no moral
qualities.

Thus, moral sentiment “raises in a manner a new creation” (EPM, Appendix I,
paragraph 21), at least when that sentiment is caused by an object under standard
conditions.
65

In short, Hume holds that moral qualities are the product of the interplay between
objects’ qualities and particular sentiments of qualified observers, that is, observers who
have “a perfect serenity of mind, a recollection of thought, [and pay] due attention to the
object” (SOT, 270).
66
Hume illustrates this interplay between object and sentiment in the
following passage:
This doctrine will become still more evident, if we compare moral beauty
with natural, to which in many particulars it bears so near a resemblance.
It is on the proportion, relation, and position of parts, that all natural
beauty depends; but it would be absurd thence to infer, that the perception

65
This explains why Hume sometimes claims that moral feelings are the cause of virtue and vice, and why
other times he claims that virtue and vice are the cause of our moral feelings (see, e.g., T 296ff). As we
shall discuss, it is also in this respect that moral qualities can be considered a “phantasm” of the senses, to
wit: they are in large part, but by no means entirely, the product of our sentiments. Although we may
mistakenly take them to be intrinsic qualities of objects, they are in fact impression-dependent qualities.

66
This account of being qualified is not to be confused with some ideal observer theories. Hume’s view is
not ideal insofar as humans, as they actually are, can count as qualified; see, e.g., G. Sayre-McCord (1994).

63
of beauty, like that of truth in geometrical problems, consists wholly in the
perceptions of relations, and was performed entirely by the understanding
of intellectual faculties... [I]n all decisions of taste or external beauty, all
the relations are beforehand obvious to the eye; and we thence proceed to
feel a sentiment of complacency or disgust, according to the nature of the
object, and disposition of our organs. (EPM, Appendix I, paragraph 13,
emphasis added)

It is noteworthy that here too Hume states that the quality “depends” in part on the
form constituted by the object’s parts, and thus partly “consists in” perceptions of
relations. In other words, moral qualities are determined in part by the nature of the
object itself. Hence Hume’s claim, quoted earlier, that “beauty is an order and
constitution of parts... fitted to give a pleasure and satisfaction to the soul” (T 299).
Consequently, moral qualities are in part an object of the understanding.
Of course Hume stresses, in opposition to the moral rationalists, that moral
qualities are also objects of feeling. They not only depend on the object’s nature; they
depend upon human nature. Natural and moral beauty and deformity consists in both the
form of the object and also the feeling—i.e., the pleasure or pain—produced. This is
why Hume states both that an object’s “figure and appearance” are “essential” to beauty
and deformity, and also that “pleasure and pain...constitute their essence” (T 299). A few
sentences later, Hume is somewhat more explicit that both form and feeling are essential
when he writes: “beauty is nothing but a form [of an object], which produces pleasure, as
deformity is a structure of parts, which conveys pain; and since the power of producing

64
pleasure and pain make in this manner the essence of beauty and deformity, all the effects
of these qualities must be deriv’d from the sensation” (ibid).
67

We can restate Hume’s view this way: virtue and vice, beauty and deformity, true
and false wit, and the like, are all dispositions of objects to cause particular pleasures and
pains in observers under certain conditions. Hence Hume writes: “Where any object has
a tendency to produce pleasure... it is always regarded as beautiful; as every object, that
has a tendency to produce pain, is disagreeable and deform’d” (T 576). Since such
qualities are nothing more than “powers” or proclivities of objects to produce particular
pains and pleasures, Hume claims that there is no way to judge whether an object has
such a dispositional quality without appealing to the particular sentiment associated with
it. “’Tis only by taste we can decide concerning it, nor are we possest of any other
standard, upon which we can form a judgment of this kind” (T 297). Nevertheless, the
sentiment is by no means identical to an object’s dispositional quality.
In Hume’s view, there is also no way to acquire the idea of the dispositional
quality without also appealing to the object itself, for the object has that quality or
tendency—i.e., the object is such as to elicit the relevant feeling—in virtue of its form.
For instance, Hume claims that “a great part of beauty [be it natural or moral] is deriv’d
from the idea of [the object’s] convenience and utility” (T 299), and this idea is derived
from our idea of the object’s form, which is a product of at least some of its intrinsic

67
A. E. Pitson puzzles over the fact that Hume makes each of these claims because he takes them to be
incompatible. He writes: “Clearly, the essence of beauty cannot consist both in pleasure and also in a
power, in virtue of its form, to produce pleasure” (1989: 74). I suggest that he misunderstands Hume’s
claims. Hume’s point is that this power, or disposition, of the object is to be conceived as a causal relation
between the object’s form and the pleasure it causes. Hence both are essential.

65
properties (ibid). Reason therefore plays a vital role in determining whether the object,
given its form, is useful, that is, whether the object is such as to bring about ends that are
pleasurable to its possessor or those who “have commerce” with her.
68
Hume writes:
“One principal foundation of moral praise being supposed to lie in the usefulness of any
quality or action, it is evident that reason must enter for a considerable share in all
decisions of this kind; since nothing but that faculty can instruct us in the tendency of
qualities and actions…” (EPM, Appendix I, paragraph 2). The tendency of which Hume
speaks is the tendency of various qualities and actions to fulfill a given end or a set of
ends. After all, according to Hume, what makes a quality or action useful is the fact that
it tends to fulfill an end that produces pleasure. He writes: “Usefulness is only a tendency
to a certain end”, namely, an end that is agreeable to its possessor or to those associated
with her (EPM, Appendix I, part II, paragraph 2).
Thus, in Hume’s view, we first consider the object in question and use our reason
to judge whether it, in virtue of its form, tends to fulfill (or foil) various ends, and thereby
cause pleasure (or pain) in its possessor or those associated with her. Our moral feelings
then respond to these reasoned evaluations about the tendency of objects. Hume writes:
“we must allow, that the reflecting on the tendency of characters and mental qualities, is
sufficient to give us the sentiments of approbation and blame” (T 577).
69

68
Concerning beauty, Hume writes: “Handsome and beautiful, on most occasions, is not an absolute but a
relative quality, and pleases us by nothing but its tendency to produce an end that is agreeable. The same
principle produces, in many instances, our sentiments of morals” (T 577).

69
And elsewhere Hume states that the “distinction between what is useful, and what is pernicious…is the
same in all its parts, with the moral distinction” (EPM, Section VI, part I, paragraph 5).

66
As we shall discuss in detail later in chapter IV, it is the psychological process of
“sympathy” that converts the appraiser’s idea of another’s pleasure (or pain) into a
feeling of approval (or disapproval) toward the object. For instance, Hume writes:
“Wherever an object has a tendency to produce pleasure in the possessor, or in other
words, is the proper cause of pleasure, it is sure to please the spectator, by a delicate
sympathy with the possessor” (T 576-7, his emphasis). This makes sympathy “a very
powerful principle of human nature” according to Hume because this psychological
process plays a central role in accounting for the interplay between form and feeling, and
this relation between form and feeling just is what constitutes moral qualities. In other
words, without sympathy there would be no moral qualities.
Of course, our moral feelings are only indicative of moral distinctions when
“standard” conditions are satisfied.
70
As we shall discuss in chapters VI and VII, the
primary purpose of many of these standard conditions, or “general rules”, is to regulate
sympathy’s natural variability so as to produce stability and uniformity in observers’
feelings and judgments.
71
Without at least some grasp of these standard conditions,
observers cannot appeal to their feelings as reliable indicators of moral distinctions. In
fact, if one does not grasp that standard conditions apply, then it appears that one cannot
have a proper understanding of moral qualities, for an object’s moral quality does not
vary in conjunction with our immediate moral impressions any more than an object’ s

70
See, for instance, T 472.

71
These general rules do not eliminate sympathy’s natural variability altogether however. On Hume’s
view, the moral view inculcates a measure of variability, but it does so in such a way that it still ensures
stability and uniformity of experience and judgment. See chapter VII for further discussion.

67
color varies in conjunction with variation in our immediate color impressions. Such
standard conditions are what require observers to reflect upon the relationship between
our impressions and the objects, along with their qualities, that cause the impression.
Indeed, the fact that Hume appeals to standard conditions to explain how we can
and do correctly attribute moral qualities to objects, even when we fail to have an
occurrent moral impression that corresponds to the moral quality, provides further
evidence that he holds a dispositional, impression-dependent account of moral qualities.
It thereby also provides evidence that Hume (at least implicitly) distinguishes between
our ideas of moral qualities and our ideas of moral impressions.
72
To acquire an idea of
an object’s moral quality, we must understand that even if we do not experience an
occurrent moral impression that corresponds to the presence of the relevant moral
property, the object would cause that impression were it observed under standard
conditions. Only by grasping this point can we judge accurately. For example, even if
our immediate moral feelings differ with respect to distinct actions due to the natural
variability in our sympathy, Hume claims that we nonetheless “blame equally a bad
action, which we read of in history, with one perform’d in our neighborhood t’other day:
The meaning of which is, that we know from reflexion, that the former action wou’d
excite as strong sentiments of disapprobation as the latter, were it plac’d in the same
position” (T 583-4).
So Hume appears to hold that an action is as equally morally bad as another if and
only if both actions are such as to “excite”, or cause, an equally strong sentiment of

72
See, for example, T 582 and T 603.

68
disapproval in us under standard conditions. Hence we can conclude that an action is
morally bad if and only if it is such as to cause a sentiment of disapproval in us—i.e.,
anyone—under standard conditions. Generalizing this claim, we get: an object (e.g., a
particular act or trait) has a moral quality if and only if it is such as to cause, under
standard conditions, moral feelings. And this latter claim is, of course, consistent with an
account of moral qualities as “powers”, or dispositions, of objects to cause particular
feelings under standard conditions. Moral qualities, as such, depend upon both the
object’s nature and human nature, both form and feeling.
To sum up, as I read Hume our ideas of these causal dispositions or “powers” of
objects consist in both the form of the object—i.e., the cause—and the feeling
produced—i.e., the effect—; for we cannot comprehend these causal dispositions without
having ideas of the form and feeling involved.
73
Thus, based upon my understanding of
Hume’s account of moral qualities, our ideas of these moral qualities are complex ideas.
They are composed of ideas of: objects, universal moral impressions, and cause and
effect, or at least an idea of a constant conjunction between the objects and impressions.
74

Hume is therefore plausibly read as implicitly drawing a distinction between two
kinds of moral idea, namely: ideas of moral impressions, which are presumably simple
ideas, and ideas of moral qualities, which are complex ideas that contain our ideas of

73
For instance, Hume writes: “’Tis only when a character is considered in general, without reference to our
particular interest, that it causes such a feeling, or sentiment, as denominates it morally good and evil” (T
472).

74
This is not to say that people cannot misconceive of moral qualities. For instance, Hume thought that the
“vulgar” frequently conceive of moral qualities as “primary” or intrinsic properties of objects. They have
“obscure” ideas of moral qualities due to various kinds of prejudice. My claim is that, in Hume’s view,
“clear” ideas of moral qualities are the complex ideas described here.

69
moral impressions among other ideas. Moreover, I have argued that this interpretation is
preferable on grounds of charitable exegesis. This interpretation, according to which
Hume holds that our ideas of moral qualities are ideas of a “relation, which nature has
placed between the form and the sentiment” (SOT, paragraph 10), best accommodates the
totality of Hume’s claims because, as far as I am aware, it is the only available account of
Hume in which he is not read as issuing either contradictory or competing claims
regarding the nature of moral qualities.
§II.7 An objection to the proffered interpretation & a reply
It might be objected that the dispositional, impression-dependency interpretation I
am espousing is incompatible with many of Hume’s claims, claims that suggest our moral
judgments are more a matter of psychological projection than genuine property
ascription. For instance, Hume writes: “Vice and Virtue, therefore, may be compar’d to
sounds, colours, heat and cold, which, according to modern philosophy, are not qualities
in objects, but perceptions in the mind” (ibid). I have already suggested that this passage
should not be read as denying that objects can rightly be attributed these qualities.
Instead, it should be read as denying that these qualities exist wholly independently of
our impressions. Hume is, in effect, claiming that they are impression-dependent
qualities.
75
Because of the necessary connection between our mental states and these
secondary qualities, our mental states determine these qualities, including moral qualities.
Indeed Hume explicitly states: “Now it has been observ’d, that our own sensations

75
Hence by heat and sound I take Hume to be talking about properties that can only be explained by appeal
to our sensations, rather than the more modern, scientific conceptions of heat and sound.

70
determine the vice and virtue of any quality, as well as those sensations, which it may
excite in others” (T 597, my emphasis). Let me now defend this suggestion.
To draw upon Hume’s own example (T 468-9), an object’s moral qualities will
entirely escape us as long as we consider the objects alone, independently of us. We will
never find them until we turn our attention to our impressions, which are obviously not
found in the object but in us. Of course we can turn to others’ impressions for assistance,
but in doing so we will not be apprehending the object’s moral properties. Hence Hume
remarks that moral qualities “are more properly felt than judg’d of” (T 470). They are
more properly felt than judged of precisely because they are feeling-dependent properties.
Consider the following passage from the Enquiry Concerning the Principles of Morals:
[C]ompare moral beauty with natural, to which in many particulars it bears
so near a resemblance… Palladio and Perrault…talk of the cornice, and
frieze, and base, and entablature, and shaft and architrave; and give the
description and position of each of these members. But should you ask the
description and position of it beauty, they would readily reply, that the
beauty is not in any of the parts or members of a pillar, but results from
the whole, when that complicated figure is presented to an intelligent
mind, susceptible to those finer sensations. Till such a spectator appear,
there is nothing but a figure of such particular dimensions and proportions:
from his sentiments alone arise its elegance and beauty. (EPM, [129-30],
emphasis added)

Compare Hume’s claim that moral properties are not qualities in objects to what I
think is the modern version of it, namely, that these so-called “secondary qualities”,
unlike “primary qualities”, are not part of the fundamental causal/explanatory structure of
the world. Under at least one type of scientific description of the world—viz., one that
makes no appeal to observers or their sensibilities or impressions—objects are not

71
attributed secondary qualities. For example, under this description, objects would not be
described as possessing color. Objects per se would be deemed colorless.
76

Hume is keenly aware that our natural “vulgar” tendency (due to various
principles of mental association) is to project an impression-independent nature onto
qualities that are impression-dependent. When Hume engages in talk that rings of
psychological projection, I suggest that his primary purpose is to warn against this
mistaken but natural tendency. Hume is, in effect, stressing that certain qualities are in
large part, though by no means entirely, the product of our sentiments. For instance,
Hume’s metaphor of “gilding or staining all natural objects with the colours borrowed
from internal sentiment” colorfully illustrates the fact that moral properties are
determined by our mental states. Without these states, there are no such properties.
External objects per se, without any appeal to observers’ sensibilities or their
corresponding mental states, would not possess moral properties. Hence moral

76
It might be objected that contemporary science construes colors, not as secondary qualities, but as
surface-reflectance properties of objects. I find this construal implausible however. For instance, we could
conceivably equip a blind person with an apparatus that accurately detects the surface-reflectance
properties of objects. Yet presumably we would not say that such a blind person apprehends the actual
color of objects; e.g., they would not directly perceive the redness of an apple. Further, in the case of color,
the surface-reflectance properties of objects must be grouped together according to the band(s) of light
reflected. However, the bands of light themselves are determined by appealing to our color impressions.
It is also noteworthy that there are objects that visually normal people operating under standard
conditions would judge to be the same color, and yet, strictly speaking, they have different reflectance
properties. (See, e.g., C. L. Hardin’s [1988] discussion of metamers.) One could insist that, in such cases,
the objects are actually quite distinct colors even though they appear to be the same color under standard
conditions. But I suggest that a more intuitive answer is that these objects are, in fact, the same color, even
though their surface-reflectance properties vary. They are the same color because color qualities are
determined by the color impressions visually normal observers experience under certain conditions, and not
simply by the surface-reflectance properties. Of course, once the requisite correlations are established
between normal observers’ visual impressions and the surface-reflectance properties of objects, we can
accurately judge the color of objects based solely upon their surface-reflectance properties.

72
properties, like all impression-dependent qualities, are, as Hume puts it, “in a manner a
new creation”.
77

Understood this way, Hume’s claims that ring of psychological projection are
compatible with the proffered dispositional, impression-dependency interpretation.
78

There are clearly instances in which Hume seems to overstate his case however. As
noted, he suggests that certain qualities such as color are “allowed to be merely a
phantasm of the senses” (On the Standards of Taste, 234). Of course, in the very same
sentence, he also boldly speaks of the “true and real color” of objects, which is
determined by “the appearance of objects in day-light, to the eye of a man in health”. At
times, Hume even speaks of such qualities being in objects—see, for instance, T 2. This
certainly seems to support the contention that Hume offers a dispositional, impression-
dependency account of some qualities. In short, the proposed dispositional, impression-
dependency account can explain why, on the one hand, Hume claims that these qualities
are a “phantasm” of the senses, to wit: they do not exist independently of our senses. It
also explains why, on the other hand, he claims that we can and do legitimately ascribe
these qualities to objects, namely: these qualities exist and genuinely belong to objects.
§II.8 A second objection to the proffered interpretation & a reply
The proffered interpretation of Hume is certainly not without its critics. For
instance, Rachel Cohon (1997) remarks: “The resulting account seems to treat moral
judgment as cognitions, specifically, beliefs (frequently counterfactual ones) about what

77
The two previous references are EPM, Appendix 1: paragraph 21.

78
David Wiggins (1987) echoes Hume’s thought when he states that, in such instances, we are “making
ourselves at home in the world.”

73
someone or anyone would feel if she occupied a [particular] point of view... This would
make moral evaluations into inductive, empirical beliefs, presumably based on past
experience of the effects of people’s character traits…” (833). She then objects that such
an account contradicts many of Hume’s claims in which he stresses that moral
evaluations are a matter of feeling and not a matter of drawing inferences (833-4).
79

Cohon offers the following citations, which she and others think are among the best
examples that speak against the sort of interpretation of Hume given by this project:
[1] “…morality…consists not in any matter of fact, which can be
discover’d by the understanding” (T 468); [2] “But can there be any
difficulty in proving, that vice and virtue are not matters of fact, whose
existence we can infer by reason?” (ibid); [3] “Morality, therefore, is more
properly felt than judged of” (T 470); and [4] “To have a sense of virtue, is
nothing but to feel a satisfaction of a particular kind from the
contemplation of a character. The very feeling constitutes our praise and
admiration... We do not infer a character to be virtuous, because it
pleases” (T 471).
80

Although all of these quotes, especially the last quote, do prima facie appear to
contradict my reading of Hume, I suggest that there is no contradiction. First, as we shall
discuss in chapters IV and V, Hume claims that our moral feelings do depend upon
particular causal inferences. So, in these passages, Hume’s main point seems to be that
moral evaluations are not merely a matter of reason but also a matter of feeling. Indeed,
this is precisely the conclusion Hume draws, namely: “that since virtue and vice are not

79
Cohon also offers a second objection, which is that “an account of the virtues which defines them as
qualities that would provoke certain sentiments in people who were in certain circumstances…is no better
able to explain moral motivation [than the moral rationalists’ account]” (1997: 835). And she points out
that one of Hume’s main objections to the moral rationalists’ account is that it cannot explain moral
motivation. In reply, first, the proffered account isn’t strictly a definition of virtue and vice or moral
qualities; second, this account is able to explain moral motivation. See chapter V for that explanation.

80
See also the so-called is-ought paragraph (T 469-70), which falls between citations (2) and (3).

74
discoverable merely by reason, or the comparison of ideas, it must be by means of some
impression or sentiment they occasion, that we are able to mark the difference betwixt
them” (T 470, emphasis added). Once again, it is noteworthy that Hume speaks of virtue
and vice occasioning, that is, causing the relevant impression. For reasons already
rehearsed, this strongly supports reading Hume as claiming that moral properties are
determined by feeling. They are feeling-dependent properties.
Second, in the four passages Cohon cites, Hume is attacking the rationalists’
account, which holds that morality is discoverable by reason alone, i.e., that “the very
essence of morality…consist[s] in an agreement or disagreement to reason” (T 460). In
the first passage Cohon cites, Hume explicitly denies this: morality consists not in any
matter of fact discoverable by the understanding alone. Contrary to what the rationalists’
claim, reason “cannot be the source” of moral distinctions (T 458, my emphasis). Rather,
feelings are the “original source” (T 460). As discussed, one must ultimately appeal to
particular feelings because the very essence of morality consists in—i.e., is determined
by—those feelings (cf. EPM 171-2). This is clearly compatible with the dispositional,
impression-dependent account of moral properties.
Similarly, in the second passage Cohon cites, Hume is rhetorically asking whether
there can be any difficulty in proving that vice and virtue are not matters of fact whose
existence we can infer merely by reason. Hume is confident that there is no difficulty in
proving this since there is no difficulty in demonstrating that moral distinctions are
response-dependent; i.e., they originate in feeling. He demonstrates this (T 468-9) by
showing that you will never apprehend the moral viciousness of an act of murder until

75
you appeal to the sentiment of disapprobation towards the act. “Here is a matter of fact,”
says Hume, “but ‘tis the object of feeling, not of reason” (T 469). It is noteworthy that
Hume explicitly states that morality is a matter of fact. Moreover, right after making this
point, Hume draws the comparison with (other) dispositional, impression-dependent
qualities and various facts regarding them, which suggests that he thought that moral
facts were very much like these non-moral facts.
So understood, Hume’s first two claims are perfectly consistent with the
impression-dependent account of moral properties. The domain of moral properties “is
not founded merely on the relations of objects, nor is perceiv’d by reason [alone]” (T
470). Instead, it is founded upon particular feelings. Consequently, since moral
properties are determined by particular feelings, the very possibility of inferring the
moral properties of objects depends upon those properties first being felt. Hence reason
alone can never discover them. In support of reading Hume this way, consider the
following passage:
[T]hose who would resolve all moral determinations into sentiment, may
endeavour to show, that it is impossible for reason ever to draw
conclusions of this nature. To virtue, say they, it belongs to being
amicable, vice odious. This forms their very nature or essence. But can
reason or argumentation distribute these different epithets to any subjects,
and pronounce beforehand, that this must produce love, and that hatred?
Or what other reason can we ever assign for these affections, but the
original fabric and formation of the human mind, which is naturally
adapted to receive them? (EPM 171-2, emphasis added)

As we’ve already discussed, it is precisely because Hume holds that moral
properties are feeling-dependent properties that he goes on to claim, in the third passage
Cohon cites, that morality is “more properly felt than judged of.” It is noteworthy that

76
Hume says that morality is “more properly” felt than judged of. This, I suggest, points to
the fact that he does not think it wholly improper, or impossible, to issue a moral
judgment without the corresponding feeling. Of course, in Hume’s view, merely
inferring the presence of a moral property based upon prior experience will not produce a
moral sentiment. And, as we shall discuss in more detail in chapters IV and V, without
actually feeling a moral sentiment, one cannot be motivated to act morally. Hence given
the fact that morality is essentially a practical matter—a fact that Hume repeatedly
stresses—moral properties are indeed more properly felt than merely judged of or
inferred.
81

However, Cohon’s objection calls into question the very idea that one can, in
Hume’s view, infer a moral fact at all. And by far the most persuasive evidence she
offers is the fourth and final passage she cites. In that passage, Hume appears to
explicitly deny that one can infer that an object has a moral property. Upon closer
inspection of the text, however, we can see that Hume is not denying the possibility of
inferring that an object has a moral property. Instead, he is only denying that when we
take up the moral view and actually feel a particular moral sentiment we then as an extra
step go on to infer that an object has a particular moral property. Hume’s point is that in

81
Chapter V discusses how, in Hume’s view, a moral sentiment can be produced upon inferring the
presence of a moral property. In such cases, the moral sentiment is, in effect, the product of our sense of
moral duty. Thus, a contingent desire to do one’s moral duty is sufficient for moral motivation. It is
noteworthy, however, that the moral sentiment is not generated by the mere inference of the presence of a
moral property. Rather, the sentiment is generated by sympathy, which works in conjunction with the
inference in such instances. (Hence, in Hume’s view, sympathy accounts for our sense of moral duty; see
chapter V.) As we shall now discuss, even in these instances there is still an important sense in which
moral properties are “more properly felt”, for they must first be felt before they can be inferred.

77
this case we sense the moral property; we do not infer it. This, I shall now argue, can be
brought to light once we include the rest of the passage that Cohon unfortunately omits:
To have the sense of virtue is nothing but to feel a satisfaction of a
particular kind from the contemplation of a character. The very feeling
constitutes our praise and admiration. We go no farther; nor do we
enquire into the cause of the satisfaction. We do not infer a character to be
virtuous, because it pleases: But in feeling that it pleases after such a
particular manner, we in effect feel that it is virtuous. (T 471, my
emphasis)

It is noteworthy that, in this passage, Hume refers to the sense of virtue, which he
holds is nothing but the capacity to feel a satisfaction of a particular kind from the
contemplation of a character from within the general moral view. Although Hume does
not make it explicit in his claim that these feelings must be had from within the general
moral view, he hints at this in the last sentence when he speaks of the “particular manner”
in which those feelings are felt. Fortunately, Hume is more explicit about his view in the
following paragraph where he writes: “’Tis only when a character is considered in
general, without reference to our particular interest, that it causes such a feeling, or
sentiment, as denominates it morally good and evil” (T 472).
The reason Hume’s reference to the sense of virtue is noteworthy is that it
highlights Hume’s focus in this passage on the actual feelings we have from within the
moral view: to sense virtue is to actually feel a feeling of satisfaction. Hume remarks that
we do not inquire into the cause of this satisfaction; that is, we inquire no further into the
cause of the feeling beyond virtue itself. Indeed, just before the passage quoted above,
Hume explicitly states that: “An action, or sentiment, or character is virtuous or vicious;
why? because its view causes a pleasure or uneasiness of a particular kind” (T 471,

78
emphasis added). So Hume has just made it clear that the virtuous character causes the
feeling of satisfaction (in those who meet the conditions of the moral view). Thus, his
point appears to be that we usually do not inquire any further into the cause. Viewing
virtue as the cause is typically good enough for practical purposes. Once again, all of this
is not only perfectly consistent with the proffered interpretation; Hume’s claim that virtue
and vice cause the relevant feelings strongly suggests that it is Hume’s account.
Consideration of the rest of the above passage also clearly indicates that Hume is
speaking of instances in which one actually feels a pleasurable sensation from within the
moral view. This is important because it serves to limit Hume’s claim that we do not
infer the moral quality of an object, such as the virtue of one’s character. Hume’s point is
that when we sense a pleasing feeling from within the moral view we do not then infer a
character to be virtuous based upon that feeling. Rather, in feeling that it pleases, we
immediately sense its virtue. In other words, that very feeling constitutes the direct
apprehension of its moral quality. Thus, we typically do not inquire further into the cause
of this feeling beyond the character’s virtue because the object of this feeling is virtue.
Comparison with our color judgments nicely illustrates Hume’s position. Our
color judgments are clearly not founded on reason but upon the actual visual impressions
we have under certain conditions. Moreover, we presumably do not infer an object to be
a particular color because of the particular color impression we are currently experiencing
under the requisite conditions. Rather, in having that particular visual impression under
those conditions, we directly apprehend the color of the object. Nevertheless, we can and
sometimes do infer an object’s color based on prior experience and the surrounding

79
circumstances. But this in no way undermines the fact that colors originate in the senses.
They must first be sensed before being inferred in cases in which they are not sensed.
So, as I read Hume, he does not deny that moral facts can be inferred. He leaves
open the possibility that we can infer moral facts in those cases in which we fail to meet
the conditions of the moral view. Indeed, as mentioned earlier, Hume explicitly states:
“We blame equally a bad action, which we read of in history, with one perform’d in our
neighborhood t’other day [even though our actual feelings toward the acts differ]: The
meaning of which is, that we know from reflexion, that the former action wou’d excite as
strong sentiments of disapprobation as the latter, were it plac’d in the same position” (T
584, emphasis added). This certainly suggests that Hume thinks that we can know moral
facts by inference from experience. And it also bolsters this project’s interpretation.
I suggest Hume holds that moral facts can be inferred, but only from other moral
facts or (what can be considered) non-moral facts that take feelings as their objects, such
as counterfactuals about how one would feel were one to satisfy certain conditions.
82

Thus, it remains the case that, pace the moral rationalists, moral facts cannot be inferred
or “deduced” solely from facts that reason alone can discover.
Hume argues this point in the now famous passage about deducing ought from is
(T 469-70). It is important to note that, when Hume takes up this issue, he still has in
mind the moral rationalists, who hold that (moral) ought is a relation discoverable by

82
See chapters IV and V for further discussion about drawing moral inferences from such non-moral facts.

80
reason alone.
83
Hume has already argued that this putative (moral) ought-relation must
be distinct from the four is-relations, namely: “Resemblance, contrariety, degrees in
quality, and proportions in quantity and number” (T 464, his emphasis). According to
Hume, holding that this (moral) ought-relation is not distinct from the four is-relations
leads to serious problems. He sums up his argument as follows:
If you assert, that vice and virtue consist in relations susceptible of
certainty and demonstration, you must confine yourself to those four
relations, which alone admit of that degree of evidence; and in that case
you run into absurdities, from which you will never be able to extricate
yourself. For as you make the very essence of morality to lie in the
relations, and as there is no one of these relations but what is applicable,
not only to an irrational, but also to an inanimate object; it follows, that
even such objects must be susceptible of merit or demerit… ’Tis
unquestionable, therefore, that morality lies not in any of these relations,
nor the sense of it in their discovery. (T 463-4)

So those who hold that (moral) ought is a relation must also hold that it is a
relation distinct from the four is-relations. But Hume objects to those who hold this view
drawing inferences from is-propositions to (moral) ought-propositions without
explaining: (i) the nature of the (moral) ought-relation putatively ascribed by ought-
propositions,
84
and (ii) “how this new relation can be a deduction from others [viz., the
four is-relations], which are entirely different from it” (T 469).

It seems “altogether
inconceivable” to Hume what a plausible answer to either question would be. He is

83
Capaldi (1989: 55-95) offers a thorough discussion of the is-ought passage in which he argues that Hume
specifically has in mind Samuel Clarke. Although my reading of the passage differs from Capaldi’s in
important respects, it has clear affinity with his interpretation.

84
This first explanatory charge is simply a reiteration of one given just a few pages earlier:
Shou’d it be asserted, that the sense of morality consists in the discovery of some
relation, distinct from these, and that our enumeration was not complete, when we
comprehended all demonstrable relations under four general heads: To this I know not
what to reply, till some one be so good as to point out to me this new relation. ’Tis
impossible to refute a system, which has never yet been explain’d. (T 464)

81
convinced that those who hold any such “vulgar” view of (moral) ought cannot give one.
Hume concludes that closer attention to the sudden transition from is to (moral) ought
would therefore “subvert all the vulgar systems of morality”, including the rationalists’
system, and let us see that moral distinctions are “not founded merely on the relations of
objects, nor…perceiv’d [purely] by reason” (T 470, emphasis added).
Careful reading of the is-ought passage reveals that contrary to one all to common
interpretation of it Hume never claims that moral ought-propositions cannot be inferred
from is-propositions.
85
The inaptly titled “Hume’s Law” is by no means Hume’s!
Instead, Hume only calls into question “vulgar” attempts to make this inference, namely:
(a) those that take moral facts—e.g., moral ought-facts—to be mere relations or (b) those
that take moral facts to consist in other “objects of reason”, in particular, primary
qualities (i.e., impression-independent qualities) that reside in objects.
Such attempts appear “altogether inconceivable” to Hume precisely because they
repudiate “the very essence of morality”, to wit: morality’s feeling-dependent nature.
86

85
A. C. MacIntyre (1959) and more recently A. Baier (1991: 176-7) also point this out, although both of
their overall interpretations of the passage differ dramatically from the one given here.

86
Thus I am in general agreement with N. Sturgeon’s claim that “Hume’s skepticism about reason in ethics
[i.e., his general skepticism about the capacity of reason to discover moral distinctions] depends…on his
reductive naturalism” (2001: 4, emphasis added). As Sturgeon notes, this is not the usual interpretation of
Hume’s views. Nonetheless, it is one that we both defend. Significantly, Sturgeon too reads Hume as
holding that moral properties are impression-dependent properties of objects. However, whereas Sturgeon
defends the common understanding of the is-ought paragraph according to which there is a “logical gap”
between is and moral ought, I do not.
This difference of interpretation is crucial because, on Sturgeon’s reading, Hume denies that moral
facts can be inferred from non-moral facts. This leads Sturgeon to conclude that Hume’s argument in the
is-ought paragraph is ultimately flawed because, given Hume’s espousal of the impression-dependent
nature of moral facts, we can know moral facts by inference from experience. As we have seen, on my
reading of the is-ought paragraph Hume does not claim that there is a logical gap between is and moral
ought. Moral facts can be inferred, but just not from facts that do not take particular feelings as their
subject matter. Hence Hume’s argument in the is-ought paragraph is not flawed, nor is it inconsistent or

82
Even though he doesn’t explicitly say so in the is-ought passage, I suggest that Hume
sees it as a virtue of his system of morality that, by taking particular feelings as its
foundation, it alone can explain how to deduce moral facts—e.g., moral ought-facts—
from particular non-moral is-facts—viz., facts about particular feelings had under certain
conditions. Consider, for instance, the following passage:
The hypothesis which we embrace is plain. It maintains that morality is
determined by sentiment. It defines virtue to be whatever mental action or
quality gives to [i.e., causes in] a spectator [who satisfies certain
conditions]
87
the pleasing sentiment of approbation; and vice the contrary.
We then proceed to examine a plain matter of fact, to wit, what actions
have this [causal] influence. We consider all the circumstances in which
these actions agree, and thence endeavour to extract some general
observations with regard to these sentiments. (EPM [127], Hume’s
emphasis)

In this passage, Hume in effect claims that an object is moral or (as the case may
be) immoral if and only if it causes in a spectator who satisfies certain conditions a
feeling of approval or (as the case may be) disapproval. Given this conception of moral
qualities as feeling-dependent properties, we discover via our feelings which objects have
such causal properties. Based on these facts, Hume says that we can then try to discover
what else, if anything, all such objects have in common. As briefly discussed, elsewhere
Hume argues that all objects that have moral properties are also either useful or
immediately agreeable to its possessor or those who associate with her. Hence our moral
feelings respond to objects that also have these properties. Significantly, it follows from

incompatible with Hume’s moral ontology. This fact alone speaks in favor of my explication of the
passage over the common understanding, or so I suggest.

87
To recapitulate, this added caveat is implied since Hume makes it quite clear that our moral feelings are
indicative of moral distinctions only when we meet certain conditions.

83
this that one can infer that an object has a moral property from facts about whether the
object is useful or immediately agreeable.
Of course the only way to establish this correlation between objects that are useful
and/or immediately agreeable and objects that are rightly judged to be moral is by
ultimately appealing to the relevant feelings that those objects cause in spectators who
satisfy certain conditions. Thus, it is possible to draw inferences about which objects
possess moral properties only after we have observed from prior experience which
objects cause the relevant feelings in those who meet the requisite conditions. Similarly,
the correlation between the surface-reflectance properties of objects and their particular
color qualities is established by appealing to the relevant visual states that those objects
cause in spectators who satisfy certain conditions. Thus, it is possible to draw inferences
about which objects possess particular color qualities only after one has observed from
prior experience which objects cause the relevant visual states in those who meet the
requisite conditions. Hence Hume’s argument that reason alone cannot discover
impression-dependent properties—in particular, moral properties—appears to be
vindicated.
In sum, all of the considerations discussed thus far strongly suggest that Hume
offers an interpersonal or intersubjective, dispositional account of moral properties,
according to which the domain of moral properties is determined by appeal to particular
“universal” sentiments had under certain conditions. Just as our color judgments refer to
intersubjective, impression-dependent color properties and not simply visual impressions,

84
so too do our moral judgments refer to intersubjective, impression-dependent moral
properties and not simply impressions of approbation or disapprobation.
Moral properties, as such, are distinct from moral impressions. A distinction
between our ideas of moral properties and our ideas of moral impressions is therefore, at
least implicitly, found in Hume’s work. And this distinction is crucial to understanding
Hume’s distinction between moral sentiment and moral judgment. Our moral judgments
are not judgments about our moral sentiments; they are judgments that ascribed moral
properties that are determined by moral sentiments. Of course, Hume acknowledges that
our sentiments of moral approval and disapproval are not always indicative of moral
qualities. Certain conditions must first be met. Consequently, our sentiments can
sometimes lead us astray, potentially resulting in belief of false moral judgments. But, as
Hume observes, it also follows from this account that we can issue true moral judgments
without actually experiencing the corresponding moral feelings. Moral facts can be
inferred.

85
Chapter III: Impressions, Ideas, and Their Relations

§III.1 Problems with Hume’s account
Thus far we have defended the claim that Hume provides a (non-minimal)
descriptive, property ascribing, truth-evaluable account of moral thought/judgment. We
have done so by appealing to Hume’s distinction between moral sentiment and judgment,
as well as the distinction between impressions and ideas. In order to fully understand the
distinction between moral sentiment and moral judgment, we need to distinguish between
our moral ideas of our moral impressions and our moral ideas of moral qualities. The
previous chapter argued that Hume offers an interpersonal or intersubjective,
dispositional analysis of the moral qualities ascribed by moral judgments, properties that
by their nature depend upon moral impressions, but are nonetheless distinct from them.
As we have seen, there is significant potential for confusion on this point. The
problem, in a nutshell, is that when Hume talks about (ideas of) morals it isn’t
immediately obvious whether he is talking about (ideas of) moral impressions or moral
properties. To illustrate this, observe that when we talk about colors it isn’t always
immediate obvious whether we are talking about the colors of objects or our color
impressions. For instance, we use the term “red” to refer to both a particular color quality
and a particular kind of visual impression. Yet the property of redness is clearly distinct
from a mere visual impression of red. An object can have the property without one
having the corresponding visual impression, and one can have the visual impression even
if the observed object does not possess the property.

86
Hume’s own account of the relation between impressions, ideas, and language,
unfortunately, only fosters further confusion about these matters. For instance, Hume
might be read as holding that a blind person cannot refer to colors, and hence cannot
judge the color of things, for the following reason. It might be thought that, on Hume’s
view, the blind cannot make color judgments—e.g., the judgment that the apple on the
table is red—because it appears that one must have had at some point an impression of a
particular color (e.g., red) in order to have an idea of that color. After all, “all our ideas
are deriv’d from correspondent impressions” (T 105). Since those blind from birth have
never had a color impression, they therefore cannot have ideas of color impressions.
Without ideas of color impressions, one cannot have ideas of colors. And without ideas
of colors—e.g., the idea of red—one cannot refer to that color. Hence blind persons
cannot refer to any colors, nor therefore can they make color judgments.
However, blind persons arguably can and sometimes do judge truly the color of
things. They, therefore, can and do make genuine color judgments, even though they
have never had an impression of any color. If so, then one’s having an experience or
impression of an object or property is not a necessary feature for possessing and using a
term that refers to (or concept about) that object or property. The public nature of
language (and concepts) makes it possible for someone to use a term (or concept) to refer
to an object or property, even though he or she may never have had any experience of, or
relating to, that object or property.
It might be thought that Hume overlooks this important fact, but he does not.
Hume thinks that the existence of words enables us to account for both genuine

87
communication and abstract ideas. So he does recognize the public nature of language.
Because language provides us with the means by which we can communicate our ideas,
including our ideas of our impressions, we can acquire ideas of others’ impressions. This
fact is significant because it enables Hume to explain how, for instance, the blind can
acquire ideas of others’ color impressions. Hence the blind can also acquire ideas of the
colors of objects even though they have never had a color impression. Thus, on Hume’s
account, the blind can indeed make color judgments. They can judge that objects are
colored, but their judgments will rely upon others’ impressions to verify those judgments.
However, this response does reveal what appears to be a serious problem with
Hume’s account of the relation between, impressions, ideas, and language. The problem
derives from the fact that, on Hume’s view, we presumably cannot acquire a simple idea
without having the requisite kind of simple impression. For instance, we cannot acquire a
simple idea of a color impression without having the requisite color impressions.
Hume does say that we can sometimes get simple ideas from complex
impressions, but the possibility of having this complex impression requires that we have
already experienced the requisite simple impressions. For example, Hume claims that we
can acquire a simple idea of a particular shade of blue of which we have not had an
impression (T 6). But in this case, which is an exception so rare that Hume hardly
thought it worth mentioning, the complex impression is constituted by one’s simple
impressions of other shades of blue; i.e., the complex impression depends upon one

88
already experiencing simple impressions of various shades of blue.
88
This is why our
ideas of shades of color are simple even when we lack the simple impression that the idea
represents. It is an idea of a missing simple impression that is formed from our simple
impressions of various shades of blue that surround the missing shade. The idea of that
simple impression, Hume concludes, must also be simple.
So it appears that, on Hume’s account, one cannot have a simple idea of a simple
impression unless one already has simple impressions of the requisite sort. This fact is
significant because it suggests that, in Hume’s view, the blind person’s ideas of others’
simple color impressions are complex rather than simple. They are ideas to the effect that
others’ have simple impressions of a sort that is wholly foreign to the blind.
On the one hand, this account has intuitive appeal. After all, those blind from
birth don’t have a clear idea of what it’s like to have a color impression. They only have
ideas to the effect that others’ have color impressions, whatever they are like. On the
other hand, the problem is that, given Hume’s account of the relation between language,
ideas, and impressions, the blind cannot make genuine color judgments, or at least they
don’t make the same color judgments as the rest of us. We, in effect, talk past each other
because, on Hume’s view, we do not appear to be referring to the same thing.
This is explained by the fact that the blind’s ideas of color impressions are not
simple ideas as are the sighted’s ideas of color impressions. Instead, they are complex

88
In the moral case, although our moral impressions do admit of degrees of intensity, we do not seem to be
able to individuate our ideas of moral impressions into distinct “shades” as we can with our ideas of color
impressions. In other words, our moral impressions don’t appear to have a robust and distinctive enough
phenomenology for this sort of individuation. It has been suggested that this is an important difference
between morals and colors. Darwall, Gibbard, Railton [1997] question, on this basis, response-dependent
characterizations of moral properties, such as the one proffered in this project. We shall address their
argument in part two of the project.

89
ideas. Hence they aren’t ideas of the same kind. Consequently, the blind’s ideas of color
qualities aren’t the same as the sighted’s ideas of color qualities. Whereas the sighted’s
ideas of colors are complex ideas that consist of simple ideas of color impressions among
other ideas, the blind’s ideas of colors do not consist of simple ideas of color impressions.
Since the sighted and the blind’s ideas of color qualities aren’t the same, neither are their
color terms.
89
The blind, as well as their color ideas and terms, lack something essential
to color, namely, a sense of the particular impressions that determine the color qualities to
which their color ideas (and ultimately their terms) refer. It is as if the blind use their
color terms in an inverted-commas sense.
90
It follows that the sighted and blind’s color
judgments differ fundamentally. The sighted and the blind, in effect, talk past each other.
This may also shed new light on Hume’s claim that moral facts cannot be inferred
by reason alone. In Hume’s view, we can only know moral facts by directly
apprehending them via particular feelings or by inference from prior experience of these
feelings. Those who have not experienced these feelings cannot grasp “the very essence
of morality”, viz., those feelings. Hume’s thought might be that one cannot infer moral
facts merely by reason because reason alone cannot apprehend the nature of morality.
The problem with Hume’s view is that the blind arguably can and do issue color
judgments just like the rest of us. Their color judgments appear to have the same truth-
conditions as our color judgments. We presumably can genuinely disagree with the blind

89
Recall that, in Hume’s view, our terms refer to our ideas.

90
Compare this with Hare’s claims that amoralist uses moral evaluative terms in an inverted-commas
sense. According to Hare, when an amoralist utters the expression “ϕ-ing is right” he actually means
something like “ϕ-ing is in accordance with what other people judge to be right” (1952: 124-6, 163-5).

90
about the color facts and the truth-value of a given color judgment/belief. Similarly, the
amoralist—viz., someone who recognizes the existence of morals yet never experiences
moral impressions
91
—can (arguably) issue moral judgments just like the rest of us.
92

Their moral judgments appear to have the same truth-conditions as our moral judgments.
And we presumably can genuinely disagree about both the moral facts and the truth-value
of a given moral judgment/belief. Hume’s conception of impressions, ideas, and
language cannot account for this however.
Because Hume bases his theory upon the construction of complex ideas from
simple ones, he subsequently overlooks the importance of analyzing any judgment as first
and foremost a propositional entity rather than a mental entity. In opposition to Hume’s
view, mental content—i.e., a belief—is analyzed in terms of its propositional content.
And the propositional nature of both thought and language makes possible direct
reference to objects and properties without the need for any corresponding experience.
For instance, depending upon the context, both the idea (or concept) red and the
term “red” refer either to precisely the same color (viz., red) or to precisely the same type
of visual impression (viz., an impression of red), regardless of whether one has ever
experienced the requisite sort of color impression or not. Similarly, the idea (or concept)
moral and the term “moral” refer to the same moral properties or the same moral
impression-types (viz., feelings of approval or disapproval), regardless of whether one

91
Compare this with David Brink’s description of the amoralist, namely: “someone who recognizes the
existence of moral considerations and remains unmoved” (1986: 30). Note that we could also describe a
blind person as someone who recognizes the existence of colors yet never experiences color impressions.

92
This is disputed by nondescriptivists (e.g., Hare [1952]) and some descriptivists (e.g., M. Smith [1994]).

91
has ever had a moral impression.
93
This proffered change enables the view to account for
the apparent fact that the amoralist and moralist’s moral judgments are synonymous.
Hence they can indeed disagree about the truth-value of any given moral judgment.
94

§III.2 The distinction between original & secondary impressions
Before concluding our discussion of impressions and ideas, and sentiment and
judgment, and move on to discuss the foundations of Hume’s moral psychology in the
following chapters, it will behoove us to acquire a better understanding of the moral
sentiments and their place within Hume’s theory of motivational psychology, including
their relation to other important impressions—viz., the passions. This will set the stage
for our discussion, in the next chapter, of sympathy and how it can produce motivation.
This section identifies and explains the kind of impression that Hume claims
moral sentiments are. In the following section, we will discuss their relation to the
passions. The passions are important in Hume’s view because they play a powerful
motivational role within our motivational psychology. In fact, they play such a powerful
role that without a close connection between the passions and the moral sentiments our
moral considerations would always be overridden by our passions. Hence a grasp of the
passions is indispensable in understanding Hume’s account of the motivational force of

93
Although the distinction between impressions and ideas is clearly important for understanding and
interpreting many of Hume’s claims, especially those found in the Treatise, it does not appear to play an
indispensable role in his moral theory. It is noteworthy that Hume does not discuss that distinction when
presenting what he considers to be the quintessential account of his moral theory in the Enquiry
Concerning the Principle of Morals. This suggests that Hume does not think that his account of how our
moral judgments refer to moral properties is essential to what he thinks is most important features of his
moral theory—namely, a commonsense account of our moral experience, one that has an empirically
verifiable ontology that is rooted in human psychology and explains the connection between our moral
considerations and moral motivation.

94
Of course, none of this changes the fact that moral properties are determined by feeling.

92
the moral sentiments. Consequently, without them, one cannot understand the connection
between moral judgment and human conduct.
Hume bifurcates impressions into original impressions (a.k.a., impressions of
sensation) and secondary impressions (a.k.a., impressions of reflection).
95
Original
impressions or impressions of sensation are impressions we get directly from our five
senses, such as colors, smells, sounds, tastes, textures, and impressions of bodily
pleasures and pains. These impressions are involved in our causal beliefs and our beliefs
about the physical world. Secondary impressions or impressions of reflection, on the
other hand, are always preceded by at least one original impression and/or by at least one
idea, which is ultimately derived from an original impression. In other words, these
impressions are special insofar as they can be caused by reflection upon other
impressions, or by ideas, or both.
96
We will discuss their causes in the following section.
This distinction between original and secondary impressions is important for the
following reasons, among others. First, besides the fact that moral apprehension is a
mode of perception that is not provided by a distinctive external sense or organ, this
distinction marks a key point at which the analogy with color breaks down: our

95
In Book I of the Treatise, Hume distinguishes between “impressions of sensation” and “impressions of
reflection”. However, in Book II he draws this same distinction but refers to it as the distinction between
“original impressions” and “secondary impressions”.

96
The fact that the secondary impressions are sometimes caused by our ideas alone is one significant
reason why Hume begins the Treatise by discussing ideas (and belief) before impressions, even though our
ideas are temporally posterior to and depend upon our impressions (T 8). Hume thought that secondary
impressions were of the utmost importance because, as we shall discuss, the passions, which are a kind of
secondary impression, are the driving motivational force within us—not to mention the fact that the moral
sentiments are also among the secondary impressions. This last point takes us to a second significant
reason, which is that if Hume can first demonstrate the inability of the rationalist’s conception of reason
and judgment to account for the intimate connection moral judgment has with motivation to action, then his
reader can better appreciate the importance of the secondary impressions in explaining this connection.

93
impressions of color are original impressions (or impressions of sensation), whereas our
moral impressions are secondary impressions (or impressions of reflection). As we shall
discuss in more detail in the next chapter, the moral sentiments are always mediated by
ideas (and hence beliefs), which is why they are among the secondary impressions. In
short, whereas we do not need to have any concepts or beliefs in order to perceive an
object’s color, on Hume’s view of morality, we do need to conceptualize the object—e.g.,
as an act that causes pain—in order to apprehend its moral qualities or moral value. This
fact clearly illustrates an important point at which the analogy with color breaks down.
97

This fact is important for various reasons. For instance, it helps explain why
reason plays such an important role in making moral distinctions—cf., the absence of the
role of reason in making color distinctions. Moreover, it also helps explain why there is
so much more disagreement concerning our judgments of moral distinctions and moral
facts than our judgments of color distinctions and color facts: a significant degree of
moral disagreement can be explained by appealing to disagreement about non-moral, but
morally relevant, beliefs. Hence, if Hume’s approach is correct, then simply appealing to
the extent of moral disagreement—e.g., as compared to color disagreement—may not be
a very persuasive way of objecting to his moral theory. We shall discuss these issues in
more detail later in part two of the project.

97
Some (e.g., Blackburn [1993]) have suggested that these differences undermine any strict comparison
between moral properties and secondary-qualities. I disagree. The distinguishing characteristic of all
secondary-qualities is that they are response-dependent properties. Neither the fact that the impressions
that determine moral qualities are secondary impressions rather than original impressions, nor the fact that
the moral “sense” does not have its own distinct external mode of perception, subtract from the response-
dependent nature of moral properties. Hence what I have suggested is the main point of Hume’s
comparison with other secondary-qualities such as color stands. For further discussion of this issue, see the
following chapter.

94
Second, understanding both how and why our moral impressions get classified as
secondary impressions serves, in effect, to mark the key role that the human capacity for
sympathy plays in Hume’s moral psychology, namely, the role of converting the ideas of
others’ pleasure and pain into impressions of pleasure and pain within oneself. This
conversion process of sympathy is crucial for at least the three reasons given below. We
will discuss these reasons in more depth later.
First, on Hume’s account, reason alone can never motivate; pleasurable or painful
impressions, or at least their prospect, are necessary.
98
Hence the prospect of the
production of pleasurable or painful impressions within oneself is crucial if others’
pleasure or pain is to have any chance of motivating us to action. Hence sympathy’s
conversion process is vital for an account of moral motivation.
99

Second, for this very reason it is indispensable in grasping the connection
between moral judgment and motivation because it is key to understanding how moral
feelings, which determine moral properties and induce moral motivation, are produced.
100

98
Hume writes:
‘Tis from the prospect of pleasure or pain that the aversion or propensity arises toward
any object: And these emotions extend themselves to the causes and effects of that object,
as they are pointed out to us by reason and experience. It can never in the least concern
us to know, that such objects are causes, and such other effects, if both the causes and
effects be indifferent to us. Where the objects themselves do not affect us, their
connexion can never give them an influence; and ‘tis plain, that as reason is nothing but
the discovery of this connexion, it cannot be by its means that the objects are able to
affect us. (T 414)

99
As revealed in the preceding footnote, Hume does think that our ideas/beliefs about the prospects
of pleasure and pain can generate motivating states within us. According to Hume, sympathy is the
process by which our ideas/beliefs about the prospects of others’ pleasure and pain can generate
motivating states in us.

100
According to Hume, there are no “original” or “natural” moral motives. Thus, in Hume’s view, there
must be some original, empirically verifiable, non-moral motive that accounts for the moral sentiments and

95
Without the conversion process of sympathy, neither the moral impressions nor other
important emotional states—viz., the passions, which on Hume’s view ultimately
motivate us to act—would be produced. Consequently, without the capacity for
sympathy, there can be no moral motivation.
Third, as we shall discuss in part two (chapter X) of this project, the (all but)
universal human capacity for sympathy plays a significant role in accounting for
morality’s normativity, including a defense of the claim that it applies to individuals
independently of their attitudes, intentions, beliefs, etc. In short, by fully understanding
the precise manner in which moral impressions are secondary impressions, one can
thereby come to grasp the importance of sympathy within Hume’s moral psychology.
The distinction between original and secondary impressions is also important
because the passions, which in Hume’s view play a central role within our motivational
psychology, fall under the secondary impressions. According to Hume’s theory of
human psychology, there is a direct causal connection between the passions and the
moral sentiments. Among other things, this connection explains why we often
experience an overwhelming motivation to act morally. In other words, in Hume’s view,
the passions play a crucial part in explaining why our moral considerations frequently
appear to have such a strong motivational force. Equally important, they also play a part
in explaining why we sometimes do not feel a compelling moral motivation: sometimes
our passions conflict with and prevail over our moral feelings, and in some cases they
actually prevent us from taking up the moral perspective.

moral motivation. As we shall discuss in chapter IV, sympathy is the original, empirically verifiable, non-
moral motive that fulfills this function.

96
In short, if moral considerations are to play a significant role within our
motivational psychology, then the passions must figure within Hume’s theory of moral
psychology. Otherwise our passions would always override our moral considerations.
So it is to the passions and their causal relation to the moral sentiments that we now turn.
§III.3 Moral impressions & the passions
In order to understand the causal relation between the moral sentiments and the
passions, we first need to know how the passions themselves are caused. Hume
subdivides the secondary impressions into two classes according to their phenomenal
“feel”: the calm and the violent. The calm impressions include the aesthetic
impressions—i.e., impressions of beauty and deformity—as well as the moral
impressions—i.e., impressions of moral approval and disapproval.
The violent impressions, or “passions”, are further bifurcated into direct and
indirect passions.
101
This distinction pertains to the manner in which the passions are
caused. Direct passions are caused immediately by reflecting upon original impressions
of pain or pleasure, or, in some cases, simply from “instinct” or predisposition.
102
These
passions come in contrasting pairs such as desire and aversion, joy and grief, and hope

101
As Capaldi (1989) observes, moral impressions are not passions. This point has been misunderstood in
the literature by, for instance, Norman Kemp Smith (1941), Pall Ardal (1966), and more recently, Christine
Korsgaard (1999) and Elizabeth Radcliffe (2000).

102
Hume writes:
Besides [“natural” or non-moral] good and evil, or in other words, pain and pleasure, the direct
passions frequently arise from a natural impulse or instinct, which is perfectly unaccountable.
Of this kind is the desire of punishment to our enemies, and of happiness to our friends; hunger,
lust, and a few other bodily appetites. These passions, properly speaking, produce good and evil,
and proceed not from them, like the other affections (T 439).
One highly plausible way to account for these “unaccountable” natural predispositions is to appeal to
natural selection.

97
and fear. Indirect passions are also caused by reflection upon original impressions of
pain or pleasure, but as with the moral impressions, they too must be mediated through
reflection upon ideas. In other words, reflection on both impressions and ideas are
necessary to cause indirect passions. The indirect passions also come in contrasting pairs
such as pride and humility, love and hate, and pity and malice.
Let us now take a closer look at the causal mechanisms involved. As mentioned
in the previous section, Hume holds that there are three general principles of
psychological association by which secondary impressions are caused. First, we have
already noted that some secondary impressions—viz., our moral impressions—are caused
by reflection on ideas.
103
According to the principle of association of ideas, the human
mind tends to naturally associate ideas based upon their resemblance, contiguity, and
cause and effect. Second, some secondary impressions are caused by reflection on
impressions according to their resemblance. Hume calls this causal principle the
principle of association of impressions. According to Hume, all secondary impressions
naturally cause other secondary impressions according to their resemblance. Hume
articulates this point in the following passage:
Grief and disappointment give rise to anger, anger to envy, envy to malice,
and malice to grief again, till the whole circle be compleated [sic]. In like
manner our temper, when elevated with joy, naturally throws itself into
love, generosity, pity, courage, pride and the other resembling affections
(T 283).

103
How reflection upon our ideas can cause impressions will be discussed in the following section. But we
have already seen that sympathy is a psychological process that converts the idea of another’s pain or
pleasure into a like impression within one’s self. This transition occurs because of the transference of
vivacity from an initial impression to an idea (and sometimes to another idea), and then to another
impression, and so on.

98
The indirect passions of love and hate, as well as pride and humility (and other
related feelings such as guilt, shame, etc.) closely resemble the moral sentiments.
Consequently, these passions are naturally causally associated with the moral sentiments
(and vice-versa). The real significance of this aspect of Hume’s account of human
psychology is that our moral motivations are of a piece with our overall motivational
psychology. Consequently, our non-moral feelings can affect our moral feelings and
vice-versa.
Finally, the third principle that can account for the cause of a secondary
impression is what Hume calls the “double relation of impressions and ideas”. This
principle is, in effect, the occurrence of the two former principles conjoining and
reinforcing each other. Hume uses this last principle of association to explain the causes
of all the indirect passions. Moreover, according to Hume, this principle accounts for the
strong motivational force of the indirect passions. And because of the association
between these passions and the moral sentiments, it also accounts for the strong
motivational force we typically associate with moral considerations.
To illustrate this latter principle of double relation and thereby how indirect
passions are caused, let’s consider one of Hume’s examples. Suppose I take pride in “the
beautiful house” (T 279). According to Hume, pride is caused when one reflects on a
pleasing quality (e.g., beauty) of a “subject” (e.g., a house), which is also somehow
related to the self (e.g., it might be related to me by ownership and/or causality supposing
I built it or designed it). Hence I can take pride in the beautiful house only if the house
(a) has a quality that is pleasing to me and (b) is somehow related to myself. Together,

99
these conditions are both necessary and sufficient for the production of pride, which is a
separate feeling from the pleasure produced by the quality itself.
For instance, if the subject does not cause a pleasing feeling in me but instead an
unpleasant feeling, then the subject’s quality cannot cause pride in me. Hume claims that
a subject with such a quality would cause the indirect passion of humility. Moreover, if
the subject does cause a pleasant quality in me but the subject is related to someone else
rather than myself, then I will not experience pride. Rather, according to Hume, the
subject and its pleasing quality would cause me to feel the indirect passion of love for that
individual. On the other hand, if the subject’s quality caused an unpleasant sensation and
was related to another self, then it would cause me to feel hatred towards that
individual.
104

In short, all indirect passions are caused by an idea of a subject with two parts:
namely, a subject that (1) possesses a certain quality—a quality that is itself either
pleasurable or painful—and (2) is related to either the self or some other self. For
instance, as discussed, pride is only produced by reflection on the subject and its pleasant
quality—e.g., the beautiful house—and its relation to oneself—e.g., I own, built, and/or
designed the house that brings me pleasure. Equally important is the fact that the subject
and its quality—e.g., observing or contemplating the beautiful house—causes an
impression of pleasure—as in the present example—or pain irrespective of any further
considerations (e.g., whether the subject is related to me, someone else, or no one at all).

104
Among other reasons, the role that the conceptions of self and other play in Hume’s account of the
passions is important because it enables Hume to distinguish between love of non-human objects and
persons.

100
In other words, the subject’s pleasant or unpleasant quality is wholly independent of the
subject’s relation to the self or some other self. The beautiful house, for example, would
cause an impression of pleasure even if no one owned it.
Given Hume’s account of the cause of the indirect passions, it turns out that the
same qualities would cause these different indirect passions simply in virtue of the
relation of the subject and its quality to the self or someone else: “the same qualities that
produce pride or humility, cause love or hatred” (T 332). As Hume points out (T331-2),
this is why, for instance, we tend to publicly exhibit those qualities that cause us to feel
pride and hide those qualities that cause us to feel humility. On the other hand, it is also
why we often feel pride in those qualities that others love in us, and why we feel humility
about possessing those qualities for which others have feelings of hatred towards us.
105

Although this may initially appear rather commonsensical, we should not
underestimate Hume’s contention. Hume states that “the pleasant or painful object [must]
be very discernible and obvious, and that not only to ourselves but to others also” (T
292). Hence, on Hume’s view, the same quality in us that causes us to feel pride must,
when possessed by others, cause us to feel love for them; that is, it must do so when we
observe or contemplate the quality in a steady and general manner.
106
If it does not, then

105
It is our capacity for sympathy that accounts for this. Through sympathy, we feel the pleasure caused in
others by an object’s quality. Hence if that object is related to ourselves, or if the quality is a quality of the
self, then we will feel pride. Similarly, if we take pleasure in a particular quality of an object (i.e., subject)
that is related to ourselves, others are likely to sympathetically feel our pleasure and this will cause them to
feel love for us.

106
James King (1999) calls this principle of parity the Principle of Self-other Parity, and uses it to
demonstrate that Hume’s sensible knave cannot feel pride for her knavery.

101
whatever feeling one may be experiencing, that feeling does not, in Hume’s view, count
as a genuine instance of the relevant indirect passion.
107

The reason why it does not qualify as an indirect passion is that Hume builds into
his account of the indirect passions, as well as the painful and pleasurable states that
cause them, an interpersonal (i.e. shared) perspective that contains various corrective
principles.
108
In other words, according to Hume, in order for a quality to be deemed
genuinely pleasant or unpleasant, others must have the same basic reaction to that quality
(with various corrective principles in place) as oneself.
109
Such corrective principles also
apply to the indirect passions themselves, since the subject and its quality are the cause of
these passions. If the subject and its quality are not genuinely pleasant or unpleasant,
then whatever feelings one might feel in response to what one mistakenly believes to be a
pleasant or unpleasant quality of a subject, it isn’t a genuine instance of an indirect
passion; nor is the feeling fitting since it rests upon a false belief.
110

No doubt this will strike many readers as a rather peculiar and perhaps
implausible account, as it appears incompatible with our ordinary usage of the terms

107
When it does not and the feeling is pleasant, King (1999) calls this a feeling of relish. King presumably
borrows the term straight from Hume (EPM, section I) who speaks of a “false relish” that must be corrected
by “argument and reflection” in order to qualify as a “proper sentiment”).

108
King (1999) also notes this. In fact, Hume applies corrective principles to all cases in which we judge
things by our senses (T 603, Enquiry 227-8); see chapter IV for further discussion. Sayre-McCord (1994)
also stresses this point.

109
However, as King observes, Hume does recognize that our responses to those qualities may come in
varying degrees. (See, for example, T 392).

110
Although I think that Hume would agree that the feeling one feels in such instances is unfitting because
it rests upon a false belief, he would not describe the feeling as an unfitting indirect passion. For example,
Hume cannot say that, in such an instance, one feels an unfitting feeling of pride because, according to his
account of the cause of the indirect passions, that feeling cannot be a genuine instance of the indirect
passion.

102
‘pride’, ‘love’, ‘humility’, etc. Typically we think that a person can feel a genuine (and
perhaps fitting) feeling of pride in an object that is related to the self, even if the object
does not cause others to feel pleasure. For example, others may simply not find my
house to be beautiful, as I do. Indeed most of us presumably think that a pleasant or
unpleasant quality such as beauty or deformity is a rather subjective or personally relative
quality. Beauty is, after all, in the eye of the beholder. Hence it is a personally relative
or idiosyncratic issue whether or not one feels an indirect passion such as pride.
Nevertheless, Hume claims that there is a non-idiosyncratic fact of the matter
concerning whether a subject really has a pleasant or unpleasant quality (e.g., beauty).
Certain standards apply, and these standards act as a corrective measure such that some
people’s opinions and preferences are objectively better than others. In other words,
according to Hume, there is as much a fact of the matter concerning whether or not my
house is indeed beautiful as there is concerning whether or not my house is a particular
color or shape. So there are facts about whether the feelings one feels are indeed genuine
and fitting instances of the indirect passions.
In both cases, we can be mistaken. For instance, we can believe that the feeling
we feel is an instance of pride when in fact it is not because we can mistakenly believe
that a subject and its quality are genuinely pleasurable when in fact they are not.
Moreover, it turns out that the feeling we feel in such instances is also unfitting insofar as
it is a response to a false belief. This aspect of Hume’s theory is perspicuously brought
out in his essay “Of the Standard of Taste”. For instance, Hume writes:

103
Whether any particular person be endowed with good sense and a delicate
imagination, free from prejudice, may often be the subject of dispute, and
be liable to great discussion and enquiry: But that such a character is
valuable and estimable will be agreed in by all mankind. Where these
doubts occur, men can do no more than in other disputable questions,
which are submitted to their understanding: they must produce the best
arguments that their invention suggests to them; they must acknowledge a
true and decisive standard to exist somewhere, to wit, real existence and
matter of fact; and they must have indulgence to such as differ from them
in their appeals to this standard. It is sufficient for our present purposes if
we have proved that the taste of all individuals is not upon an equal
footing, and that some men in general, however difficult to be pitched
upon, will be acknowledged by universal sentiment to have a preference
above others. (‘Of the Standard of Taste,’ paragraph 25)

This is a point at which many readers of Hume will likely demur and arguably for
good reason.
111
For instance, one would presumably be hard pressed to demonstrate the
requisite sort of objective standard and corresponding facts about the beauty of objects.
112

If so, then are we to conclude that an object’s (e.g., my house’s) putative beauty cannot

111
However, we should be careful not to understate the power that public opinion and agreement does tend
to have over our own feelings and preferences. Largely because of our capacity for sympathy, we are
sensitive to all sorts of social pressures. Hence, even if Hume does overstate its role, I suggest that there is
at least something to Hume’s privileging the common or interpersonal point of view. For instance, upon
discovering that virtually no one else finds one’s house to be beautiful, one may very well cease to take
pride in it. Or perhaps upon discovering that others take pleasure in some quality in oneself, one may then
feel pride for possessing that quality, which previously gave oneself no pleasure whatsoever. On the other
hand, one may not. But, in Hume’s view, whether or not these feelings, or lack thereof, are justified
depends entirely upon whether or not anyone else would (mutatis mutandis) share those feelings—hence
Hume’s (perhaps implausibly) strict adherence to the common, objective point of view. (Regarding the last
statement, see also Gabriele Taylor [1980: 401].)

112
This is not to say that just anything will count as an instance of beauty however. As Hume stresses, we
all share the same physiological underpinnings (or “original existences”), which is, in effect, why we quite
naturally expect there to be at least general agreement in our responses. There is, for example, statistically
significant cross-cultural agreement in the ranking of human faces according to their attractiveness
(although many of the faces ranked exhibit various deformities). The relevant standard of attractiveness in
this case appears to be a rather precise symmetrical relation among facial features. The ability to recognize
this particular symmetrical relation has a neurological basis—in fact, there is a specialized area of the
frontal cortex devoted purely to human facial interpretation and recognition. This standard of attractiveness
is most likely the result of the process of natural selection based upon correlations between genetic viability
and facial symmetry. Regardless, this clearly will not produce the sort of standards for beauty that Hume’s
position requires. (Part of the problem is that, in this example, the faces are ranked on a scale of their
comparative attractiveness rather than in terms of beauty per se.)

104
therefore be the cause of a genuine instance of pride? Hume’s view necessitates this
conclusion, and this is presumably an outcome that, for most of us, runs contrary to our
intuitions on the matter.
In any case, whatever worries or reservations one may have regarding this aspect
of Hume’s account of the indirect passions, I argue that they do not apply to the concerns
of this project. The passions only interest us insofar as they are causally related to the
moral sentiments, and significantly, this project claims to provide evidence for holding
that an interpersonal or intersubjective perspective, which contains particular corrective
principles, does govern our moral sentiments. In this respect, our moral judgments are
more like color judgments than aesthetic judgments. Hence, if this project can discharge
this promissory note, then the indirect passions, at least in as much as they relate to the
moral sentiments,
113
will have the (perhaps otherwise troublesome) features Hume
claims.
Let us therefore return back to the main thread of the discussion, namely: the
cause of the indirect passions, their motivational strength, and their relation to the moral
sentiments. As we have discussed, the subject’s quality independently causes a
pleasurable or painful impression. The subject and its quality together, when related to
the self or another self, also cause the indirect passions, impressions that are likewise
either pleasurable or painful. In other words, we have a double relation of pleasant or
unpleasant impressions, which is based on the principle of association of impressions.

113
As briefly discussed two footnotes after this one and then again more thoroughly in part two, this does
mark an important difference from Hume’s approach.

105
For example, in the case of the beautiful house, the house has a quality—viz., its
beauty—that causes a pleasurable impression independently of any other consideration.
And the association of the subject and its quality with myself—e.g., my reflecting on the
fact that I own the beautiful house—also causes the pleasurable passion of pride in me.
Supposing, however, that the house is in disrepair, it would independently cause an
unpleasant feeling. And upon reflection of the association of that displeasing house with
myself, I would subsequently experience the displeasing feeling of humility.
Furthermore, we also have a double relation of ideas, which is based on the
principle of association of ideas. Not only is the cause of an indirect passion—viz., the
idea of the subject and its quality—associated with an idea of the self or someone else,
the indirect passion itself, which is the effect of that cause, is also likewise associated
with the same idea. For instance, the passion of pride is always associated with the self.
Thus, Hume nicely sums up the occasion of pride when he writes: “Any thing, that gives
a pleasant sensation, and is related to the self, excites the passion of pride, which is also
agreeable, and has self for its object” (T 288). Hume calls the self, in its association with
pride, the “object” of pride. The object of an indirect passion—viz., the self or some
other self—is always causally related by association back to the cause of that passion:
namely, the idea of the subject—i.e., an item and its quality—such as the beautiful house.
Thus, according to Hume, the cause of the passion is its subject. This is why, for
instance, I take pride in the beautiful house (viz., the subject) that I (the object) own.
The passion of humility also has the self for its object, which is why Hume groups
pride and humility together. But these two passions are a contrasting pair because,

106
whereas pride is agreeable, humility is a disagreeable feeling. Similarly, Hume (T 329ff)
classifies love and hatred together as a contrasting pair of passions based on the fact that
they share the same object: they each have some other self for its object. But whereas
love is a pleasurable feeling, the passion of hatred is painful. Hence these passions are a
contrasting pair.
Moreover, just as one would expect, the cause, or subject, of love—e.g., the
beautiful house—is associated with (i) a pleasant sensation, and (ii) an idea of some other
self—e.g., my neighbor. And the cause (i.e., the idea of the subject) of hatred—e.g., the
thought of the dilapidated house—is associated with (i) an unpleasant sensation, and (ii)
an idea of some other self—e.g., another neighbor of mine.
114
So, in the former case, the
pleasure caused by the beautiful house is carried over to those related to it such as its
owner (e.g. my neighbor) or its architect. Similarly, in the latter case, the displeasure
caused by the thought of the dilapidated house is carried over to those related to it.
In sum, the cause of any indirect passion is a complex idea that is related to both
an idea of the self or another self and an impression of pleasure or pain that exists
independently of that relation and independently of the passion. The effect of the cause,
namely the indirect passion itself, is separately but similarly either pleasurable or painful
(depending on whether the aforementioned impression is pleasurable or painful), and it is
also related to the same idea as its cause: namely, either the self or another self. Hence,
due to the active principles of mental association of impressions and ideas, the cause is
doubly related to the effect. Together, the double impulse—i.e., experiencing twice over

114
As with “pride” and “humility”, Hume doesn’t appear to use the terms “love” and “hatred” in the way
that they are commonly used.

107
a pleasurable or painful feeling with respect to the same item object—and double
relation—i.e., experiencing twice over, through mental association, the relation of that
object to the self or some other self—not only account for the motivational strength, or
“vivacity”, of the indirect passions; in Hume’s view, they also thereby account for the
overall motivational force of all pleasurable or painful feelings associated with any object
that is related to the self or some other self.
Significantly, this includes the pleasurable and painful feelings associated with
acts and traits that are related to the self or some other self. And as we’ve discussed,
Hume conceives of moral properties as qualities of objects, such as acts and traits, that
are both “placed in”, or related to, the self or someone else and cause feelings of pleasure
or displeasure—viz., the moral sentiments—in anyone who satisfies certain conditions.
Thus, objects that possess moral qualities—e.g., virtues and vices—also cause the
passions of pride or love, or humility or hate respectively.
115
As Hume puts it:
Pride and humility, love and hatred are excited, when there is any thing
presented to us, that both bears a relation to the object of the passion [viz.,
others or oneself], and produces a separate sensation related to the
sensation of the passion. Now virtue and vice are attended with these
circumstances. They must necessarily be plac’d either in ourselves or
others, and excite either pleasure or uneasiness; and therefore must give
rise to one of these four passions. (T 473)

Hume conceives of virtues and vices as moral and immoral character traits
respectively. Hence, as traits, they must be “placed either in ourselves or others”; i.e.,
they are traits of persons. And in virtue of possessing moral qualities, they must “excite
either pleasure or uneasiness” respectively. In other words, these traits will cause

115
Hume observes that we can use this feature to distinguish our moral feelings from “the pleasure and pain
arising from inanimate objects” (ibid).

108
feelings of moral approval or (as the case may be) disapproval in anyone who satisfies
certain conditions. More precisely, Hume’s view is that virtues and vices are dispositions
of persons to behave in particular ways, behavior that tends to produce pleasurable or
painful effects. Consequently, when these dispositions are exercised, they cause, via
sympathy, feelings of moral approbation or (as the case may be) disapprobation in
anyone who satisfies certain conditions while observing or contemplating the typical
effects of those dispositions. And this is what makes them moral dispositions/traits. Of
course this reading strongly supports the proffered dispositional, impression-dependency
interpretation of Hume’s moral ontology, and the claim that our ideas of the moral
qualities virtue and vice possess must be distinct from our ideas of the moral impressions
that virtue and vice cause.
In any case, the fact that virtue and vice always cause the passions of pride or
love, or humility or hate respectively is, says Hume, “the most considerable effect that
virtue and vice have upon the human mind” (T 473). This is because these passions play
a crucial role in accounting for the motivational force of objects that possess moral
qualities; i.e., moral properties acquire much of their motivational force by generating
other feelings, namely, passions.
Thus, not only do the double impulse and double relation associated with the
passions account for the powerful motivational force of the moral impressions, as well as
all other pleasurable or painful feelings associated with an object related to the self or
some other self; it also accounts for the overall motivational force of moral properties. In
other words, both our moral impressions and moral properties acquire their strong

109
motivational impetus from their direct causal association with the indirect passions, a
causal association that, as just discussed, occurs according to the active mental principle
of “the double relation of impressions and ideas to pain and pleasure” (T 574). Hence
Hume writes:
We have already observ’d, that moral distinctions depend entirely on
certain particular sentiments of pain and pleasure, and that whatever
mental quality in ourselves or others gives us a satisfaction, by the survey
or reflexion, is of course virtuous; as everything of this nature, that gives
uneasiness, is vicious. Now since every quality in ourselves or others,
which gives pleasure, always causes pride or love; as every one, that
produces uneasiness, excites humility or hatred: It follows, that these two
particulars are to be consider’d as equivalent, with regard to our mental
qualities, virtue and the power of producing love or pride, vice and the
power of producing humility and hatred. In every case, therefore, we must
judge of the one by the other; and may pronounce any quality of the mind
virtuous, which causes love or pride; and any one vicious, which causes
hatred or humility.
116
(T 575)

According to Hume, the indirect passions can transfer their vivacity to other
secondary impressions that resemble them. For instance, recall the following quotation
from Hume, which illustrates the affect of the principle of association of impressions:

116
See also T 473. It is noteworthy that the conclusion Hume draws in the above passage is the result of
the fact that, as we have already briefly discussed, Hume thinks that there are interpersonally objective
standards that apply not only with respect to our moral judgments but to all judgments about qualities that
are “discovered” by our senses, including judgments about the size, shape, color, and beauty of objects. I
have called the plausibility of Hume’s approach into question, at least with respect to some qualities such
as the beauty of my house. However, I contend that there is an interpersonally, or intersubjectively,
objective moral standard. I shall defend this contention in part two.
The result of this difference in our approaches is significant. According to the view to be
proposed, not all mental qualities or character traits that give us satisfaction or uneasiness will turn out to
be morally virtuous or morally vicious respectively. Instead, only those objects that are evaluated in
accordance with the moral standard, or perspective, will be correctly evaluated as moral or immoral. In
other words, this project does not simply lump all virtues together, as Hume arguably does, because it, in
contrast to Hume’s approach, has a way to clearly distinguish moral approbation from plain old non-moral
admiration. This, consequently, enables the proposed view to distinguish among different kinds of virtues
and vices, that is, between moral and non-moral virtues and vices. For instance, some qualities or traits
that engender pleasure within oneself and/or others (e.g., wit) may turn out to be prudentially and/or
socially virtuous but not morally virtuous. We shall discuss these issues in much greater detail in part two.

110
“Grief and disappointment give rise to anger, anger to envy, envy to malice, and malice
to grief again, till the whole circle be compleated [sic]. In like manner our temper, when
elevated with joy, naturally throws itself into love, generosity, pity, courage, pride and
the other resembling affections” (T 283). Hence, because our moral impressions
“resemble” the passions, moral impressions can acquire the indirect passions’ intense
motivational vivacity. Moral properties, insofar as they cause those moral impressions,
thereby have a strong motivational force built into them as it were.
We can now better appreciate the significance of this association. Through the
double impulse and double relation, the indirect passions garner significant motivational
vivacity. This vivacity can then be transferred to other impressions according to their
resemblance to those indirect passions. The end result is that the principles of mental
association not only account for the production of the indirect passions as well as their
motivational strength; they also account for their motivational influence on other mental
states that motivate in the same general direction. And this subsequently explains the
powerful motivational role the indirect passions play not only within our overall
motivational psychology but also within our moral psychology. And this, in turns,
explains how moral properties acquire a strong motivational impetus, and why Hume
thinks that engendering the passions of love, hatred, pride, and humility is “the most
considerable effect that virtue and vice have upon the human mind” (T 473).
§III.4 Some concluding remarks
It is precisely because the indirect passions play such a powerful motivational role
in Hume’s account that we have taken time to discuss them. If moral considerations are

111
to play a significant role within our motivational psychology, then it seems the passions,
particularly the indirect passions, must figure within Hume’s theory of moral
psychology—otherwise moral considerations would always be overridden by our direct
and indirect passions. As we have seen, this is indeed the case. The moral sentiments are
closely associated with the indirect passions such that the indirect passions are produced
by these impressions (and vice-versa) according to the principle of the double association
of impressions and ideas.
There is, to be sure, much more we could say about Hume’s theory of
psychology, including his theory of the passions. Unfortunately, many of those details
would take us outside the purview of the present project.
117
Hume’s theory of
psychology and the passions concern us only insofar as they are requisite for
understanding how Hume accounts for the motivational force of our moral
considerations. And this is crucial to appreciating the similarities, as well as the
differences, between Hume’s moral theory and the view to be proposed in part two of this
project.
It is an interesting question whether or not Hume is right that the passions,
especially the indirect passions, play an indispensable role in accounting for the
motivational impetus of moral considerations. However, this too is beyond the pale. It is
at heart an empirical question whether the indirect passions are the driving force behind

117
Further discussion of Hume’s theory of the passions can be found in Smith (1941), Albert Glathe
(1950), Ardal (1966), which arguably has become the locus classicus on Hume’s theory of the passions,
and Capaldi (1989: 153-93), which offers an excellent discussion, including a brief overview of the
literature.

112
moral motivation, an empirical question ultimately to be settled by cognitive science.
118

The present project will consequently remain non-committal about whether or not Hume
is right about this. In other words, as far as the present project is concerned, the moral
sentiments of approval and disapproval, in conjunction with sympathy, may turn out to be
sufficient on their own to account for the strong motivational force of our moral
evaluations. Having said that, there is, I think, something to be said in favor of Hume’s
introspective account: those emotions, or “passions”, do after all seem to figure
prominently in our moral experience.
119

It is noteworthy that Hume not only attempts to explain morality’s motivational
component by appealing to the passions; as we shall discuss later, he also attempts to
explain its normative component in a similar manner. Although appeal to the passions

118
Here I dramatically part ways with Hume who, in Part II of the Treatise, is willing to play the cognitive
scientist insofar as he commits himself to a particular substantive analysis of the motivational force of the
direct and indirect passions based on introspective “experiments”. But it is noteworthy that Hume did not
confuse his introspective experiments with the physiological processes that underlie them. Moreover, as
we shall discuss in part two, it is to Hume’s credit that his introspective experiments concerning the
sympathy process offer an astonishingly accurate account of our current neurological understanding of that
psychological process. Hume hoped for a physiological confirmation of his introspective experiments, and
he now appears to have it, at least with respect to his general account of sympathy.

119
Philip Mercer (1972: 49ff) claims that “it is easy to pick holes in Hume’s position here”. “For instance,”
Mercer continues, “although a fairly plausible case can be made out for a connection between, on the one
hand, pride and humility (in so far as the latter is equivalent to shame and remorse) and, on the other, moral
disapproval, it is clear that we do not necessarily love where we approve or hate where we disapprove. The
awkwardness of Hume’s position is due not to faulty observation but to the rigidity of his psychology of
association.” Although I too have just questioned Hume’s theory of human psychology, I do not think
Mercer’s critique is correct. Hume’s psychology of association is not nearly as rigid as Mercer’s critique
suggests. It is quite clear, for example, that our feelings of approval and disapproval do not causally
necessitate feelings of love and hatred respectively. Rather, Hume claims that it is only when those
feelings are sufficiently strong do they cause such feelings as love and hatred. Hume observes that the
force of our feelings can vary depending on all sorts of factors, and as a result, their causal efficacy can also
vary. In this way, Hume combats the sort of rigid psychology Mercer suggests is latent in his view. As we
shall discuss in the next two chapters, Hume’s empirical observation regarding the variations in the vivacity
of our feelings will have important consequences for his account of sympathy and its relation to the moral
sentiments.

113
may be able to account for certain aspects of moral normativity, that appeal cannot fully
explain the commonplace view that moral evaluations have a kind of attitude-
independent normative authority. Nonetheless, I do think that the commonplace view of
moral normativity can be accommodated in a manner generally consistent with Hume’s
moral ontology and his account of normativity, or so I shall argue in part two.
In any case, this project does not attempt to defend all of the particulars of
Hume’s theory of psychology, nor the particulars of his theory of the passions, nor
therefore the entirety of his theory of moral psychology. Its aims are much more modest.
It only attempts to defend two key aspects of Hume’s moral theory.
First, it defends what I have argued is Hume’s moral ontology, namely, the view
that moral properties are dispositional, impression-dependent properties of objects. To
recapitulate, as I read Hume, he holds that an object has a moral property if and only if it
is such as to elicit particular feelings in anyone who satisfies certain conditions. Second,
this project defends the basic foundations of Hume’s moral ontology, to wit: the notion of
a “common” or universal moral view, which is comprised of various corrective principles
that regulate the capacities upon which the moral view depends, namely, sympathy and
causal reasoning. It is to these foundations—viz., sympathy, causal reasoning, and the
moral view—that we now turn.

114
Chapter IV: The Foundations of Hume’s Moral Ontology

§IV.1 The general principle of morals
Recall that our main goal in part one of this project is to explain Hume’s account
of the nature of moral properties and facts. By employing the distinction between moral
sentiment and moral judgment, chapter II, has defended this project’s contention that
Hume provides an account of moral judgment as descriptive, property-ascribing, truth-
evaluable judgments, some of which are true. According to that account, both error and
moral disagreement, including disagreement about the truth-value of a given moral
judgment, can and do occur. This marks an important first step toward our goal of
resolving the accommodation problem.
Chapter II has also argued that Hume offers an account of the moral properties
ascribed by moral judgments as intersubjective, dispositional, impression-dependent
properties. More precisely, Hume conceives of moral properties as causal properties of
objects: an object has a moral property if and only if the object causes particular
motivational states within those who satisfy various conditions. As discussed, it follows
from this account of moral properties that one can infer moral facts from non-moral
facts—e.g., facts about whether an object is either useful or agreeable to oneself or
others. §II.8 noted that this implication is widely thought to be inconsistent with some of
Hume’s claims, but it went on to argue that this putative inconsistency is due to common
misunderstandings of those claims.

115
So interpreted, this project is in full agreement with Hume regarding all of the
general points expressed above. However, as discussed, there are also important points
of disagreement between Hume and this project. For instance, §III.1 calls into question
Hume’s account of the relation between impressions, ideas, and language. Because
Hume bases his theory upon the construction of complex ideas from simple ones, he
overlooks the importance of analyzing moral judgments as first and foremost a
propositional entity rather than a mental entity. I have suggested that the propositional
nature of both thought and language makes possible direct reference to objects and
properties without the need for any corresponding experience, a position that appears to
run contrary to Hume’s view. Moreover, as discussed in §III.4, this project also remains
neutral with respect to Hume’s claim about the direct causal connection between moral
sentiments and the passions (see §III.3), as well as his claim that the moral sentiments
acquire their strong motivational force from the passions via principles of psychological
association (again see §III.3).
Chapter II (§II.8) also raised one important challenge to the proffered and
somewhat novel interpretation of Hume that we have yet to address. If, as I have argued,
Hume does indeed hold that one can infer a moral fact from (the right sort of) non-moral
facts without experiencing the relevant corresponding moral feeling at the time of the
inference, then it appears prima facie impossible in such instances for the judge to be
motivated to act in accordance with his or her belief in that moral fact—impossible
because, in Hume’s view, moral feelings are requisite for the production of moral
motivation. And this result clearly threatens to undermine the intimate connection that

116
Hume insists upon between moral thought/judgment and human conduct. Hume
repeatedly, and I think rightly, stresses that morality is essentially a practical matter, and
criticizes the rationalists’ system of morality on the grounds that it cannot account for this
fact. Yet the proffered account of Hume is in danger of falling prey to this same criticism.
The following chapter will address this issue among others. It argues that it is
possible, in Hume’s view, for one to be motivated by a moral judgment that is accepted
by inference from experience, that is, accepted without experiencing the moral sentiment
that confirms the judgment. As we shall see, sympathy can induce a moral sentiment in
response to an inferred moral judgment. In order to explain how this is possible, we will
first need to get a better grasp of the capacity for sympathy, which Hume views as the
principal foundation of morality.
Thus far we have been emphasizing the vital role that our moral feelings play in
Hume’s moral theory. These feelings are not only requisite for moral motivation; they
are also the primary means by which we draw moral distinctions, for the domain of moral
properties is determined by these feelings.
120
Of course these feelings accurately track
moral distinctions only if we satisfy the conditions of the moral view. In other words, it
is not just what we feel that matters on Hume’s view; equally important is how we come
to feel what we feel because how we come to feel affects what we feel. In short, Hume’s

120
These feelings are the primary means but not the only means because, as just mentioned, we can still
make moral distinctions without experiencing those feelings at the time of the judgment. In such instances,
we typically rely upon memories of our own past impressions to guide our judgment. But we can also rely
upon the present or past moral feelings of others, in much the same way that we sometimes rely on the past
or present color impressions of others. In either case, we must ultimately appeal to our own or others’
moral feelings, and this is due to the fact that moral distinctions are determined by these feelings.

117
moral system cannot be fully grasped without an account of the conditions that comprise
the moral view, for those conditions regulate our feelings, as well as our judgments.
However, as we shall see, the conditions of the moral view are explicated
according to the means by which the moral sentiments are engendered. In other words, if
we want an explanatory account of the moral view’s “standard conditions”, then we must
first grasp the source of our positive feelings of moral approval and our negative feelings
of moral disapproval. Hume puts the question at issue as follows:
It may now be asked in general, concerning this pain and pleasure, that
distinguishes moral good and evil, From what principles is it derived, and
whence does it arise in the human mind? (T 473)

Answering this question is, in effect, the second and final task in part one of our
project. Recall from §II.1 that the primary task is to provide insight into Hume’s moral
ontology. Hume’s view is that this is at heart a matter of better understanding the mental
principles responsible for producing the moral sentiments that determine moral
properties. Only after we have a sufficient grasp of the psychological processes and
causal mechanisms at play can we determine how best to regulate them so as to produce
stability in feeling and uniformity of judgment, which is the job of the conditions that
comprise the moral view. With a better understanding in place, Hume is confident that
we will not only provide an explanation of the moral view; we will be able to justify our
moral judgments, which ascribe moral properties. Whether Hume’s confidence is itself
justified will be addressed at the end of part one and again in part two.
Chapter III laid down some important groundwork by identifying moral
sentiments as secondary impressions (a.k.a., impressions of reflection) and briefly

118
discussing the significance of this identification. Among other things, this distinction
marks a key point at which the analogy with color breaks down. Moral sentiments,
unlike our color impressions, are mediated by beliefs. As we shall discuss throughout
this chapter, these beliefs are formed in large part as a result of causal reasoning. Thus,
reason plays an indispensable role in our drawing moral distinctions. It is the capacity of
reason that paves the way for the sentiments by which we make moral distinctions.
Nevertheless, according to Hume, it is sympathy that generates the moral
sentiments. And these feelings are, of course, responsible for motivating us to act
morally (primarily, in Hume’s view, by causing the passions). It is because there would
be no moral sentiments and hence no moral motivation without the capacity for sympathy
that Hume refers to sympathy as “the general principle of morals”.
121
Sympathy turns out
to be the key to grasping the intimate connection between moral judgment and human
conduct because it is the key to understanding how the moral sentiments and thereby
moral motives are produced.
122

In sum, Hume holds that there are two primary mental processes at play when we
make moral evaluations, to wit: sympathy and causal reasoning. Sympathy will be our

121
It is true that, in the Treatise, Hume does hold sympathy to be the general principle of morals, but it has
been suggested that, in the Enquiry, Hume adds, or even replaces it with, another general principle of
morals, namely, “the sentiment of humanity.” (Cf. Capaldi [1989].) This apparent discrepancy calls for
explanation. K. Abramson (2001) and R. Vitz (2004) argue for the explanation that I favor, namely, that
the change is a matter of marketing rather than one of substance. In other words, the discrepancy is merely
superficial. Indeed, in the Enquiry, Hume appears to view sympathy and the sentiment of humanity as
identical. He writes: “The former sentiment, to wit, that of general benevolence, or humanity, or sympathy,
we shall have occasion frequently to treat of in the course of this enquiry” (EPM, Appendix II, fn. 1).

122
According to Hume, there are no “original” or “natural” moral motives. Thus, in Hume’s view, there
must be some original, empirically verifiable, non-moral motives that help account for the moral
sentiments and moral motivation. As we shall discuss in this chapter, sympathy is also responsible for
producing one of the requisite original, empirically verifiable, non-moral motives that fulfills this function.

119
starting point and the focus of this chapter. Chapter V will focus on the role of both
sympathy and causal reasoning in our moral evaluations, as well as the important
relationship between sympathy and causal reasoning. This will set the stage for our
discussion in the following two chapters (VI and VII) about the “moral view”, which is
constituted by the principles that regulate both sympathy and causal reasoning.
§IV.2 Sympathy: the non-moral basis of morality
Hume denies that humans have any “natural” or “original” moral motive: “the
first virtuous motive, which bestows a merit on any action, can never be a regard to the
virtue of that action, but must be some other natural motive or principle” (T 478). In
other words, “no action can be virtuous, or morally good, unless there be in human
nature some motive to produce it, distinct from the sense of its morality” (T 479, Hume’s
emphasis). For instance, Hume denies that we, by nature, have benevolent concern for
the public or humanity: “there is no such passion in human minds, as the love of
mankind, merely as such, independent of personal qualities, of services, or of relation to
ourself” (T 481). Instead, he sees our sense of moral duty and “extensive benevolence”
(i.e., a general concern for others’ wellbeing) as falling out of more basic, innate non-
moral motives.
Hume also denies that our moral motives are, or can be, mere products of artifice
and education. He claims that “had not men a natural sentiment of approbation and
blame, it cou’d never be excited by politicians; nor wou’d the words laudable and praise-
worthy, blameable and odious, be any more intelligible, than if they were a language
perfectly unknown to us, as we have already observ’d” (T 579). That is, in order for

120
education and other artifices to have any motivational effect on us, we must already have
some sentiments or motives that these artifices exploit. Hume thought that these motives,
upon which the moral motives depend, must have a non-moral basis within human
nature.
According to Hume, the “natural” or “original” non-moral motives upon which
moral motives depend are self-interestedness and “limited” or “private” benevolence—
i.e., benevolent concern for family members, friends, and others who are close to us,
though we can also at times be benevolent to strangers. Significantly, then, Hume denies
that we are purely self-interested creatures, as others have claimed. Hume writes:
I am sensible, that, generally speaking, the representations of this quality
[viz., self-interestedness] have been carried too far; and that the
descriptions, which certain philosophers delight so much to form of
mankind in this particular, are as wide of nature as any accounts of
monsters, which we meet in fables and romances. So far from thinking,
that men have no affection for any thing beyond themselves, I am of the
opinion, that tho’ it be rare to meet with one, who loves any single person
better than himself; yet ‘tis as rare to meet with one, in whom all the kind
affections, taken together, do not overbalance all the selfish. (T 486-7)

In Hume’s view, we do indeed have a natural inclination toward others’ feelings
and ends, but this natural inclination is too partial and variable to constitute a moral
motive. Our naturally occurring limited benevolence is the result of our naturally
occurring sympathy, which is also partial and variable. In other words, limited
benevolence inherits its partiality and variability from sympathy. Thus, sympathy is the
origin, or source, of one of the non-moral motives upon which our moral motives depend.
In fact, as we shall see, sympathy turns out to be the origin or source of our moral
motives as well. Hume writes: “Now we have no such extensive concern for society but

121
from sympathy; and consequently ‘tis that principle, which takes us so far out of
ourselves, as to give us the same pleasure or uneasiness in the characters of others, as if
they had a tendency to our own advantage or loss” (T 579).
In order to understand how sympathy can be the source of our moral motives, and
yet it still be the case that there are no “natural” or “original” moral motives, we must
first understand Hume’s account of sympathy. Hume initially describes sympathy as that
process in which “the ideas of the affections of others are converted into the very
impressions they represent” (T 319). So Hume sees sympathy as a kind of psychological
process by which we are able to share in others’ feelings.
Hume’s basic account of sympathy is rather commonsensical. Since we do not
directly observe (i.e., feel) the emotional states of others, Hume suggests that we must
initially observe their behavior: “When any affection is infus’d by sympathy, it is at first
known to us by its effects, and by those external signs in the countenance and
conversation, which convey an idea of it” (T 317). We then infer from the observation of
another person’s behavior a belief—i.e., an idea—about what that person feels. From
this belief, we then infer (the idea) that others are experiencing pleasure or pain.
After a cause and effect relationship is established between an act or trait and
others’ emotional states—i.e., after the tendency of an act or trait to produce painful or
pleasurable results is confirmed—we can then infer others’ emotional states from various
causes. Through custom and habit, our imagination naturally takes us from the cause to
the effect or from the effect to the cause. In either case, whether the idea of another’s

122
emotional state is inferred from the cause or the effect, sympathy converts that idea into
an impression within oneself. Hume illustrates and sums up his account as follows:
We may begin considering a-new the nature and force of sympathy…
When I see the effects of passion in the voice and gesture of any person,
my mind immediately passes from these effects to their causes, and forms
such a lively idea of the passion, as is presently converted into the passion
itself. In like manner, when I perceive the causes of any emotion, my
mind is convey’d to the effects, and is actuated with a like emotion. Were
I present at any of the more terrible operations of surgery, ‘tis certain, that
even before it begun [sic], the preparation of the instruments, the laying of
the bandages in order, the heating of the irons, with all the signs of anxiety
and concern in the patient and assistants, wou’d have a great effect upon
my mind, and excite the strongest sentiments of pity and terror. No
passion of another discovers itself immediately to the mind. We are only
sensible of its causes or effects. From these [viz., our ideas of an object’s
causes and/or effects] we infer the passion: And consequently these [viz.,
our ideas of the painful or pleasurable feelings had by others] give rise to
our sympathy. (T 575-6, Hume’s emphasis)

So, according to Hume, an observer’s ideas of others’ emotional states are the
result of causal inferences. These inferences come about naturally from the observer’s
own previous experiences and observations as well as various causal beliefs formed due
to prior experience. These ideas of others’ emotional states, which are generated by
causal inference, “give rise to our sympathy”, which converts those ideas into
impressions in the observer. An important consequence of this account is that, as claimed
in chapters II and III, causal inference plays a key role in Hume’s moral theory. Causal
inference is requisite for moral motivation since it is typically requisite to trigger

123
sympathy.
123
And as we shall see, when various corrective conditions are put in place,
sympathy is the process by which moral feelings and thus moral motives are produced.
124

In short, sympathy involves an observer, another individual, that individual’s
emotional state, and finally the observer’s idea of the individual’s emotional state, be it
either a painful or pleasurable state. Through sympathy the observer’s idea of the
individual’s pain (or pleasure) is converted into an impression of pain (or pleasure)
within the observer. These very features account for why sympathy is an instance of the
association of ideas and impressions discussed in §III.3. The foregoing description of
sympathy includes: (i) the self—the source of vivacity—and someone else—the “object”
of sympathy; (ii) the idea of pain (or pleasure)—the “cause” of sympathy; and finally,
(iii) an impression of pain (or pleasure)—the “quality”.
According to Hume, sympathy’s conversion process occurs by “enlivening” the
idea of another individual’s painful or pleasurable emotional state, which transforms that
idea into an impression. Recall from chapter II (§II.2) that one way in which Hume
marks the distinction between ideas and impressions is by observing the difference with
respect to their phenomenal “force” and “vivacity”. In Hume’s view, sympathy is able to
convert our ideas of others’ emotional states into resembling impressions within us
simply by increasing the force and vivacity of those ideas.
125

123
We will say more about the role of causal reasoning within our moral evaluations later in this chapter.

124
As mentioned in §III.2, this marks one respect in which the analogy with color breaks down. Our color
impressions, unlike our moral impressions, are not mediated through causal inference and belief.

125
Or as Hume puts it elsewhere, those ideas “communicate” their vivacity to like impressions via
sympathy. This basic principle of the communication of vivacity can be seen as the proverbial glue that
holds the Treatise together. Hume discusses it with respect to his moral theory given in Book III, with

124
But this raises the question: From whence does this force and vivacity come?
Surely sympathy doesn’t create the requisite force and vivacity ex nihilo. Indeed it does
not. According to Hume, the requisite force and vivacity are provided by the self, or
more precisely, an impression of the self.
126
Sympathy, in effect, transfers the force and
vivacity of the impression of the self to particular kinds of ideas that are related to the
self. In doing so, sympathy converts those ideas into resembling impressions. Thus,
Hume writes:
In sympathy there is an evident conversion of an idea into an impression.
This conversion arises from the relation of objects to ourself [sic]. (T 320)

This explains why sympathy is naturally variable and partial. Sympathy’s natural
variability and partiality is due to the variability and partiality associated with the relation
of others to oneself. In other words, sympathy naturally varies in accordance with the
relations of resemblance, contiguity, and causality (i.e., blood-relation) between oneself
and others.
127
One’s sympathy strengthens as these relations between oneself and others
strengthen. Likewise, one’s sympathy weakens as those relations between self and other
weaken. In short, the extent of sympathy’s natural variability and partiality is explained
by Hume’s account of sympathy as a particular psychological process that converts ideas
into impressions.

respect to the passions and sympathy in Book II, and with respect to belief in Book I. For instance,
regarding its role in his account of belief, Hume writes:
…the effect, then, of belief is to raise up a simple idea to an equality with our
impressions, and bestow on it a like influence on the passions. This effect it can only
have by making an idea approach an impression in force and vivacity. (T 119)

126
See, for instance, T 317.

127
Chapters V and VI discuss this matter in much greater detail.

125
§IV.3 An objection to Hume’s account of sympathy
Hume’s account of sympathy raises an important objection to his contention that
sympathy is the non-moral basis of morality. Hume addresses this objection in the
Treatise (580-4). The argument is basically as follows. If sympathy were the source of
our moral sentiments and thereby moral distinctions (since moral distinctions are
determined by our moral sentiments) and true moral judgments, then moral sentiments,
moral distinctions, and true moral judgments would vary in accordance with the natural
variability of sympathy. But true moral judgments, moral distinctions, and the moral
sentiments indicative of moral distinctions, do not vary in accordance with the natural
variability of sympathy. Hence sympathy cannot be the source of moral distinctions or
the moral sentiments that determine them. Hume states the objection as follows:
But as… sympathy is very variable, it may be thought, that our sentiments
of morals must admit of all the same variations. We sympathize more
with persons contiguous to us, than with persons remote from us: With our
acquaintance, than with strangers: With our countrymen, than with
foreigners. But notwithstanding this variation of our sympathy, we give
the same approbation to the same moral qualities in China as in England.
They appear equally virtuous, and recommend themselves equally to the
esteem of a judicious spectator. The sympathy varies without a variation
in our esteem. Our esteem, therefore, proceeds not from sympathy. (T
580-1)

In this passage and others, Hume readily acknowledges that our sympathetic
feelings do indeed tend to vary according to the degree of contiguity, resemblance, and
causal (i.e., blood) relations between others and ourselves. It is a matter of psychological
fact that our sympathetic feelings naturally tend to be stronger when we sympathize with
those close to us, and weaker when we sympathize with strangers. Hume even admits

126
that when various comparisons such as resemblance to oneself are judged to be
sufficiently deficient or incongruous (as is often the case with complete strangers or our
enemies), our beliefs about such individuals’ emotional states may not produce any
concern for their wellbeing at all. For instance, Hume claims that “when a sympathy
with uneasiness is weak, it produces hatred and contempt” for others rather than a
concern for their wellbeing (T 385ff). In fact, as we shall discuss in chapter VI, Hume
even argues that consideration of others’ easiness—i.e., other’s pleasurable emotional
states—can sometimes result in feelings of hatred and contempt for those individuals.
It may initially seem counterintuitive that sympathy with others’ uneasiness can
produce hatred and contempt for those individuals rather than concern. But given
Hume’s account of the conversion process of sympathy and the principle the double
association of impressions and ideas (discussed in chapter III), this result is not at all
surprising.
Recall from §III.3 that hatred for an individual is caused when a “subject”
possesses a painful “quality” (i.e. it causes pain in the observer), and the subject and its
quality are related to the individual who is the “object” of hatred. In the present case, all
these conditions are met. Sympathy, in Hume’s view, converts the observer’s idea of an
individual’s uneasiness into a resembling impression within the observer, namely, an
impression of uneasiness. The individual is subsequently causally associated with one’s
own felt uneasiness produced by a sympathy with that individual. Hence we have a
“subject” (viz., the individual) and its “quality” (viz., the felt uneasiness that the subject
causes in oneself). The subject and its quality are related back to the individual via the

127
relation of identity; i.e., in this case, the “subject” and “object” of hatred are one and the
same. The result is the indirect passion of hatred toward the individual with whom one
sympathizes.
It is therefore no wonder that the observer’s sympathetic response to another’s
uneasiness might be hatred and contempt for that individual. Indeed it turns out to be
much more surprising that, on Hume’s account, sympathy for others’ uneasiness could
ever produce anything but hatred for them, much less a concern for their wellbeing that
would motivate us to help them. The answer to this conundrum is that the difference in
the outcome of a weak verses strong sympathy turns out to be slightly more complex than
a mere difference in the vivacity of the feeling that each produces.
On the one hand, when sympathy is weak, a weak force and vivacity is
communicated to the observer’s idea of the other individual’s emotional state, which
results in a weak impression in the observer. The influence of the idea of the individual’s
emotional state is not strong enough to extend beyond the individual’s present emotional
state to ideas of future emotional states of the individual; e.g., we only consider our
enemy’s immediate misery or uneasiness. And for the reasons discussed above, this
produces hatred toward that person.
On the other hand, when sympathy is strong, a strong force and vivacity is
communicated to the observer’s idea of the individual’s emotional state, and this results
in a strong impression in the observer. Because of the strength of the idea’s force and
vivacity, its influence extends to associated ideas, namely, ideas of the individual’s future
feelings, as well as ideas of the past, present, and future circumstances that will more than

128
likely causally contribute to those feelings. In Hume’s view, consideration of the
prospect of the individual’s present and future feelings thereby generates, via sympathy, a
concern for that individual’s wellbeing. Hume writes:
When the present misery of another has any strong influence upon me, the
vivacity of the conception is not confin’d merely to its immediate object,
but diffuses its influence over all the related ideas, and gives me a lively
notion of all the circumstances of that person, whether past, present, or
future; possible, probable or certain. By means of this lively notion I am
interested in them; take part with them; and feel a sympathetic motion in
my breast, conformable to whatever I imagine in his. If I diminish the
vivacity of the first conception, I diminish that of the related ideas… By
this diminution I destroy the future prospect, which is necessary to interest
me perfectly in the fortune of another. I may feel the present impression,
but carry my sympathy no farther… But as I am not so much interested as
to concern myself in his good fortune, as well as his bad, I never feel the
extensive sympathy, nor the passions related to it. (T 386)

Before we discuss Hume’s response to the aforementioned objection to his
account in the next section, it is worth noting that in this passage Hume in effect draws a
distinction between two different kinds of sympathy: “extensive sympathy” and “limited
sympathy”.
128
This leads to an ambiguity in the unqualified use of the term ‘sympathy’.
Hume uses the term to refer to limited sympathy in some cases and extensive sympathy in
others;
129
and sometimes he even seems to use it to refer to both kinds of sympathy. This
is noteworthy because the two kinds of sympathy produce quite different outcomes.
Hume typically uses the term ‘sympathy’ to refer to limited sympathy, which is
nothing more than the naturally occurring but highly variable and partial psychological
process described above in which an idea of another’s pain or pleasure is “enlivened” into

128
Elsewhere Hume is quite explicit about this distinction; see T 387.

129
Kate Abramson (1999: 342) also observes this ambiguity in Hume’s use of ‘sympathy’.

129
an impression within oneself.
130
Limited sympathy is, at best, only able to produce
limited benevolence, which is also highly variable and partial. Due to the extent of its
variability and partiality, limited benevolence cannot serve as a moral motive.
Extensive sympathy also includes the former psychological process in which the
idea of another’s emotional state is converted into an impression within oneself. But in
the case of extensive sympathy, this process is regulated such that the force and vivacity
communicated to the idea results in an impression that is both strong and extensive
enough to produce the moral sentiments and extensive benevolence—i.e., a general
concern for others’ wellbeing—both of which serve as moral motives.
131
In other words,
it is this regulated, extensive sympathy that accounts for production of moral motivation.
Unfortunately, the ambiguity in Hume’s use of the term ‘sympathy’ is worsened
by the fact that he also sometimes uses the term to refer to either the psychological
process of sympathy plus the feeling it produces or simply to the feeling itself. Consider,
for instance, the passage (T 386) above.
132
Potential for further confusion arises because,
as we have just observed, extensive sympathy produces moral feelings and extensive
benevolence, whereas the naturally occurring unregulated limited sympathy does not.
Also noteworthy is the fact that Hume doesn’t explicitly say whether he conceives
of the sympathetic feelings that extensive sympathy produces as identical to moral
feelings, or whether he conceives of moral sentiments as direct responses to these

130
See, for instance, T 317, 319-20, and 427.

131
See, for instance, T 586.

132
Also, recall the passage noted earlier in which Hume explicitly refers to sympathy as a sentiment: “The
former sentiment, to wit, that of general benevolence, or humanity, or sympathy, we shall have occasion
frequently to treat of in the course of this enquiry” (EPM, Appendix II, fn. 1).

130
sympathetic feelings via principles of mental association. We will not dwell on this issue
at this point however because it is clear that Hume treats our extensive sympathetic
feelings and our moral feelings as at least equivalent in much the same way that our
moral feelings are equivalent to our feelings of pride and humility, and love and hatred
(when had under certain conditions).
133
These feelings are all equivalent insofar as they
are all indicative of moral distinctions. For instance, recall the following passage:
We have…observ’d, that moral distinctions depend entirely on certain
peculiar sentiments of pain and pleasure… Now since every quality in
ourselves or others which gives pleasure, always causes pride or love; as
every one, that produces uneasiness, excites humility or hatred: It follows,
that these two particulars are to be consider’d as equivalent, with regard to
our mental qualities, virtue and the power of producing love or pride, vice
and the power of producing humility or hatred. In every case, therefore,
we must judge of the one by the other; and may pronounce any quality of
the mind virtuous, which causes love or pride; and any one vicious, which
causes hatred or humility. (T 574-75, emphasis added)

§IV.4 Hume’s reply & the transition from non-moral to moral motives
It is a distinct advantage of Hume’s conception of sympathy that it can account
for the extent of the variation in the sympathetic pleasures or pains we experience; i.e., it
adequately describes this feature of our experience. As we have just briefly discussed,
Hume can even explain why sympathy can sometimes produce hatred and contempt for
others rather than a concern for their wellbeing.
134
But the question at issue, which is
raised by the objection, is whether Hume’s theory can account for the fact that we
nevertheless judge that relevantly similar acts or traits are equally virtuous or vicious,

133
§V.5 argues that sympathetic feelings are not identical to moral feelings of approval and disapproval.

134
We will say more about this in chapters V and VI.

131
morally good or bad, right or wrong, etc., regardless of any variation in the degree of the
pleasure or (as the case may be) pain that sympathy produces within us.
Hume confidently claims that “’tis the easiest matter in the world to account for
it” (T 581). He simply appeals to various regulatory principles, or “general rules”, that
comprise one part of the “general” moral view.
135
When applied successfully, these
principles eliminate, or at least sufficiently mitigate, the variability and partiality of our
naturally occurring limited sympathy in order to ensure an extensive sympathy and
thereby extensive benevolence.
136
Thus, besides the important difference in outcome,
extensive sympathy differs from limited sympathy insofar as it is the product of artifice.
This is significant because it explains why there are no “natural” moral motives.
However, there are many instances in which we do not successfully apply those
regulatory principles. In these instances, the process of extensive sympathy will not
occur. Nevertheless, in such cases, we still can, and presumably often do, issue moral
judgments that do not reflect the variability and partiality of our unregulated sympathy.
Hence, as things stand, this reply does not fully answer the objection. As we shall see in
a moment, the general rules to which Hume appeals can fully answer the objection. But
before we discuss how they do so, I would like to quickly point out that we have now
acquired the materials to lay out the basic structure of Hume’s position regarding the
transition from non-moral motives to moral motives.

135
As we shall discuss in detail in the next chapter, there is a second part to the moral view, and this part is
comprised of general rules that govern causal reasoning, which is usually requisite to produce sympathy.
Together both sets of regulatory principles—viz., the set governing causal reasoning and the set governing
sympathy—constitute the entirety of the moral view.

136
Much more will be said about this in chapters VI and VII.

132
Recall that all moral motives, such as extensive benevolence, depend upon the
non-moral motives of self-interestedness and limited benevolence. Our naturally
occurring limited benevolence is the result of our naturally occurring but limited
sympathy. This unregulated psychological process of sympathy is “limited” due to the
extent of its natural variability and partiality.
We now know that Hume transforms this natural but limited sympathy into
extensive sympathy by subjecting it to various corrective conditions. Extensive
sympathy is, of course, responsible for producing extensive benevolence and the moral
sentiments, which are moral motives. Thus, Hume accounts for the transition from non-
moral motives to moral motives by appealing to various regulatory principles that govern
sympathy’s natural variability and partiality. In Hume’s view, there are no “original” or
“natural” moral motives precisely because moral motives are, at least in part, a product of
artifice. They are in part a product of regulatory principles that do not occur naturally.
But this raises the following question: How and why do the non-moral motives of
self-interestedness and limited benevolence motivate us to regulate the psychological
process of sympathy with general rules? This question becomes all the more pressing
once it is apparent that an extensive sympathy and the principles that govern it sometimes
undermine limited benevolence and quite often undermine our own self-interest.
I have brought up Hume’s position regarding the transition from non-moral
motives to moral motives and raised the above question at this point because Hume’s
answer to this question is closely related to his full answer to the objection against his
account of sympathy as the non-moral basis of morality. According to Hume, our moral

133
judgments do not vary with variations in our sympathy precisely because, as we shall
discuss in a moment, the general rules that regulate our sympathy serve to establish a
“common”, or shared, perspective, namely, the general moral view. This shared moral
point of view brings uniformity and thereby stability to our moral judgments. And it
turns out that the uniformity and stability brought to our moral judgments by the moral
view are in our own interest, as well as the interest of our family and friends.
Let’s tackle the question concerning the transition from non-moral motives to
moral motives first. Hume provides a practical answer to this question. We are initially
motivated by both self-interestedness and limited benevolence to establish a common
perspective so that we can better communicate with one another and benefit from the
practical consequences that this shared perspective provides.
137
Hume writes:
[E]very particular man has a peculiar position with regard to others; and
‘tis impossible we cou’d ever converse together on any reasonable terms,
were each of us to consider characters and persons, only as they appear
from his peculiar point of view. In order, therefore, to prevent those
continual contradictions, and arrive at a more stable judgment of things,

137
This answer presupposes the formation of society. Hume explicitly states that individuals initially chose
to construct and submit to societal regulations according to the non-moral motives of self-interestedness
and limited benevolence. Such rules are necessary for a stable society, and a stable society offers plenty of
benefits for oneself, as well as one’s family and friends, that could not be had otherwise (T 485ff). Thus, to
a large degree, society itself comes about from the same non-moral motives as morality. But Hume also
stresses that human beings are social creatures by nature, just as we are to some extent naturally
benevolent. He consequently rejects the notion of a state of nature as a mere philosophical fiction (T 493).
Moreover, Hume goes on to argue that our capacity for sympathy is what accounts for our sense of duty
towards justice, which arises only after that system has been put in place. Hume writes:
Nay when the injustice is so distant from us, as no way to affect our interest, it still
displeases us; because we consider it as prejudicial to human society, and pernicious to
every one that approaches the person guilty of it. We partake of their uneasiness by
sympathy…this is the reason why the sense of moral good and evil follows upon justice
and injustice…Thus self-interest is the original motive to the establishment of justice: but
a sympathy with public interest is the source of the moral approbation, which attends that
virtue. (T 499-500, Hume’s emphasis)
Later in this chapter I shall argue that Hume makes a similar move with regard to our sense of duty toward
morality itself.

134
we fix on some steady and general points of view; and always, in our
thoughts, place ourselves in them, whatever may be our present situation.
(T 581-2, Hume’s emphasis)

Without fixing on some “steady and general” points of view, our judgments
would lack uniformity and would consequently be unstable. That is, without shared
perspectives from which to issue our judgments, we would always talk past one another.
Although, in such cases, our judgments are not strictly speaking “contradictions”,
Hume’s point is that they feel like contradictions.
138
According to Hume, judgments that
lack uniformity and stability tend to produce feelings of uneasiness within us. Moreover,
the lack of uniformity and stability in our unregulated impressions and judgments has
important practical implications that can also produce uneasiness in us. For instance, it
may lead to dire consequences for one’s loved ones and/or oneself.
139

Any lack of uniformity and stability in our judgments is primarily due to the
constant variability in our impressions. Steady and general points of view function as
corrective measures for this variability in our impressions, thereby bringing uniformity
and stability to our judgments. This is why Hume claims that we seek out such corrective
measures for all of our senses and corresponding judgments.
140
For instance, Hume

138
See also Louis Loeb (2002 & 2004a), who forcefully argues this point.

139
Loeb and I agree that our feelings of uneasiness are what motivate us to seek out and take up shared
perspectives, but he appeals only to the uneasy feelings associated with the pseudo-contradictions. He
overlooks the uneasy feelings that result from perceived (or imagined) negative consequences. These
feelings are important because they help explain why self-interestedness and limited benevolence motivate
us to choose the moral point of view over other shared perspectives. (See the following discussion.) It is
also noteworthy that the uneasy feelings associated with the perceived harm that may come to family and
friends—harm that results from the lack of a shared perspective—are the product of (limited) sympathy.

140
Hume clearly views our capacity for sympathy—i.e., the capability to partake in that psychological
process—as a kind of sense; see, for instance, T 499. It is the mental faculty by which we apprehend moral
properties.

135
writes: “Such corrections are common with regard to all the senses; and indeed ’twere
impossible we cou’d ever make use of language, or communicate our sentiments to one
another, did we not correct the momentary appearances of things, and overlook our
present situation” (T 582).
For reasons that we will discuss shortly, shared perspectives enable us to talk
about the same thing rather than only being able to talk about our own fluctuating
experiences. For instance, without a common perspective, we could not give and receive
reliable advice to others about what to do, pursue, or choose. A shared perspective is, in
this way, beneficial to all those who wish to communicate. Indeed Hume appears to
insinuate in the above passage that the very possibility of using language presupposes
that we have already naturally taken up at least some sort of shared point of view.
In short, self-interestedness and limited benevolence motivate us to seek out and
take up common points of view in order to curtail all of the negative feelings that would
otherwise arise in ourselves and our family and friends. Such shared viewpoints curtail
those feelings by bringing uniformity and stability to our feelings and judgments. And
they bring this uniformity and stability by implementing “general rules”, which
standardize the conditions by which we issue our judgments. In other words, steady and
general points of view are, in effect, nothing more than various sets of normative
principles that we use to govern both our senses and our judgments.
141

141
Loeb (2002) argues that Hume holds a stability account of normativity, according to which the stability
provided by the uniformity of our judgments positively contributes to their normative status. On this view,
our feelings turn out to be the original source of normativity. Korsgaard (1996), on the other hand, argues
that Hume holds a reflective endorsement account of normativity, according to which moral judgments
have a positive normative status only in virtue of reflective endorsement. I basically side with Loeb on this

136
These normative principles govern our senses and our judgments by drawing a
vital distinction between what appears to be the case and what actually is the case. For
instance, Hume observes that, as we move further away from an object, its size appears to
decrease; yet we do not judge that the object’s size diminishes according to the distance
we stand from it (T 582). The object remains the same size despite appearances to the
contrary. In this case, as in most cases concerning judgments that rely upon our senses,
we apply regulatory principles that “correct the momentary appearances of things”, or at
least govern our judgments of how things actually are as opposed to how they appear to
us at the moment. In other words, we may need to “overlook our present situation” and
focus on how things would appear under the appropriate conditions for judging the
object—e.g., the appropriate conditions for judging an object’s size or color, such as
being a normal visual observer who is fairly close to the object and observing it in broad
daylight or similar lighting conditions.
It is this distinction between appearance and fact that enables us to talk about the
same thing rather than simply talk past one another. Of course, as we have just seen,
there are facts about the way things appear to us under various conditions. And as the
conditions vary so will appearances. This fact helps explain Hume’s claim that our
judgments about the same object can sometimes seem, or “feel”, like they contradict one
another. For example, under the conditions in which you observe a particular object, it
might appear small and dull colored, but under the conditions in which I view the object,

issue, although I suggest that stability is also produced by the positive practical consequences that result
from taking up the moral view, as opposed to some other common perspective. This account, however,
does not discount the relevance of reflection for normativity because reflection can add stability. But it
does abandon the notion that reflection is the only source of normativity.

137
it might appear large and bright colored. If our judgments about that object simply reflect
these “momentary appearances” of the object from our own “peculiar position”, then our
judgments can seem to be in tension with one another, and thereby “feel” like
contradictions. After all, in this example, you would judge that the object is small and
dull colored, and I would judge that it is large and bright colored.
To avoid any uneasy feelings associated with such pseudo-contradictions and any
perceived negative consequences that arise from the lack of uniformity and stability in
our unregulated impressions and judgments, we “fix on some steady and general points of
view.” In other words, we introduce and adopt a set of principles that establish a
standard of correctness for both our sentiments and our judgments by providing the
resources needed to distinguish between the momentary appearances of things and how
things really are. These “standard conditions” thereby bring uniformity and stability.
So, in Hume’s view, things are as they would appear to a qualified observer—i.e.,
anyone who is observing things under the requisite conditions. Of course what ends up
counting as the requisite set of standard conditions will, on Hume’s view, depend upon
the nature of what is being judged, as well as the nature of those who are doing the
judging.
142
Take, for instance, our color judgments. The set of standard conditions for
judging the color of an object is determined by the nature of color and our color
sensibilities, e.g., how particular spectrums of light, the human eye, and brain causally
interact so as to produce color sensations.

142
Cf. Sayre-McCord (1994).

138
This set of conditions acts as a standard of correctness for our impressions insofar
as we try our best to satisfy those conditions so that our impressions, such as our color
impressions, are indicative of the relevant properties—e.g., an object’s color. After all,
the color of an object is (arguably) determined by the color the object would appear to an
observer who satisfies a particular set of conditions that is requisite for apprehending the
color of objects.
But equally important, this set of conditions also acts as a standard of correctness
for our judgments. That is, the set of conditions serve to regulate our judgments, even in
those cases in which we fail to regulate our sentiments because we do not or cannot meet
all of those conditions. Our judgments are to reflect what sentiments we believe we
would experience were we to satisfy the requisite conditions, and thus our judgments are
to reflect what judgments we believe we would arrive at were we to satisfy those
conditions. Hence, when standard conditions are put into place, our judgments will no
longer vary in conjunction with our occurrent sentiments. In this way, the standardizing
conditions of shared points of view ensure uniformity and stability of our judgments.
We now have a complete answer to the objection that sympathy cannot be the
source of moral sentiments, moral distinctions, and our moral judgments because true
moral judgments, moral distinctions, and the moral sentiments that determine them, do
not vary along with the variations of our sympathy and the feelings it produces. The
general rules of the moral view set a standard of correctness for our moral judgments
even in cases in which we fail to correct our occurrent sentiments via those rules. In
short, even in those cases in which we do not or cannot satisfy the conditions of the moral

139
view, our moral judgments are supposed to reflect what we would feel were we to satisfy
those conditions and thereby experience extensive sympathy.
For instance, in taking up the moral view, Hume claims that we are to “consider
not whether the persons, affected by the qualities [in question], be our acquaintance or
strangers, countrymen or foreigners”; for the conditions of the moral view require that we
“over-look our own interest in those general judgments” (T 582). This is why, when
judging in accordance with the moral view, “[w]e blame equally a bad action, which we
read in history, with one perform’d in our neighbourhood t’other day” (T 584), even if
we fail to satisfy the conditions of the moral view; for “we know from reflexion, that the
former action wou’d excite as strong sentiments of disapprobation as the latter, were it
plac’d in the same position” (ibid). And we therefore judge accordingly.
§IV.5 Shared perspectives, objectivity, & the proffered account
I suggest Hume thinks that, in this way, the standardizing conditions of the shared
moral perspective are responsible for giving moral thought and practice the sort of
objectivity that they are usually presumed to have. They accomplish this primarily by
regulating sympathy and our causal reasoning, which ensures uniformity and stability of
judgment and produces uniformity of experience among those who meet those
conditions. The result is a dispositional, impression-dependency account of objects’
moral properties. This account is analogous to the account of objects’ colors. Various
standardizing conditions produce uniformity and stability in our color judgments, and
uniformity of our experience of color, at least among those who satisfy those conditions.
The result is a dispositional, impression-dependency account of objects’ colors.

140
The above account raises the following crucial question however: Why choose
one particular shared perspective rather than another as revealing the facts, such as, for
instance, the actual or “real” color of things? The short answer, which I have already
given, is that the set of standard conditions that comprise a shared perspective is
determined by the nature of the property in question (e.g., color) and the nature of the
relevant sensibility (e.g., our color sensibility). Let me now say a little more about this.
First, the sensibility must be widely shared because the standard conditions must
be widely accessible, otherwise those conditions would not comprise a common point of
view; hence they would not bring uniformity and stability. Second, the particular
conditions for judging color—e.g., broad daylight, being relatively close to the object,
etc—are adopted because they are mutually accessible conditions that lead to the richest
color impressions and thereby the finest color distinctions. Finally, taking up a widely
accessible perspective that offers us the most refined color discriminations not only best
regulates our color language, it also has important practical implications. It can amount
to the difference between life and death, as for example when learning (or teaching
others) how to use the color of flora to identify a specimen as medicinal or poisonous.
143

143
Korsgaard (1999) overlooks such practical implications and their importance when she writes:
All we need is to establish some convention about the point of view we will use for
making these [viz., color] judgments; and the fact that sunlight enables us to make the
most discriminations seems sufficient reason to favor it. But in this case, all that we are
determining is how it is best to talk… Yet in the moral and aesthetic cases, more seems to
be at stake…Presumably we are determining the direction in which we should cultivate
our tastes, who is entitled to our love and services, and what we ourselves ought to try to
be like. (16, her emphasis)
In other words, choosing a particular common perspective in the moral and aesthetic cases is more
than simply determining how it is best to talk; it has important practical implications as well. But
choosing a particular common perspective in the color case also clearly has important practical
implications. For example, we can use color to determine what we should and should not ingest.
Moreover, the standard conditions for judging color do indeed determine the direction in which we

141
Similarly, in the moral case, the standard conditions of the moral view are
adopted because they are mutually accessible conditions that lead to the richest moral
impressions, in particular, a concern for the wellbeing of others. In other words, the
regulative principles of the moral view are those that are most likely to cause prosocial
behavior within a wide range of subjects under a wide range of conditions. We are
motivated toward this perspective not only because it brings uniformity and stability; the
refined moral discriminations it offers also have important practical implications. For
instance, it provides greater social cohesion and cooperation, which benefit both one’s
loved ones and oneself.
§IV.6 Sympathy, the principles of the moral view, & disinterestedness
We shall say much more about the regulative principles of the moral view in the
next two chapters. But the following quotation about that shared perspective will serve
our present purposes:
…[T]is impossible men cou’d ever agree in their sentiments and
judgments, unless they chose some common point of view, from which
they might survey their object, and which might cause it to appear the
same to all of them. Now, in judging of characters, the only interest or
pleasure, which appears the same to every spectator, is that of the person
himself, whose character is examin’d; or that of persons, who have a
connexion with him. And tho’ such interests and pleasures touch us more
faintly than our own, yet being more constant and universal, they counter-
balance the latter even in practice, and are alone admitted in speculation as
the standard of virtue and morality. They alone produce the particular
feeling or sentiment, on which moral distinctions depend. (T 591)

cultivate color sensibilities; one need only ask a professional artist, an interior or exterior
decorator, etc. Hence I suggest that the color case and the moral case are much more analogous
than Korsgaard recognizes.

142
As this passage reveals, our moral feelings are indicative of moral distinctions
only when we satisfy the standard conditions of the common moral point of view. This is
because moral properties are determined by the moral feelings that we would feel upon
fulfilling the conditions of the moral view. As Hume puts it: “All morality depends upon
our sentiments; and when any action, or quality of the mind, pleases us after a certain
manner, we say it is vicious; and when the neglect, or non-performance of it, displeases
us after a like manner, we say that we lie under an obligation to perform it” (T 517,
Hume’s emphasis). The “manner” of which Hume speaks is disinterestedness—defined
(for now) as not being influenced by considerations of personal advantage.
144
According
to Hume: “Moral feelings must come from Disinterestedness, not self-interest, nor
limited benevolence, for only then can we have a common point of view that can serve as
a common standard for agreement and insure uniformity of judgment” (602-3).
It should come as no great surprise that morality is so closely related to
disinterestedness. After all, morality requires that we put aside our own motives and
interests and consider the feelings of others (e.g., how our actions affect others).
Although it may initially seem peculiar that the non-moral motives of self-interest and
limited benevolence would give rise to disinterestedness, we have already discussed how
this occurs. Roughly, disinterestedness brings uniformity and stability to our moral
judgments, refined moral discriminations, and important positive effects. And all of
these factors are in our own interest as well as the interest of our family and friends.

144
In chapters V and VI, I shall argue that Hume’s conception of disinterestedness turns out to be more
robust than what this definition suggests. As we shall also discuss in chapter VI, it is important not to
equate Hume’s conception of disinterestedness with impartiality, although these two notions are closely
related.

143
Thus, in Hume’s view, limited sympathy plays a key role in producing
disinterestedness. And disinterestedness is the means by which limited sympathy is
transformed into extensive sympathy. Indeed, extensive sympathy can be understood as a
kind of disinterested sympathy. This disinterested sympathy produces the moral motives
of extensive benevolence and the moral sentiments. Hence Hume offers an account of
sympathy that explains the vitally important transition from non-disinterested, non-moral
motives to disinterested, moral motives—vitally important because there are no “natural”
moral motives. Sympathy is the “general principle of morals” precisely because it
produces disinterestedness and thereby the moral sentiments and moral motivation.
§IV.7 A second objection to Hume’s account & Hume’s reply
Hume considers a second objection to his account of sympathy as the non-moral
basis of morality. The objection is basically as follows. As we know, Hume holds that
the domain of moral properties are determined by our moral feelings, at least when we
satisfy the conditions of the disinterested moral view. These moral feelings are
engendered by (extensive) sympathy, and sympathy is triggered by causal inference, in
particular, our beliefs about the effects of particular actions, character traits, etc.
However, we often make moral distinctions even when one is prevented from performing
a particular act or exercising a particular character trait. We often, in effect, judge an
agent’s intentions to act rather than the agent’s actions themselves or their outcomes.
But, in this case, when an act has somehow been foiled, there are no positive or
negative effects to which our sympathy can respond. In other words, because no act has
been performed, there are no painful or pleasurable feelings with which we can

144
sympathize. Hence it seems no moral feelings can be engendered in such cases since
moral feelings are a product of (extensive) sympathy. Hence it appears Hume is
committed to the view that there are no moral distinctions in such cases since moral
distinctions are determined by moral feeling and no moral feelings are produced.
This conclusion is clearly counter-intuitive however. For instance, commonsense
suggests that a person or a trait can be virtuous or vicious even if the person is prevented
from acting as such, and the trait is prevented from being exercised. As Hume writes:
Virtue in rags is still virtue; and the love, which it procures, attends a man
into a dungeon or desart [sic], where the virtue can no longer be exerted in
action, and is lost to the world. Now this may be esteem’d an objection to
the present system. Sympathy interests us in the good of mankind; and if
sympathy were the source of our esteem for virtue, that sentiment of
approbation cou’d only take place, where the virtue actually attain’d its
end, and was beneficial to mankind. There it fails of its end, ‘tis only an
imperfect means; and therefore can never acquire any merit from that end.
The goodness of an end can bestow a merit on such means alone as are
compleat, and actually produce the end. (T 584)

Hume answers the objection by once again appealing to general rules.
Significantly, general rules allow us to assess and respond to the causal tendency of
objects—e.g., acts, character traits, persons, etc.—to bring about particular effects, or
consequences. These general rules regarding the tendency of objects to cause particular
effects are formed according to the uniformity of our experience and the predominance of
positive over negative outcomes. In other words, the outcomes experienced tend to
support a general rule about probable future outcomes. As Hume puts it, “…in order to
establish a general rule, and extend it beyond its proper bounds, there is requir’d a certain

145
uniformity in our experience, and a great superiority of those instances, which are
conformable to the rule, above the contrary” (T 362). Earlier he writes:
Shou’d it be demanded why men form general rules, and allow them to
influence their judgment, even contrary to present observation and
experience, I shou’d reply, that in my opinion it proceeds from those very
principles, on which all judgments concerning causes in the facts depend.
Our judgments concerning cause and effect are deriv’d from habit and
experience; and when we have been accustom’d to see one object united to
another, our imagination passes from the first to the second, by natural
transition, which precedes reflection, and which cannot be prevented by it.
(T 147)

Hume quite plausibly holds that the mere contemplation of the tendency of objects
to produce painful or pleasurable effects in others can have an impact on our beliefs.
This is primarily due to the workings of the imagination in conjunction with habit and
custom. The resulting effect on our causal beliefs can trigger (extensive) sympathy.
Hence the application of general rules can explain the presence of moral feelings of
approval or (as the case may be) disapproval toward the object in question, even in those
instances in which the object fails to actually produce painful or pleasurable effects. As
Hume puts it:
Where a character is, in every respect, fitted to be beneficial to society, the
imagination passes easily from the cause to the effect, without considering
that there are still some circumstances wanting to render the cause a
compleat one. General rules create a species of probability, which
sometimes influence the judgment, and always the imagination. (T 585,
Hume’s emphasis)

In this way, general rules can account for moral distinctions in the cases at issue.
But there is still a small but related issue to be addressed. Hume claims that the imagined
effects of, for example, a benevolent trait do not typically produce as strong a sentiment

146
as would be produced were the effects of the trait actually observed. Nevertheless, our
moral judgments about the trait do not vary along with those variations in our sentiment.
According to Hume, this is because “[w]e know, that an alteration of fortune may render
the benevolent disposition entirely impotent; and therefore we separate, as much as
possible, the fortune from the disposition” (T 585).
Consequently, when issuing moral judgments, we are to apply rules that tell us to
focus as much as possible on the typical effects of an object, rather than its actual effects
(if any). Of course, we will not always manage to overlook the actual effects of an act or
trait and respond merely to its usual effects. In those cases in which the actual effects
correspond to what we judge to be the usual effects of the act or trait our moral sentiment
tends to be stronger, or more “vivacious”, than it is in those cases in which its actual
effects fail to correspond to what we judge to be the usual effects.
145
This fact of human
psychology provides one—but only one—invaluable explanation as to why we
sometimes fail to act upon our moral judgments. Hume writes:
Upon these principles we may easily remove any contradiction, which
may appear to be betwixt the extensive sympathy, on which our sentiments
of virtue depend, and that limited generosity which I have frequently
observ’d to be natural to men…My sympathy with another may give me
the sentiment of pain and disapprobation, when any object is presented,
that has a tendency to give him uneasiness; tho’ I may not be willing to
sacrifice any thing of my own interest, or cross any of my passions, for his
satisfaction…Sentiments must touch the heart, to make them control our
passions: But they need not extend beyond the imagination, to make them
influence our taste. (T 586)

Hume immediately goes on to illustrate this using the following example:

145
See, for instance, T 585.

147
When a building seems clumsy and tottering to the eye, it is ugly and
disagreeable; tho’ we be fully assur’d of the solidity of the workmanship.
‘Tis a kind of fear, which causes this sentiment of disapprobation; but the
passion is not the same with that which we feel, when oblig’d to stand
under a wall, that we really think tottering and insecure. The seeming
tendencies of objects affect the mind: And the emotions they excite are of
a like species with those, which proceed from the real consequences of
objects, but their feeling is different [viz., less vivacious]. (T 586)

The important point to take away from this is that, when our moral feelings are
weak, the passions they produce will also be weak. Consequently, these passions are
often too weak to sway other conflicting passions. When this occurs, we are much more
likely to be motivated by those conflicting passions rather than the weaker passions
produced by our moral sentiments. In short, the moral motives produced by extensive
sympathy are sometimes so weak that they are outweighed by the non-moral motives of
limited benevolence and self-interest. Hence there is no guarantee that our moral motives
will always trump our non-moral motives. Nonetheless, using the general rules of the
moral view, we can still make the right moral judgment even without the corresponding
moral motivation.
§IV.8 Two final points about the general principles of the moral view
Before we conclude this chapter, there are two other noteworthy points to make
about Hume’s use of the “general rules” of the moral view. First, Hume appeals to these
general rules to explain how one is able to issue moral judgments about oneself and one’s
own actions and traits. For instance, Hume writes: “And tho’ this sense [viz., extensive
sympathy], in the present case, be deriv’d only from contemplating the actions of others,
yet we fail not to extend it even to our own actions. The general rule reaches beyond

148
those instances, from which it arose” (T 499, emphasis is Hume’s). In other words, one
can morally evaluate oneself by extensively, or disinterestedly, sympathizing with oneself
using the general rules of the moral view. Thus, the foundations of Hume’s moral
theory—viz., sympathy and the general rules that comprise the moral view—not only
explain why morality is inherently other-regarding; significantly, they also account for
the fact that morality can be self-regarding in a disinterested manner.
Second, it is also noteworthy that Hume’s replies to the two objections that he
considers against his account of sympathy demarcate an important division between the
general rules that comprise the moral view. On the one hand, when Hume replies to the
first objection (§IV.4), he appeals to general rules that govern sympathy. Recall that
these are the regulatory principles that function as a standard of correctness for both our
natural but limited sympathy and our moral judgments. When we issue moral judgments,
we are to attempt to satisfy the principles that ensure our sympathy will be disinterested,
or “extensive.” And in those cases in which we fail to regulate our sympathy, we are still
to judge according to what we would feel were our sympathy disinterested.
On the other hand, Hume’s reply to the second objection (§IV.7) appeals to
general rules that govern causal reasoning—viz., our reasoning about the causes and
effects of various objects, as well as the causal tendencies of those objects to produce
various effects. When we issue moral judgments, the general rules of the moral view
require us to focus on the tendency of the object under moral evaluation to cause
emotional effects.

149
Taken together, both of Hume’s replies also make it clear that these two sets of
regulative principles are integrally related to one another because sympathy and causal
reasoning are integrally related to each other. The following chapter (chapter V) will
discuss this relationship and its significance. Chapters VI and VII will then focus on the
two sets of general rules that together comprise the disinterested moral view and govern
both sympathy and causal reasoning.

150
Chapter V: Sympathy, Moral Motivation, and Causal Reasoning
§V.1 A problem for moral motivation: an objection to the account
Let us take stock at this point. I have argued that, in Hume’s view, our moral
judgments ascribe moral properties. Hume holds that these moral properties are
determined by the feelings we have when we successfully take up the moral view. And
these moral feelings are produced by sympathy governed by a set of rules that partially
comprises the moral view, i.e., extensive sympathy. By deploying these rules, we
regulate the natural variations in our sympathetic feelings in order to acquire uniformity
and stability of moral feeling and judgment, which is desirable for non-moral reasons. In
Hume’s view, our moral judgments presuppose that, ceteris paribus, psychologically
normal human observers would have the same basic moral experience were they to
evaluate the object in question under the standardizing conditions, or rules, that govern
our moral judgments. These ceteris paribus conditions involve agreement in belief about
the non-moral facts of the matter, such as whether an act or trait causes, or tends to cause,
painful or pleasurable effects.
146
The other set of rules that comprise the moral view
regulate such causal beliefs.
Because of this account, Hume is able to hold that, even if I fail to adequately
govern my sympathy and the moral value (or virtue) of the object under evaluation
consequently appears (i.e., feels) different to me due to variations in contiguity,
resemblance, and other relations, I am to overlook these variations and continue to judge

146
To recapitulate, these ceteris paribus conditions mark an important difference between our color and
moral impressions: namely, our moral impressions are mediated through our beliefs.

151
according to the rules of the moral view. The fact that our moral judgments can remain
constant, even when our sentiments vary, presupposes the vital distinction between moral
sentiment and judgment, which we discussed at some length in chapter II. This
distinction is crucial because it enables us to secure uniformity and stability of judgment
even in cases in which we lack uniformity and stability of sentiment. Hume writes:
In general, all sentiments of blame or praise are variable, according to our
situation of nearness or remoteness, with regard to the person blam’d or
prais’d, and according to the present disposition of our mind. But these
variations we regard not in our general decisions, but still apply the terms
expressive of our liking or dislike, in the same manner, as if we remain’d
in one point of view. Experience soon teaches us this method of
correcting our sentiments, or at least, of correcting our language, where
the sentiments are more stubborn and inalterable. (T 582)

As we’ve discussed in the previous chapter, if we fail to adequately follow all of
the moral view’s general rules, our moral sentiments may be too weak to trump our non-
moral motives. Nevertheless, based on these normative principles, we can recognize
what moral judgment we ought to make. Significantly, in the above passage, Hume
appears to go one step further and claim that, in some cases in which we fail to govern
our sympathy in accordance with the rules of the moral view, no moral sentiment may be
produced at all. Yet, even in these instances, we can still recognize what moral judgment
we ought to make.
The veracity of both of these normative judgments is based on counterfactuals
about what we would feel were we to observe or contemplate the relevant circumstances
upon satisfying the conditions of the moral view. And such counterfactual judgments
are, in turn, based upon habit and prior experiences, as well as other causal judgments. In

152
short, Hume holds that a moral sentiment that confirms a given moral judgment is not
requisite. As argued in chapter II, we can derive moral facts from (the right sort of) non-
moral facts. This account falls directly out of a proper understanding of Hume’s
distinction between moral sentiment and judgment.
Some have questioned this account however, noting that in the now famous is-
ought paragraph Hume claims that no moral ought can be deduced from what is the
case.
147
This reading of Hume has become so common that it is now known as “Hume’s
Law.” But I have already argued in §II.8 that “Hume’s Law” is by no means Hume’s.
Hume never claims that a moral ought cannot be deduced from an is.
In the is-ought paragraph, Hume was concerned to refute the moral rationalists’
contention that moral ought-claims make reference to some peculiar moral relation.
Hume’s view, I have argued, is that (i) a correct analysis of our moral ought-judgments
must ultimately appeal to the moral sentiments and not merely to relations of ideas, and
that (ii) our moral ought-judgments, as such, can indeed be inferred from what can be
construed as non-moral facts, such as facts about what we would feel under certain
conditions, or facts about which acts or traits are useful or agreeable.
However, the proffered account has also been questioned on slightly different
grounds. As mentioned in a footnote in §II.8, Cohon argues that “an account of the
virtues which defines them as qualities that would provoke certain sentiments in people
who were in certain circumstances…is no better able to explain moral motivation [than

147
As discussed in §II.8, Cohon (1997) gives a closely related argument against this account. She appeals
to other passages in the Treatise that suggest that, in Hume’s view, we cannot infer moral facts. In
response, I have argued for a somewhat novel interpretation of the passages she cites, according to which
Hume suggests no such thing.

153
the moral rationalists’ account]” (1997: 835). And she points out that one of Hume’s
main objections to the moral rationalists’ account is that it cannot explain moral
motivation. Although the proffered account is not strictly speaking a definition of moral
qualities, the charge that the account is unable to adequately explain moral motivation is
a pressing one that has also been echoed by others.
For instance, according to Elizabeth Radcliffe, the challenge is that “any plausible
interpretation of Hume, if it is to explain how people experience what counts as moral
motivation, must explain this: How do people who generally do not experience the exact
feelings upon which their moral assessments are based frequently find that those
assessments furnish motives, when motives are feelings?” (1994: 49, her emphasis).
Radcliffe agrees with Cohon insofar as she does not think that the sort of account
proposed by this project can answer the interpretive challenge. They both think that the
proffered interpretation of Hume cannot account for moral motivation in those instances
in which, according to that interpretation, a moral judgment is issued and no occurrent
moral sentiment is produced. Both Radcliffe and Cohon conclude that this interpretation
must be mistaken. Consequently, they offer a different account of Hume’s view.
According to Radcliffe, Hume holds that we experience a sentiment of approval
or (as the case may be) disapproval each and every time we issue a moral judgment.
Thus, she attempts to answer the interpretive challenge by appealing to those sentiments
that we actually experience. Although Radcliffe claims that these feelings are not exactly
the feelings one would have if one were to satisfy all of the conditions of the moral view,
she suggests that they are typically sufficient to motivate us to act morally. Radcliffe

154
writes: “I approximate to the proper feeling, even if I do not actually experience it; on
this proposal, the approximation is sufficient to render the correct judgment, and the
degree of approval or disapproval I in fact feel is sufficient to provide the purported
motivation” (1994: 51-2).
Cohon (1997) also offers a similar answer to the interpretive challenge. She too
claims that we must experience either a sentiment of approval or disapproval every time
we issue a moral judgment.
148
In Cohon’s view, Hume has a “two-sentiment” account,
according to which “we feel certain passions from our particular vantage point, and
whenever we contemplate the same character from the common point of view we feel
another, weaker sentiment [viz., a moral sentiment]. That is, we feel two sentiments
toward the same character trait” (1997: 836). She goes on to suggest that we can use the
weaker moral sentiment to “correct or adjust” the stronger one, making it “exactly like”
the calmer one (ibid). When successful, these two sentiments “merge” together to
produce moral motivation. Sometimes, however, the passions are stubborn and resist our
attempted corrections and adjustments. Unfortunately, Cohon neither explains how, on
Hume’s account, the calmer moral sentiment corrects or adjusts our non-moral passionate
feelings, nor how these adjustments transform them into moral feelings.
Capaldi (1989), on the other hand, thinks that the proffered interpretation of
Hume, or something similar to it, is correct. According to Hume’s Treatise, we can
indeed issue a moral judgment even when no moral sentiment is produced. However,
Capaldi argues that this creates an “irresolvable” problem of moral motivation for

148
Baier (1991) and Sayre-McCord (1994) also hold similar views.

155
Hume’s account of sympathy. He writes: “That the sympathy mechanism is incapable of
accounting for the presence of a moral sentiment [in every instance a moral judgment is
issued] is, by far, the most serious shortcoming in Hume’s moral theory” (ibid, 231).
Capaldi goes on to argue that the only way for Hume to resolve this shortcoming
is “the postulation of a broader kind of benevolence” (232) that can produce an occurrent
moral sentiment each and every time one issues a moral judgment. Only then, says
Capaldi, will Hume be able to account for the production of a moral sentiment and
thereby explain moral motivation. He suggests that Hume ultimately comes to the same
conclusion, and so Hume consequently makes the necessary changes when he writes An
Enquiry Concerning the Principles of Morals. Capaldi takes various textual differences
in the Enquiry to be “a covert admission of failure on Hume’s part” (ibid) concerning his
account of sympathy given in the Treatise. According to Capaldi, “the final result is that
the sentiment of humanity”, which Capaldi sees as a natural disinterested (i.e., extensive)
benevolence, “replaces sympathy as the general principles of morals” (237).
In the following section, I will briefly respond to these latter claims made by
Capaldi. But for now, the important point is that Capaldi agrees with Cohon and
Radcliffe that moral motivation is an irresolvable problem for the proffered
interpretation. Of course, unlike Radcliffe and Cohon, he thinks that that interpretation,
or something very similar to it, is the correct account of Hume’s position in the Treatise.
Regardless, according to all three scholars, the only way in which Hume can avoid this
problem of moral motivation is to hold that each and every time one issues a moral
judgment one must also experience a corresponding moral sentiment.

156
According to Radcliffe, the approximating sentiment, which is experienced every
time we issue a moral judgment, is a moral sentiment. However, it is not the moral
sentiment that she contends is referred to by our moral judgments. Rather, as she
interprets Hume, our moral judgments refer to the moral sentiments approximated by the
sentiments we actually feel. As Cohon interprets Hume, he holds that moral judgments
either refer to or express—she remains non-committal about which option is Hume’s—
the felt moral sentiment generated by the two-sentiment account. Finally, according to
Capaldi’s interpretation, in the Enquiry Hume holds that moral judgments refer to the
corresponding moral sentiments engendered by the sentiment of humanity. On the other
hand, as Capaldi understands the view set out in the Treatise, Hume must hold that, in
those instances in which no moral sentiment is produced, one’s moral judgments refer to
moral ideas rather than moral sentiments.
149
According to Capaldi, Hume is forced to
hold this view because, in such instances, the speaker fails to have an occurrent moral
sentiment to which she can refer.
150

149
See, for instance, Capaldi (1989: 230).

150
Oddly enough, at one point Capaldi admits that Hume’s account does not always necessitate the
presence of a confirming impression, and hence a judgment need not refer to that impression. He writes:
…[E]very judgment contains an idea which refers in a general way to an impression.
When we make the judgment, the impression need not necessarily be present. For
example, I might describe the color of a shirt which I am not now wearing but which is at
home in a closet. The impression of color is understood although one need not have the
impression at the present in order to comprehend my meaning. The same holds true for a
moral judgment. (1989: 127)
It is therefore far from clear why Capaldi thinks that moral judgments must, in certain cases, refer to moral
ideas rather than referring “in a general way” to moral impressions, or why he claims that “all judgments,
including moral judgments, rely upon a present impression, and we cannot have a confirmable judgment
without an impression or with a contrary impression” (1989: 228). I suggest that Capaldi is confused, in
part, because he does not fully appreciate Hume’s distinction between moral judgment and moral
sentiment, and he consequently overlooks the distinction between our ideas of moral impressions and ideas
of impression-dependent moral properties. (See chapter II.) But part of his confusion is, I suggest, also

157
One significant problem for Hume holding this view is that he explicitly denies
that ideas alone can motivate us to act morally. A problem of moral motivation arises
precisely because a moral sentiment is requisite to produce moral motivation. For
instance, Hume writes:
The sentiments of others can never affect us, but by becoming, in some
sense our own; in which case they operate upon us, by opposing and
increasing our passions, in the very same manner, as if they had been
originally deriv’d from our own temper and disposition. While they
remain conceal’d in the minds of others, they can never have any
influence upon us: And even when they are known, if they went no further
than the imagination, or conception; that faculty is so accustom’d to
objects of every different kind, that a mere idea, tho’ contrary to our
sentiments and inclinations, wou’d never alone be able to affect us. (T
593)

I have already argued, in effect, that all of the interpretations given by the authors
above are mistaken. First, our moral judgments neither refer to our moral sentiments nor
to our moral ideas. As argued in chapter II, Hume’s view is that moral judgments contain
complex moral ideas that refer to moral properties, which are impression-dependent
properties of objects that (under the right conditions) “excite”, or cause, moral
sentiments. To recapitulate, an object has a moral property if and only if it is such as to
cause a sentiment of approval, or (as the case may be) disapproval in anyone that
observes or contemplates a given object while satisfying all of the regulative principles
that comprise the moral point of view.
Second, given this account of moral properties, our moral judgments do not
always necessitate the presence of a confirming moral sentiment or even a moral

due to his failure to see how Hume can resolve the problem of motivation in those cases in which one
issues a moral judgment without an occurrent moral impression. (See the following discussion.)

158
sentiment that approximates a confirming sentiment. A moral judgment necessitates the
presence of a confirming moral sentiment only when the judge satisfies all of the moral
view’s standardizing conditions. But there is no guarantee that the judge will always
satisfy those conditions. As discussed, Hume observes that our psychological
constitution is such that it can be rather difficult in some instances to extensively
sympathize with others—e.g., cases in which we are to extensively sympathize our
enemies or perfect strangers.
According to Hume’s moral theory, at least as I understand it, the possibility that
we can issue a moral judgment and yet we fail to have a confirming moral sentiment, or a
sentiment that approximates it, cannot be negated. Hence, as I read Hume, we can
indeed have a confirmable moral judgment without an occurrent confirming moral
sentiment or an occurrent sentiment that approximates the confirming sentiment (or even
with an occurrent sentiment that is contrary to a confirming moral sentiment). What we
cannot have is the direct confirmation of that judgment without a corresponding
occurrent moral sentiment.
151
Not only is this perfectly consistent with Hume’s account,
it is in fact entailed by it.
Consider once again an analogy with color. Suppose I see a red cup on the table
in front of me. And then all of a sudden the lights go out. Even supposing that there is
pitch-blackness and I therefore have no occurrent color impressions at all, I can still
judge that the cup in front of me is red. Based upon prior experience, I can know (or at

151
I say “direct confirmation” because there are, of course, indirect ways to confirm a moral judgment,
such as, for instance, appealing to past experience or the sentiments of those whom we believe satisfy the
conditions of the moral view while observing or contemplating the object under evaluation.

159
least believe) that I would have the relevant color impression were I to look at the cup
under standard conditions for viewing objects’ color. Similarly, given the impression-
dependent nature of moral properties, I can issue a moral judgment even in cases in
which I fail to have a moral sentiment. Based upon prior experiences, I can know (or at
least form a belief/judgment about) what moral feeling I would experience were I to meet
the conditions that comprise the moral view.
I suggest that this is the primary reason why Hume draws analogies with distance
perception and aesthetic beauty in stating his reply to the first objection against his
account of sympathy as the non-moral basis of morality. Both analogies indicate inter
alia that judgments, such as judgments of beauty, size, color, morality, etc., do not
necessitate a corresponding occurrent sentiment. For example, I can judge that grass in
green in the pitch black of night, or when I am not currently looking at green grass.
According to Capaldi, however, Hume’s analogies fail to accomplish their
intended purpose because they “still preserve some reference to a present perception,
however distorted, whereas in the case [at issue] we actually face the possibility of no
sentiment or perception at all” (224). Although strictly speaking Hume’s analogies do
preserve reference to an occurrent perception, it is clear that the point of the analogies is
to demonstrate how standard conditions stabilize our judgments and ensure uniformity. It
is not difficult to see how one can easily extend these analogies to cover those cases in
which there is no corresponding occurrent perception, as I have done in my examples
above. In fact, while engaging in a similar discussion within the Enquiry, Hume once
again invokes the analogy with distance perception when he writes: “Virtue, placed at a

160
distance, is like a fixed star, which, though to the eye of reason it may appear as luminous
as the sun in his meridian, is so infinitely removed as to affect the senses neither with
light nor heat” (E 230fn, emphasis added). Significantly, in this instance, Hume clearly
does extend the analogy with distance perception to the point at which the relevant
sensations are absent.
In sum, once standard conditions are put into place, there will always be the
possibility that one fails to meet those conditions. Regardless, one can still issue a
judgment that ascribes the relevant property. According to Hume’s account, one can
ascribe a moral property to an object even if one fails to have the relevant moral
impression as long as: (i) one is aware that one fails to meet all of the conditions of the
moral view, and (ii) one believes that one would experience the impression that confirms
the judgment were one to meet those conditions.
152

Neither Cohon, nor Radcliffe, nor Capaldi’s interpretation can adequately explain
how one can, and presumably sometimes does, issue moral judgments without
experiencing a corresponding moral sentiment. For instance, they cannot account for the
apparent fact that the amoralist can issue true moral judgments without experiencing a
moral sentiment.
153
Instead they attempt to deny this possibility. They appear to do so
primarily on the grounds that the proffered account of Hume cannot explain how it is
possible to produce moral motivation in those cases in which we fail to experience an

152
Making this determination is therefore crucial in making correct moral assessments. Hence having a
fairly good grasp of the requisite conditions is extremely important. For this reason, the following two
chapters in which the standardizing conditions are set out may be the most important chapters for this
Humean project.

153
Recall that the amoralist was briefly discussed in chapter II; see also part two of this project for further
discussion.

161
occurrent moral sentiment when a moral judgment is issued. Cohon and Radcliffe go so
far as to conclude that this therefore must not be Hume’s account. So the objection based
on moral motivation appears to be quite serious for the proposed interpretation.
§V.2 The solution to the objection
Unfortunately, Hume does not explicitly answer the objection against (what I
have argued is) his account of sympathy as the non-moral basis of morality. He is not
entirely explicit about how it is possible, on his view, for extensive sympathy to generate
a moral sentiment in any and every case. For instance, I have argued that, according to
Hume’s view, one can infer moral facts from various non-moral facts. Yet Hume does
not explicitly explain how it is possible for someone to be motivated by a moral judgment
that he or she accepts by inference from experience. This is unfortunate because if Hume
had been more explicit about his answer, then the objection would not be so serious for
the proffered interpretation since the answer would be more obvious. I suggest that,
consequently, that interpretation would be more widely accepted.
In any case, although Hume does not explicitly address this problem of moral
motivation, fortunately there are a few passages that point to its solution. I shall now
argue that Hume provides the materials, as well as an example that illustrates how he can
use those materials, to resolve the problem by invoking his account of sympathy and
thereby answer the objection. Thanks to Hume’s precedent, it turns out that the solution
is rather straightforward. It is analogous to the answer Hume offers to the question
concerning our moral approval of justice.

162
As noted the previous chapter, Hume argues that individuals initially chose to
construct and submit to societal regulations due to the non-moral motives of self-
interestedness and limited benevolence. Such rules are necessary for a stable society, and
a stable society satisfies both our self-interest and limited benevolence by offering plenty
of benefits both to oneself and to one’s family and friends that could not be had
otherwise.
154
Upon reflecting on the societal benefits of the system of justice, feelings of
moral approval are engendered toward that system by way of extensive sympathy. In this
way, concludes Hume, our sympathy with the public interest is what accounts for our
sense of duty towards justice, which arises only after that system has been put in place.
Thus, by producing moral feelings of approval toward justice, our sympathy accounts for
moral motivation towards that system even in those instances in which we personally do
not benefit from—and may in fact be harmed by—the system of justice. Hume writes:
Nay when the injustice is so distant from us, as no way to affect our
interest, it still displeases us; because we consider it as prejudicial to
human society, and pernicious to every one that approaches the person
guilty of it. We partake of their uneasiness by sympathy… this is the
reason why the sense of moral good and evil follows upon justice and
injustice… Thus self-interest is the original motive to the establishment of
justice: but a sympathy with public interest is the source of the moral
approbation, which attends that virtue. (T 499-500, Hume’s emphasis)

My suggestion is that Hume can resolve the problem of moral motivation at issue
by making an analogous move with regard to our sense of duty toward morality itself. In
those instances in which no moral sentiment is produced due to previously discussed
features of human psychology—and in fact even if a conflicting sentiment is produced—
a moral sentiment can nevertheless subsequently be generated through sympathy by

154
See T 485ff.

163
reflecting on, and sympathetically responding to, the fact that morality has important
benefits for society and humanity. Because Hume grounds the notion of duty or
obligation within sympathy, recognition of one’s moral duty can produce a moral motive.
Our sense of moral duty can therefore serve as a moral motive even in cases in which no
moral sentiment is initially produced in conjunction with our moral judgment.
For instance, suppose I come to accept the moral judgment that it is morally
wrong to torture my enemies even if I hate them. Although my hatred toward them
might, in some instances, prevent me from having an occurrent (confirming) moral
sentiment of disapproval toward torturing my enemies, I can still recognize that I have
failed to have that sentiment because I have not satisfied the conditions imposed by the
moral view. I have failed to disinterestedly—i.e., extensively—sympathize with them.
Supposing this is a fact that I do recognize, I would presumably still believe the moral
judgment—viz., that it is morally wrong to torture my enemies even if hate them—even
though I have no occurrent (confirming) sentiment of moral disapproval. In short, I
believe the judgment but lack the moral motivation to act upon it. Consequently, ceteris
paribus, it is impossible for me, in Hume’s view, to act morally in this case.
However, a moral sentiment can nonetheless be engendered by sympathy if I
recognize that I have a moral duty not to perform acts that are morally wrong. That is,
based upon my observations overtime, I may come to realize—and am likely to have
already been socialized to accept—that moral practice is beneficial and advantageous to
humankind, society, my loved one’s, and myself. Upon reflecting on the positive general
effects of morality, by way of the sympathy process subject to the regulatory principles of

164
the moral view, moral feelings of approval toward the practice of morality (and
disapproval of neglecting it) will be engendered.
155
I will thereby have at least some
motivation to obey its dictates and some motivation against disobeying.
156
I have, in
effect, internalized the subject matter of morality. Or as Hume puts it:
According to the latter system [viz., Hume’s moral system], not only must
virtue be approv’d of, but also the sense of virtue: And not only that sense,
but also the principles, from which it is deriv’d. So that nothing is
presented on any side, but what is laudable and good. (T 619)

It might be objected that this account of moral duty implausibly requires a focus
on the consequences, or effects, of morality. In reply to this worry, I offer the following
two points. First, once the connection between sympathy and causal reasoning is better
appreciated, Hume’s focus on emotional effects will seem much more plausible than it
might initially appear. (This connection and the role of utility in Hume’s theory will be
the main topic of discussion in the proceeding sections.) For instance, consideration of
the effects of morality can include things such as the violation of the respect for persons.
Second, Hume’s account of moral duty given above, which clearly does focus on
consequences, is an account of the origination of our sense of moral duty (as well as our
sense of duty towards justice). Failure to act morally engenders negative feelings in us
(at least under certain conditions). According to Hume, these feelings are produced by an

155
In the Enquiry, Hume refers to our attitudinal responses to the general effects of an object upon
humanity as the “sentiment of humanity”. Thus, the sentiment of humanity (i.e., extensive benevolence)
plays a key role in accounting for the moral motivation we experience due to our moral duty. Our moral
duty is of course the result of the psychological process of sympathy. Hence the “sentiment of humanity”
(like extensive benevolence) is a byproduct of sympathy. (See the following section for further discussion.)

156
However, in these cases, we may be less likely to be motivated to act on such considerations since there
may be other conflicting feelings that arise due to sympathy’s partiality and indifference. See the following
two chapters for further discussion of this matter.

165
extensive sympathy with the negative effects that tend to be caused by failing to act
morally. But Hume also observes, and indeed stresses, that we are creatures of habit.
Hence our sympathetic feelings can be affected by custom and the force of habit. Our
sense of moral duty can therefore be greatly impacted by our social upbringing. We are
likely to be taught and habituated to apply and follow various general rules about how to
conduct ourselves, especially if they tend to produce beneficial effects. Through such
teaching and social habituation, our sense of (moral) duty can be regulated and refined.
For instance, we can be taught and habituated to act in accordance with our moral
judgments.
157
Consequently, our negative feelings that result from failing to do what we
believe morality requires of us may, in fact, tend not to be engendered from conscious
deliberation about the typical negative effects that result from failing to act in accordance
with our moral judgments. Rather, these feelings may tend to be due to applying and
following rules that result from our social upbringing and habituation, which is why focus
on consequences may seem prima facie implausible. Moreover, this helps explain why
we also tend to presume that those who sincerely issue moral judgments also have a sense
of moral duty. We consequently not only assume that people’s moral judgments will be
expressive of their actual feelings (as claimed in §II.5); in virtue of those feelings, we
expect people to be motivated to act in accordance with the moral judgments they accept.
In sum, thanks in large part to the tendency of morality to benefit society and
humanity, sympathy is indeed able to account for the production of a moral sentiment in
all of the problematic cases. According to Hume’s account, sympathy is responsible for

157
Of course, as discussed in §IV.2, the possibility of being taught and habituated in this way depends upon
our already possessing a capacity for sympathy.

166
generating our sense of moral duty, which in turn explains how a moral motive is
produced even when: (i) we issue a moral judgment but the conjoining occurrent moral
sentiment is too weak to move us to action (because our extensive sympathy is weak);
and (ii) we issue a moral judgment but there is no conjoining occurrent moral sentiment,
such as when we infer a moral fact from various non-moral facts. Of course, there is still
no guarantee that we will be motivated to act in accordance with our moral judgments,
but Hume’s account does demonstrate that it is at least possible to be morally motivated
in these problematic cases. Hence the objection to (what I have argued is) Hume’s
account of sympathy fails.
As mentioned, Hume does not explicitly tackle this objection. However, there is
fairly persuasive textual evidence in support of my contention that Hume would respond
to that objection in the manner that I suggest. This is significant because, apart from the
fact that what is arguably one of the strongest objections against the proffered
interpretation of Hume fails, the following passage serves to further bolster this
interpretation. Hume writes:
But may not the sense of morality or duty produce an action without any
other motive? I answer, It may: But this is no objection to the present
doctrine. When any virtuous motive or principle is common in human
nature, a person, who feels his heart devoid of that principle, may hate
himself upon that account, and may perform the action without the motive,
from a certain sense of duty, in order to acquire by practice, that virtuous
principle, or at least, to disguise himself, as much as possible, his want of
it. A man that really feels no gratitude in his temper, is still pleas’d to
perform grateful actions, and thinks he has, by that means, fulfill’d his
duty… But tho’, on some occasions, a person may perform an action
merely out of regard to its moral obligation, yet still this supposes in
human nature some distinct principles, which are capable of producing the

167
action, and whose moral beauty renders the action meritorious. (T 479,
emphasis added)

This passage is significant for our purposes because it demonstrates two key
features in support of the proffered interpretation. First, the passage strongly suggests
that it is possible in Hume’s view to issue a moral judgment without experiencing a moral
motive. This is clearly suggested by the second key feature, to wit: Hume holds that
one’s sense of moral duty, in conjunction with sympathy, can generate a moral motive
when there initially isn’t one. In other words, one’s motivating sense of moral duty can
come after the moral judgment is already made.
This account can therefore answer Cohon’s (1997: 835) objection that the
proffered interpretation is not able to fully explain moral motivation. And it also serves
as a solution to Elizabeth Radcliffe’s interpretive challenge, namely, “to reconcile the
claim that (a) our moral distinctions are directly based on feelings not [necessarily]
identical to the feelings we in fact experience, with the claim that (b) the motivating force
of morality comes from the internal nature of its grounds” (1994: 47).
It is noteworthy that my interpretation agrees with Cohon’s proposal insofar as
there can be two conflicting sentiments, namely, a non-moral sentiment that results from
one’s own particular vantage point and a moral sentiment that results from taking up the
common moral point of view. However, the production of the moral sentiment, on the
proffered solution, does not always result from one’s extensive sympathy with the
putative typical effects of the character trait, as Cohon suggests. Sometimes it is the
product of one’s extensive sympathy with the effects of acting, and failing to act, in

168
accordance with that which morality requires; i.e., the felt moral sentiment is, in effect, a
second-order motivation that results from a sense of duty to morality itself.
Moreover, the proffered interpretation also agrees with Radcliffe’s proposal
insofar as one can experience a moral sentiment that sometimes is an approximation of
the moral sentiment that one would feel were one to satisfy all of the conditions that
comprise the moral view. In other words, there are cases in which the felt sentiment will
be an approximation of the moral sentiment that determines the moral property ascribed
by the moral judgment one issues—viz., the moral sentiment that one would feel were
one to sufficiently regulate one’s sympathy according to the conditions of the moral view.
Finally, this interpretation answers Capaldi’s charge that: “The major defect in the
sympathy mechanism is that it is unable to account in every case for the production of a
moral sentiment which is necessary for the confirmation of a moral judgment. The irony
is that sympathy was inaugurated precisely to account for the moral sentiment” (226, his
emphasis). As discussed, Hume’s view only necessitates that sympathy produce an
occurrent moral sentiment that confirms one’s moral judgment in those instances in
which one satisfies all of the conditions that comprise the moral view. Hence the issue is
not the production of an occurrent confirming moral sentiment. Rather, the key issue is
whether a moral sentiment can somehow be produced even when standard conditions are
not met so that moral motivation is possible in every case. As we have just seen,
sympathy is indeed able to account for the production of a moral sentiment in such cases.
Hence moral motivation is at least possible in such cases.

169
§V.3 A reply to Capaldi
At this point, I would like to quickly respond to some of the textual evidence that
Capaldi offers in support of his claims that (i) Hume covertly admits in the Treatise and
the Enquiry that there are problems with his account of sympathy, and (ii) in the Enquiry,
Hume replaces sympathy with “the sentiment of humanity” as the general principle of
morals. Regarding Hume’s putative covert admissions in the Treatise, Capaldi writes:
When Hume initially raised the [first] objection to [his] account of
sympathy his immediate answer was that the objection “must have equal
force against every other system, as against that of sympathy” (Treatise, p.
581). Nevertheless, Hume… passed over his immediate reply and went on
to introduce his reference to general rules. When he came to the
conclusion of Book III…he once again found himself forced to defend his
view by falling back on the immediate answer. “If we compare all of
these circumstances, we shall not doubt, that sympathy is the chief source
of moral distinctions; especially when we reflect, that no objection can be
rais’d against this hypothesis in one case, which will not extend to all
cases” (Treatise, p. 618). All of this seems to indicate that Hume was
aware of a difficulty or some difficulty in his [account] of sympathy. (223)

As I read Hume, those passages indicate no such thing. After stating that the
objection “must have equal force against every other system, as against that of
sympathy”, Hume immediately writes: “But to consider the case a-right, it has no force at
all; and ‘tis the easiest matter in the world to account for it.” This hardly suggests that
Hume suspected there is a problem. As for the second passage that Capaldi cites above,
it is quite clear that Hume takes himself to have already unequivocally demonstrated that
sympathy is “the chief source of moral distinctions” in many cases. Since “no objection
can be rais’d against this hypothesis in one case, which will not extend to all cases”, there
should be no doubt that the objection does not hold water in any case. This too hardly

170
indicates that Hume was aware of some problem with his account of sympathy. To the
contrary, it seems to suggest that Hume was quite confident in that account.
Regarding Hume’s putative covert admissions in the Enquiry, Capaldi writes:

In making humanity the general principle of morals, Hume said the
following: “everything which contributes to the happiness of society,
recommends itself directly to our approbation and good will. Here is a
principle which accounts in great part for the origin of morality: And what
need we seek for abstruse and remote systems, when there occurs one so
obvious and natural” (EPM, p. 219). What Hume means is explained in a
footnote: “It is needless to push our researches so far as to ask why we
have humanity or a fellow-feeling with others” (Ibid.). In short, there will
be no attempt to explain our humanity by generating it from the sympathy
mechanism. (239-40)

I find it quite puzzling that Capaldi takes this passage and footnote to show that
Hume does not appeal to sympathy as a general principle of morals since “a fellow-
feeling with others” is the direct result of the sympathy process—indeed Hume explicitly
discusses sympathy for the rest of the section. Hence the footnote simply indicates that
Hume is not going to discuss the origins of sympathy itself. Furthermore, the passage
that Capaldi claims “leaves us in no doubt” that Hume does not attempt to explain the
sentiment of humanity by generating it from sympathy is the following:
It is but a weak subterfuge, when pressed by these facts and arguments, to
say that we transport ourselves by the force of imagination, into distant
ages and countries and consider the advantage, which we should have
reaped from these characters, had we been contemporaries; and had any
commerce with the persons. It is not conceivable, how a real sentiment or
passion can ever arise from a known imaginary interest; especially when
our real interest is still kept in view, and is often acknowledged to be
entirely distinct from the imaginary, and even sometimes opposite it.
(EPM, 217; [64])

171
Capaldi claims: “The foregoing, of course, is a perfect description of the
sympathy mechanism in the Treatise” (ibid). Capaldi’s support for this claim comes
from his suggestion that “in speaking of the ‘force of imagination’ Hume is reminding us
of his earlier description of the sympathy mechanism in the Treatise [(T 576)]” (ibid). In
other words, Capaldi interprets this passage as an explicit rejection on Hume’s part of any
attempt to explain the sentiment of humanity by generating it from sympathy.
Once again, however, I disagree with Capaldi’s interpretation of the text. The
section of the Enquiry from which the above passage is taken—viz., section V—is titled
“Why Utility Pleases”. In this section, among other things, Hume addresses the
contention that utility and the moral virtues only please us insofar as they are products of
“self-love”, i.e., self-interestedness. Not only does the above passage occur within that
discussion, I suggest it addresses that very issue. For instance, in the key passage above,
when Hume speaks of “the force of our imagination”, he states that we use it to “consider
the advantage, which we should have reaped from these characters” (emphasis added).
This, I suggest, clearly indicates that Hume is not discussing the force of imagination as it
pertains to sympathy—i.e., a fellow-feeling with others—but rather as it pertains to
ourselves.
In short, the “imaginary interest” of which the passage speaks is not a shared
interest of others but instead our own interest. Hence Hume’s point in this passage is that
no real moral sentiment can be produced from an imaginary self-interest, especially when
our real self-interest conflicts with the imaginary one. Thus, in opposition to Capaldi’s
claim, the passage is a critique of self-interest as the source of moral distinctions rather

172
than a critique of sympathy as the source.
158
To the contrary, Hume’s overall point seems
to be that sympathy, not self-interest, must be the source. If so, then this clearly echoes
the claims Hume makes in the Treatise.
Indeed, contrary to Capaldi’s claim, I suggest that Hume offers the same account
of sympathy in the Enquiry as he set forth in the Treatise. He does not replace sympathy
as the general principle of morals and posit the sentiment of humanity as a natural moral
motive, as Capaldi suggests. Capaldi is not alone in reading Hume as tacitly renouncing
his former claim that there are no natural moral motives. Alasdair MacIntyre (1965: 15),
for instance, argues that the Treatise’s theory of psychology “is both more interesting and
more complex than that of the Enquiry.” He continues, stating:
This can be bought out by comparing the following two quotations: “In
general, it may be affirm’d, that there is no such passion in human minds,
as the love of mankind, merely as such, independent of personal qualities,
of services, or of relation to ourself.” (Treatise III, ii, 1.) “It appears also,
that, in our general approbation of characters and manners, the useful
tendency of the social interests moves us not by any regards to self-
interest, but has an influence much more universal and extensive. It
appears that a tendency to public good, and to the promoting of peace,
harmony, and order in society does always, by affecting the benevolent
principles of our frame, engage us on the side of the social virtues.”
(Enquiry, V, ii.) That is, in the Enquiry, a native disposition to sympathy
and benevolence is the effective basis of moral judgment and action.

158
Further textual evidence for this interpretation can be found throughout the Enquiry. For instance, in the
following section of the Enquiry, Hume reiterates roughly the same point:
Now as these advantages are enjoyed by the person possessed of the character, it
can never be self-love which renders the prospect of them agreeable to us, the spectators,
and prompts our esteem and approbation. No force of imagination can convert us into
another person, and make us fancy, that we, being that person, reap benefit from those
valuable qualities, which belong to him. Or if it did, no celerity of imagination could
immediately transport us back, into ourselves, and make us love and esteem the person,
as different from us…All suspicion, therefore, of selfish regards, is here totally excluded.
It is a quite different principle [viz., extensive sympathy or benevolence], which actuates
our bosom, and interests us in the felicity of the person whom we contemplate.” ([78-9],
Hume’s emphasis)

173
While in the Treatise there is a problem as to how, granted that by nature
we are not disposed to the public good, we nonetheless have developed a
morality in which the public good is given priority over the private. (Ibid)

Crucial to understanding why these passages are not in conflict as MacIntyre
suggests is the previously discussed distinction between limited and extensive sympathy
and the corresponding distinction between limited and extensive benevolence. Hume
denies, in both the Treatise—as, for example, in the passage above—and the Enquiry,
that there is a natural disposition for extensive sympathy and extensive benevolence. But
he does hold in both the Treatise and the Enquiry that there is a natural inclination for
limited sympathy and limited benevolence. In both works, sympathy and benevolence are
the effective basis of moral judgment and moral motivation.
We have already discussed in the previous chapter how extensive sympathy and
extensive benevolence originate from limited sympathy and limited benevolence, as well
as self-interestedness. So, when Hume speaks of “the benevolent principles of our
frame” in the passage above, he is talking about our innate capacities for limited
sympathy and limited benevolence, which when properly regulated by the principles of
the moral view produce extensive sympathy and the moral motive of extensive
benevolence (i.e., “the sentiment of humanity”) respectively. Hence there is no tension
between these passages.
I have argued that there is no need for Hume to change the account he offers in
the Treatise because, according to that account, sympathy can produce a moral sentiment
each and every time we issue a moral judgment. But if I am right and Hume offers the
same theory of psychology and the same account of sympathy in the Enquiry as he does

174
in the Treatise, then why does he, in the Enquiry, speak of the sentiment of humanity a
general principle of morals? In answer to this question, I agree with Kate Abramson
(2001), who argues that this discrepancy between the Enquiry and the Treatise is a matter
of marketing rather than one of substance.
159

As I read Hume, he conceives of the sentiment of humanity as equivalent to
extensive benevolence. Recall, for instance, Hume’s claim that: “The former sentiment,
to wit, that of general benevolence, or humanity, or sympathy, we shall have occasion
frequently to treat of in the course of this enquiry” (EPM, Appendix II, fn. 1).
160
The
difference in marketing between the Treatise and the Enquiry is manifested, at least in
part, as a difference of emphasis. For example, when Hume discusses, in the Enquiry, the
sentiment of humanity, he tends to conceive of it as a response to our evaluations of
general utility or (as they case may be) general disutility. That is, the sentiment of
humanity is a feeling of approval toward that which we judge to be useful to ourselves
and/or others, or (as the case may be) a feeling of disapproval toward that which we
judge to be pernicious to ourselves and/or others. This is equivalent to extensive
benevolence since the result is the same, namely, a general concern for others. Moreover,
in both cases, that general concern for others is a product of our natural sympathy
subjected to the regulatory principles that comprise the moral view. Hume writes:

159
See also Rico Vitz (2004) who holds a similar view.

160
It may initially seem somewhat odd that Hume here treats (extensive) sympathy as a sentiment rather
than a psychological process, and equates it with (extensive) benevolence. But recall from §IV.3 the
ambiguity in Hume’s use of the term ‘sympathy’ in which he sometimes uses it to refer to the sentiments
produced by sympathy, which of course include benevolence.

175
[T]here must…be a choice or distinction between what is useful, and what
is pernicious. Now this distinction is the same in all its parts, with the
moral distinction, whose foundation has been so often, and so much in
vain, enquired after. The same endowments of the mind, in every
circumstance, are agreeable to the sentiment of morals and to that of
humanity; the same temper is susceptible of high degrees of the one
sentiment and of the other; and the same alteration in the objects, by their
nearer approach or by connexions, enlivens the one and the other… [I]n
each particular, even the most minute, they are governed by the same
laws, and are moved by the same objects. (EPM, [80])

According to Hume, the sentiment of humanity (or extensive benevolence) and
the moral sentiments are equivalent sentiments because “they are governed by the same
laws”. In other words, they share the same basic source or origin, namely, the capacity
for sympathy, which explains why they are moved by the same objects.
161
Thus,
sympathy remains the general principle of morals in the Enquiry. It is a necessary
precondition for the production of any moral motive, including the sentiment of
humanity/extensive benevolence.
162

While Hume’s change of focus, or emphasis, in the Enquiry does not constitute a
change in substance from the Treatise, it does turn out to be a rather significant change

161
However, there might be a difference of scope or generality between extensive benevolence and the
sentiment of humanity. Extensive benevolence can respond to the particular pleasures and pains of
particular individuals. But it isn’t clear whether Hume conceives of the sentiment of humanity in this way.
That is, the sentiment of humanity might only respond to people’s pleasures and pains conceived more
generally. It would, in effect, treat all of humanity as an undifferentiated object of pleasure and pain. After
all, the sentiment of humanity just is the feeling of approval toward that which has social utility. Such
generality is important for Hume’s theory because it enables it to account for moral motivation in the case
of artificial virtues, such as justice. For instance, we morally approve of the blindness of justice. But this
approval of its blindness cannot come from a benevolent concern for a particular person. Instead it must
come from an approval of the beneficial affects that justice’s blindness has upon humanity more generally.
Of course, extensive benevolence also includes this more general concern. So perhaps Hume conceives of
the sentiment of humanity as a kind of general extensive benevolence.

162
It is noteworthy that Hume also refers to the “sentiments of humanity” in the Treatise—see, for instance,
p. 518. And in the Treatise, it is clear that such a sentiment of humanity is ultimately dependent upon the
capacity for sympathy.

176
for the purposes of this project. As we shall see (§V.6), one important effect of including
the sentiment of humanity as a general principle of morals is that it serves to emphasize,
as vital, the intimate connection between sympathy and our judgments of social utility.
This is something that Hume arguably neglects to do in the Treatise.
163
He at least
doesn’t emphasize that link in the Treatise to the degree that he does in the Enquiry.
This change of emphasis helps answer another objection to Hume’s account of
sympathy that we will now consider. But as we shall see, it also raises pivotal questions
about the connection between sympathy, causal reasoning, and utility. The proceeding
section addresses the objection and its reply. This will set the stage for the section that
follows it, which discusses the relationship of sympathy, causal reasoning, and utility.
§V.4 Mercer’s objection to Hume’s account of sympathy & a reply
Philip Mercer (1972: 69ff) objects to Hume’s account of sympathy on the grounds
that it is “value-neutral”; i.e., it isn’t responsive to judgments of value. He observes that
merely sympathizing with someone’s (past, present, or future) pleasurable or painful
feelings is not sufficient to generate moral feelings that are indicative of actual moral
distinctions. According to Mercer, in order to generate that sort of moral feeling, “we
must also know whether this pain or pleasure has any value beyond itself” (p. 70). This
determination of value involves an evaluative judgment that is “logically independent”
of—i.e., distinct from—one’s sympathetic feelings. Hence one cannot rely upon one’s

163
In the “Advertisement to the Collected Essays”, of which the Enquiry is one essay, Hume states that
“some negligences in his former reasoning [of the Treatise] and more in the expression are, he hopes,
corrected.” Perhaps emphasis on the crucial connection between sympathy and our judgments of utility is
one of the “negligences” of which he speaks. Regardless, Hume’s focus on the sentiment of humanity is in
no way incompatible with his views in the Treatise, or so I have argued.

177
sympathetic feelings alone. One must also rely upon a value judgment, i.e., a judgment
about people’s interests or ends. Mercer offers the following example:
Suppose that I see A stopping B from doing something which it is obvious
to me that B very much wants to do. I may sympathize with B’s
frustration and misery and this sympathy will cause me to disapprove of
A’s conduct. But on deeper investigation I may find that appearances are
deceptive and perhaps either that A did not realize he was causing B pain
or that he did realize but was aiming at B’s long-term happiness (say, the
case of a father withholding consent from what he thought to be an
unsuitable marriage) or that he did realize but was aiming at other people’s
benefit or even the long-term benefit of the community (say, in the case of
punishment). On being in full possession of the facts my disapproval of A
may turn to approval. If this did happen, then my approval of A would be
perfectly compatible with my immediate sympathy with B’s pain. Now it
should be clear from this that it is not just a case of sympathizing with the
feelings of the agent and/or those affected by his conduct together with the
adoption of a neutral point of view—evaluation must also involve the
recognition of the true interests of the people concerned. (Ibid)

As Mercer sees it, this sort of evaluation—viz., one independent of sympathetic
feelings—is incompatible with Hume’s account of sympathy. The example above
illustrates that simply sympathizing with the pleasure and/or pain of those affected by the
agent’s act or trait is not enough to produce the moral feelings that indicate actual moral
distinctions within a given set of circumstances. Moral appraisers also need to determine
whether their pain or pleasure has any “value beyond itself” (70). In other words, they
need to make a “utility judgment”, which in the present example is a judgment about the
tendency of A’s act or trait to fulfill the “true interests”, or ends, of those affected by A’s
act or trait. The problem, according to Mercer, is that we cannot sympathize with the
interests, or ends, of others, and hence we cannot sympathize with the overall value of
their painful or pleasurable feelings. We can only sympathize with others’ pleasure or

178
pain. There is, in effect, a critical disconnect between sympathy and our utility/value
judgments. As Mercer puts it, sympathy is “blind to value”. Hence sympathy cannot
serve as the non-moral basis of morality.
Mercer is led to his conclusion in part because he understands Hume’s account of
sympathy as a “non-cognitive”, “infectious fellow-feeling” account, according to which
the feelings of one person involuntarily infect the feelings of another by producing the
same feeling of the former person within the latter person. Such an account of sympathy
clearly does not allow a cognitive, value judgment to affect one’s sympathetic feelings;
much less does it allow us to sympathize with the value of others’ feelings.
164

164
It is noteworthy that Mercer also puzzles over how sympathy can generate a moral motive in accordance
with the moral judgments we issue. He writes:
It seems to me that the whole point of a spectator’s adoption of a general point of view
must be to arouse within him (through the medium of sympathy) the feelings of approval
and disapproval in accordance with which moral judgments are made. If he reflects as to
what his feelings would be if he were truly disinterested then the outcome of this
reflection has to be a judgement which commits him to having certain feelings (75-6)…
[However] the judgements which are made as results of such questions as ‘What feelings
would I have towards X if I were truly disinterested?’ and ‘To what extent is X of benefit
to [the agent] and/or the community in general?’ do not in fact commit the maker to any
particular feelings or actions. These judgements are theoretical rather than practical; that
is, although they may affect feelings and actions, they need not. That these judgments are
only theoretical is in contradiction, though, with what I said earlier…where I suggested
that the judgements which are implicit in the sympathy theory must be practical if the
theory is to be made viable. This contradiction can be seen as a reflection of the
intractability of Hume’s material. (80-1)
In reply, chapter II has, in effect, already argued that Mercer is mistaken in thinking that one must
always be motivated to act in accordance with one’s moral judgments. Moral motivation is necessitated
only when one meets the conditions of the moral view. Although moral judgments do not always commit
the issuer of those judgments to having moral feelings that motivate her to act according to those moral
judgments, they do always commit the issuer to the moral belief that she (morally) ought to have
motivating moral feelings to act accordingly. Mercer is correct that our moral judgments are not merely
theoretical. They are practical, i.e., action-guiding. As also discussed in chapter II, Hume can explain how
one’s sense of moral duty, which is produced by sympathy, can always generate a motivating moral feeling
to act according to one’s moral judgment, even in those cases in which one fails to have the moral feeling
that confirms the moral judgment. Thus, the putative contradiction Mercer claims to find in Hume’s
material rests upon a misinterpretation of those materials, or so I have argued.

179
In opposition to Mercer’s contention, I have argued that Hume does not merely
embrace a non-cognitive, involuntary infectious feeling account of sympathy. While
there is an involuntary, non-cognitive element to Hume’s account of sympathy, Hume
also clearly holds that we can sympathize with the feelings we believe others will or
would have based upon our causal beliefs about the tendency of certain acts, dispositions,
etc, to produce those feelings in others. Of course, we can be wrong about the tendencies
of objects to produce such feelings in those with whom we are sympathizing. But the
fact that our sympathetic feelings respond to, and are mediated by, such causal beliefs,
clearly undermines Mercer’s interpretation of sympathy as a purely non-cognitive
process. After all, in Hume’s view, sympathy is a psychological process that converts
our ideas into impressions. Hence there must be a cognitive component at work.
To recapitulate, Hume clearly emphasizes the role that reason plays in producing
beliefs that impact sympathy and thereby our moral feelings. For example, recall part of
the passage quoted at the very beginning of the project in which Hume claims that “in
order to pave the way for such a sentiment [viz., a moral sentiment], and give proper
discernment of its object, it is often necessary, we find, that much reasoning should
precede, that nice distinctions be made, just conclusions drawn, distant comparisons
formed, complicated relations examined, and general facts fixed and ascertained.” Hume
continues stating that sometimes “it is necessary to employ much reasoning, in order to
feel the proper sentiment; and a false relish may be corrected by argument and reflection”
(EPM, section I).

180
While Mercer is indeed correct that, strictly speaking, we cannot sympathize
directly with others’ interests or ends, Hume can and does allow for our judgments of
others’ interests or ends to affect our beliefs about what their feelings will or would tend
to be under a given set of circumstances. Indeed, this seems implicit in Hume’s claim
that the spectator’s causal beliefs about the causal tendency of objects to produce certain
outcomes will affect his or her sympathetic feelings: if the spectator did not take into
account at least some of the interests, or ends, of those affected by an agent’s act or trait,
then it would be impossible to assess what people would feel in many situations. In
short, although we cannot directly sympathize with others’ interests or ends, we can
sympathize with the typical emotional effects caused by the fulfilling or foiling of many
of their ends/interests.
Hume is quite explicit that sympathy and thus our moral feelings are responsive to
our assessment of others’ interests/ends. Reconsider, for example, the following passage:
…[E]very particular person’s pleasure and interest being different, ‘tis
impossible men cou’d ever agree in their sentiments and judgments, unless
they chose some common point of view, from which they might survey
their object, and which might cause it to appear the same to all of them.
Now, in judging of characters, the only interest or pleasure, which appears
the same to every spectator, is that of the person himself, whose character
is examin’d; or that of persons, who have a connexion with him. And tho’
such interests and pleasures touch us more faintly than our own, yet being
more constant and universal, they counter-balance the latter even in
practice, and are alone admitted in speculation as the standard of virtue
and morality. They alone produce that particular feeling or sentiment, on
which moral distinctions depend.
165
(T 591, emphasis added)

165
Mercer also cites this passage. It is noteworthy that he does not see a tension between his interpretation
of Hume’s account of sympathy as merely a kind of non-cognitive emotional infection and what Hume
actually writes. Instead, Mercer takes the tension to reside within Hume’s own position.

181
The reason I have called attention to Mercer’s objection is because, as Mercer
observes, it nicely illustrates that “the spectator’s comprehensive utility judgement is in
some sense capable of exerting an influence on his feelings of sympathy” (p. 72). In fact,
the connection is stronger than this. Mercer’s objection brings to light a vital link
between sympathy (and hence our moral feelings) and our utility/value judgments.
However this intimate link between sympathy and our utility/value judgments does not
pose the sort of problem for Hume’s account of sympathy that Mercer suggests.
Sympathy is not “value-neutral” because, in Hume’s view, our sympathetic
feelings are responsive to some of our judgments of value; in particular, it is responsive
to our judgments of moral value. In other words, as I read Hume, our moral feelings are
responsive to judgments about the causal tendency of objects (e.g., acts or traits) to fulfill
or foil particular ends and thereby produce either pleasurable or painful effects
respectively.
166
Recall that we’ve already discussed how extensive sympathy is
necessary for generating the moral motive of extensive benevolence. Mercer’s objection
brings to light the fact that our utility/value judgments are necessary in assessing the
tendency of an act, trait, etc. to cause painful or pleasurable feelings in those with whom
we extensively sympathize.
Mercer is by no means alone in overlooking or misunderstanding this connection
between causal reasoning and sympathy, and thus our moral feelings. For example,
Alasdair MacIntyre (1965: 16) claims that Hume overlooks or ignores the fact that

166
This, however, is not to say that these evaluative judgments directly cause our moral feelings. Our
judgments of moral value are an integral part of the causal chain, but strictly speaking it is the sympathy
process that converts the belief (or idea) into the relevant feeling. (See the following section.)

182
“feelings do not just happen… They can be modified, criticised, rejected, developed, and
so on.” MacIntyre takes this fact to illustrate the following point against Hume: “Feeling
and judgment cannot be identified apart; indeed this is why moral feeling is not
something in terms of which moral judgment can be elucidated” (ibid.).
It is true that, strictly speaking, Hume does not think that it makes sense to simply
modify, criticize, reject, or develop the feelings one has. Our feelings, in and of
themselves, aren’t the kind of thing that can be criticized, rejected, modified, etc. because
feelings are “original existences”; they simply occur. But MacIntyre, like Mercer,
overlooks the fact that Hume sees our moral feelings, via sympathy, as responsive to
many of our reasoned judgments. Significantly, Hume does think that we can modify,
criticize, reject, or develop the judgments, intentions, decisions, and reasoning that give
rise to those feelings. Since sympathy and thereby our moral feelings are responsive to
such judgments/reasoning, any occurrence of modification, criticism, rejection,
development, etc. of our reasoned judgments will subsequently impact these feelings.
Hence, contrary to MacIntyre’s claim, we do find in Hume an important sense in which
our moral feelings are responsive to criticism, rejection, development, modification, and
so on.
MacIntyre is also mistaken to think that this fact—viz., the fact that our moral
feelings are responsive to criticism, rejection, development, modification, and so on—
illustrates that moral judgments cannot be elucidated in terms of moral feelings. To the
contrary, Hume holds that a plausible analysis of our moral judgments must ultimately
appeal to the moral feelings experienced by those who satisfy certain conditions. This is

183
because all moral judgments ascribe intersubjective, dispositional, impression-dependent
moral properties. As discussed in chapter II, the domain of these properties is determined
by reference to the moral feelings psychologically normal observers would experience
when observing or contemplating the object in question while satisfying the conditions of
the moral view.
167
In short, in Hume’s view, moral properties are analyzed in terms of
moral feelings (had under certain conditions). Hence moral judgments, which ascribe
these properties, can be and indeed are elucidated in terms of moral feeling.
§V.5 Another objection to Hume’s account of sympathy & a reply
Mercer’s example discussed above helps to illustrate another objection to Hume’s
claim that sympathy is the source of our moral feelings and hence the “primary origin” of
moral distinctions. Recall the first two lines of Mercer’s example:
Suppose that I see A stopping B from doing something which it is obvious
to me that B very much wants to do. I may sympathize with B’s
frustration and misery and this sympathy will cause me to disapprove of
A’s conduct.

In the example, Mercer nonchalantly passes over the fact that my sympathetic
feelings are directed toward B’s frustration and misery, whereas my feelings of moral
disapproval are directed toward A’s conduct. However, this transition is prima facie
problematic. How can sympathy be the source of our moral feelings if the sympathetic
feelings produced by sympathy are directed at different objects than our moral feelings?
In the present example, the object of my sympathetic feelings is B’s emotions, but the

167
I have already argued that (i) implicit in Hume’s view is a distinction between our ideas of moral feeling
and our ideas of moral properties, and (ii) our ideas of moral properties are complex ideas that contain our
simple ideas of moral impressions.

184
object of our moral feelings is A’s conduct. Hence it seems that sympathy is neither the
source of our moral feelings nor therefore the origin of moral distinctions.
This conclusion is too hasty however. Presumably I, the appraiser, am aware of
the causal connection between the misery and frustration of person B, with whom I
sympathize, and A’s conduct. In other words, I am presumably well aware that A (or A’s
conduct) caused B’s misery and frustration. And this knowledge, in conjunction with the
sympathetic pain I feel for B, will presumably result in my having feelings of disapproval
toward A or at least toward A’s conduct. Working backward, the causal chain is roughly:
My felt sympathetic pain is a result of the sympathy process, which enlivens my idea of
B’s painful feelings (viz., misery and frustration). My idea of B’s feelings is basically the
result of my observation, namely, A’s conduct causing B’s pain. Due to various
principles of mental association (see §III.3), my sympathetic pain in conjunction with my
belief that A’s conduct is the cause of both B’s pain and ultimately my own sympathetic
pain (since my sympathetic pain is simply a response to my observation of B’s pain)
subsequently enliven in me a feeling of disapproval directed at A, or at least A’s conduct.
Hence sympathy is the source of the moral feelings of approval and disapproval
and thus the source of moral distinctions. It is noteworthy that, according to the
interpretation above, our sympathetic feelings engender these moral feelings. As
mentioned in chapter IV, Hume clearly holds that our extensive sympathetic feelings are
equivalent to these moral feelings insofar as both are indicative of moral distinctions.
However, up to this point I have remained neutral regarding whether or not Hume holds
that our sympathetic feelings are identical to these moral feelings. It is now clear that

185
they cannot be identical because (i) the object of our extensive sympathetic feelings in a
particular case can be, and typically is, distinct from the object of our moral feelings of
approval or disapproval, and (ii) extensive sympathetic feelings cause those feelings.
§V.6 The relationship of sympathy, reason, utility, & moral judgment
Although sympathy (and hence moral feelings) and our utility judgments (and
hence causal reasoning) can be identified apart from one another, we now know that they
are also intimately connected. But we need to know more precisely how sympathy,
reason, and our utility judgments are connected because Hume’s account of this
connection can have a significant impact upon an analysis of moral judgment/belief. For
instance, if our sympathy is simply tracking a particular independently specifiable
property of utility, then it may be unclear why we cannot identify this property as the
moral property ascribed by moral judgments. And if we can make this identification,
then the proffered dispositional, impression-dependent account of moral properties would
be rendered obsolete, as would the role of sympathy (and hence moral feeling) in
analyzing moral judgment/belief. So a lot rides on how Hume understands the
relationship between moral judgment/belief and our utility judgments/beliefs. It is
therefore crucial that we acquire a better grasp of the relationship between sympathy,
causal reasoning, our judgments of utility, and our moral judgments.

186
Hume’s emphasis on utility in discussing morality has led some to suggest that
Hume is among the first (if not the first) utilitarian theorist(s).
168
For example, in the
Treatise, Hume writes:
Meekness, beneficence, charity, generosity, clemency, moderation, equity,
bear the greatest figure among the moral qualities, and are commonly
denominated the social virtues, to mark their tendency to the good of
society. (T 578)

As mentioned, this emphasis is much more pronounced in the Enquiry. For instance,
Hume writes:
It appears to be a matter of fact, that the circumstances of utility, in all
subjects, is a source of praise and approbation: That it is constantly
appealed to in all moral decisions concerning the merit and demerit of
actions: That it is the sole source of that high regard paid to justice,
fidelity, honour, allegiance, and chastity: That it is inseparable from all the
other social virtues, humanity, generosity, charity, affability, leniency,
mercy, and moderation: And in a word, that it is the foundation of the
chief part of morals, which has reference to mankind and our fellow
creatures. (E 231, Hume’s emphasis)

In all determinations of morality…public utility is ever principally in
view. (E 180)

In these passages, and in many others, Hume certainly sounds like a utilitarian. In
fact, Jeremy Bentham—the true founder of Utilitarianism—claimed that it was “as if the
scales had fallen from my eyes” upon reading Hume, who showed him that “the
foundations of all virtue are laid in utility.”
169
But if Bentham and others are right and
Hume really is a utilitarian, then this would clearly conflict with the interpretation of
Hume proffered by this project. According to any and all utilitarian views, our moral

168
See, for instance, J. Rawls (1971: 22, fn 9); R. Glossop (1967: 527-36); J. Plamenatz (1947: 22); E.
Albee (1901: 112); and L. Stephen (1881: 87).

169
J. Bentham (1977: 440, n. 1 & 2).

187
judgments ascribe to objects moral qualities that are not impression-dependent properties.
So, while Hume would presumably still hold that sympathy accounts for the transition
from non-moral motives to moral motives, both sympathy and moral feeling would be
rendered obsolete in an analysis of moral judgments/beliefs and the moral properties they
ascribe. In short, the account given by this project would clearly not be Hume’s.
The proffered interpretation of Hume is not under threat, however, because
Hume’s moral theory is distinctly non-utilitarian. I shall now briefly explain why it is
non-utilitarian. As we shall soon see, in giving this explanation, we will also have
thereby explained the relationship between sympathy, causal reasoning, our judgments of
utility, and moral judgments.
As a start, it is noteworthy that Hume does not think that utility is the sole
standard for determining moral value.
170
This is illustrated by his distinction between our
approval of an object due to its utility to others or oneself, on the one hand, and our
approval of an object due to the fact that it is immediately agreeable to others or oneself,
on the other. For instance, Hume explicitly claims that while “some qualities produce
pleasure [in particular, moral approbation], because they are useful to society… others
produce it more immediately” (E 261). According to Hume, virtue can therefore be
“desirable on its own account” (E 293-4). This theme continually reappears both in the
Treatise and in the Enquiry. For instance, Hume writes: “there are other virtues and vices
beside those which have this tendency to the public advantage and loss” (T 578-9); and
“there is another set of mental qualities which, without any utility or any tendency to

170
Similarly, G. Sayre-McCord observes that Hume “resists treating utility for all as the final arbitrator of
moral questions” (1996: 282).

188
farther good, either of the community or the possessors, diffuse a satisfaction on the
beholders and procure friendship and regard” (E 86-7).
One might object that none of Hume’s claims above are incompatible with a
utilitarian reading. For instance, hedonic utilitarianism too holds that there is something
that has moral value, not because it is useful for something else, but in and of itself,
namely: pleasure. Pleasure is not a means to an end; it is an end in itself. The whole
reason objects are useful is because, by bringing pleasure to others and/or oneself, they
contribute to an increase in overall pleasure. Perhaps this is the main point of Hume’s
distinction between the useful and the immediately agreeable, to wit: all ultimate ends are
immediately agreeable. If so, then Hume’s account may indeed be indistinguishable from
a utilitarian account after all. Consider, for instance, the following passage:
Ask a man why he uses exercise; he will answer, because he desires to
keep his health. If you then enquire, why he desires health, he will readily
reply, because sickness is painful. If you push your enquiries farther, and
desire a reason why he hates pain, it is impossible he can ever give any.
This is an ultimate end, and is never referred to any other object. (E
Appendix I, sec. V [131], Hume’s italics).

There is one crucial difference between Hume’s account and utilitarianism
however. In Hume’s view, an object does not have moral value in virtue of the fact that
the object maximizes overall (units of) pleasure, nor simply because it tends to cause
pleasure per se in oneself or others. Rather, an object has moral value due to the fact that
it tends to fulfill certain ends/interests, namely, ends/interests that when anyone who
satisfies the conditions of the moral view then contemplates the object’s tendency to

189
fulfill those ends/interests thereby experiences, via sympathy, feelings of approval
toward the object. Hume explains this two-stage process as follows:
But though reason, when fully assisted and improved, be sufficient to
instruct us in the pernicious or useful tendency of qualities and actions; it
is not alone sufficient to produce any moral blame or approbation. Utility
is only a tendency to a certain end; and were the end totally indifferent to
us, we should feel the same indifference towards the means. It is requisite
a sentiment should here display itself, in order to give a preference to the
useful above the pernicious tendencies… Here therefore reason instructs
us in the several tendencies of actions, and humanity makes a distinction
in favour of those which are useful and beneficial. (EPM, Appendix I,
Hume’s emphasis)

Hence an object has neither moral value nor utility simply because it produces
(maximal) pleasure. Rather it has utility because it tends to fulfill ends that are useful to
others and/or oneself. It is true that the fulfilling of these ends will typically have
pleasurable effects; indeed, this is why the object is deemed useful. But the object’s
moral value is not due merely to those positive emotional effects. On Hume’s account,
the object has (positive) moral value because contemplation of the object’s causal
tendency to fulfill the relevant ends (and thereby typically produce positive emotional
effects) elicits feelings of approval, via sympathy, within anyone who satisfies the
conditions of the moral view. It just so happens that it is the nature of extensive
sympathy to elicit feelings of approval toward objects deemed useful. In fact, utility
itself has moral value only because we approve of it via extensive sympathy.
Thus, in opposition to utilitarianism, Hume’s moral theory appeals to the typical
effects of an object upon other individuals and/or oneself rather than its actual or

190
expected effects on everyone.
171
It also follows from Hume’s account that an object can
have moral value even when it does not maximize overall (units of) pleasure. This is due
to the fact that Hume’s moral theory, unlike utilitarianism, views moral properties as
dispositional, impression-dependent properties. Moral value is determined by the
feelings engendered by a disinterested (i.e., extensive) sympathy.
172

Nevertheless, in Hume’s view, there is no way to acquire the idea of this
dispositional, impression-dependent quality without also appealing to the object itself, for
the object has that moral quality—i.e., the object is such as to elicit the relevant feeling—
in part because of its tendency to fulfill particular ends. For instance, an object can elicit
moral approval in virtue of its causal tendency to fulfill ends that are useful to its
possessor or those who “have commerce” with her. Hence Hume claims that, on most
occasions, beauty “is not an absolute but a relative quality, and pleases us by nothing but
its tendency to produce an end that is agreeable. The same principle produces, in many
instances, our sentiments of morals” (T 577).
Consequently, causal reasoning too plays a vital role in morality because it
discerns whether or not an object is useful. Hume writes: “One principal foundation of
moral praise being supposed to lie in the usefulness of any quality or action, it is evident

171
Certain ceteris paribus conditions do apply to Hume’s account. In Hume’s view, an object has utility if
and only if ceteris paribus it tends to fulfill ends that useful to others and/or oneself. For instance, Hume
must hold that an object does not have utility if it tends to fulfill ends that are useful to oneself but also
tends to foil ends that are useful to others. This caveat is requisite; otherwise, for reasons to be discussed in
the following chapter, extensive sympathy would not always elicit approval of objects that have utility.

172
Sayre-McCord (1996: 293ff) too points out that Hume’s moral theory differs from utilitarianism insofar
as it need not rely on actual or expected utility as the determinants of evaluation, nor does it require that the
moral appraiser take into account all of the consequences on everyone effected. But he does not emphasize
that this is due to the nature of sympathy in conjunction with the general rules of the moral view.

191
that reason must enter for a considerable share in all decisions of this kind; since nothing
but that faculty can instruct us in the tendency of qualities and actions…” (EPM,
Appendix I, paragraph 2). The tendency of which Hume speaks just is the causal
tendency of various qualities and actions to fulfill (or foil) a given end or a set of ends.
After all, in Hume’s view, what makes a quality or action useful is the fact that it tends to
fulfill ends that produce pleasure in its possessor or those associated with her.
173

As it stands, however, this account appears to be compatible with the utilitarian
position that our ascriptions of moral qualities are really nothing more than judgments of
utility. So it might be argued that Hume turns out to be some sort of utilitarian after all. I
disagree. As I read Hume, there is an important distinction between utility and disutility,
on the one hand, and moral properties on the other. I suggest that Hume conceives of the
utility, or usefulness, of an object not as a moral property but as, in effect, a moral-
making property distinct from the moral quality itself.
To illustrate this, consider once again Hume’s account of the relation between the
“form” of an object—e.g., the surface structure of its parts—and its beauty. The object’s
form is not identical to its beauty. The form is not a sentiment-dependent quality of the
object, whereas its beauty is a sentiment-dependent quality. Recall from chapter II that
Hume views the pleasure that the object’s form produces as an essential component of
beauty. For instance, Hume writes: “we may conclude, that beauty is nothing but a form,
which produces pleasure, as deformity is a structure of parts, which conveys pain; and
since the power of producing pain and pleasure make in this manner the essence of

173
“Usefulness is only a tendency to a certain end” (EPM, Appendix I, part II, paragraph 2).

192
beauty and deformity, all the effects of these qualities must be deriv’d from the
sensation” (T 299). In other words, the form is in effect a beauty-making quality of the
object insofar as it is the object’s form that engenders the pleasurable feeling.
As discussed in chapter II, Hume conceives of moral qualities (and beauty) as a
“relation…between the form [of an object] and the sentiment [in the judge]” (SOT, par.
10). In moral cases, I suggest that Hume conceives of the form of the relevant objects
(e.g., mental qualities and actions) as that in virtue of which they possess the tendency to
fulfill (or foil) ends and thereby cause the sentiment. Only causal reasoning can tell us
about an object’s form since, as Hume points out, nothing but reason can “instruct us in
the tendency of mental qualities and actions”. Moreover, in moral cases, an object’s form
is thereby directly related to the object’s usefulness or agreeableness (or perniciousness
or disagreeableness) since it is in virtue of its form that it has these qualities.
The object’s moral quality, on the other hand, is a relation of both the form of the
object and the moral feelings it thereby produces (via extensive sympathy). It is therefore
a sentiment-dependent property, for the feeling is “essential” to the property. Thus,
according to Hume’s view, an object’s utility or disutility is distinct from its moral
quality. Hume’s moral theory is therefore clearly not a form of utilitarianism.
Nevertheless, it is clear why Hume’s moral theory can be easily misinterpreted as
espousing a kind of utilitarian position. Hume repeatedly observes the close connection
between our judgments of utility and our moral judgments and sentiments. For instance,
he writes: “we must allow, that the reflecting on the tendency of characters and mental
qualities, is sufficient to give us the sentiments of approbation and blame” (T 577). And

193
elsewhere Hume appears to overstate the connection when he claims that the “distinction
between what is useful, and what is pernicious…is the same in all its parts, with the
moral distinction” (EPM [80]).
Hume’s point in these passages is, I think, clear once one grasps the relationship
between sympathy, causal reasoning, our utility judgments, and our moral judgments and
feelings. Causal reasoning “paves the way” for moral sentiment by determining whether
the object in question is such as to fulfill (or foil) various ends.
174
The appraiser’s
observation or mere contemplation of this produces an idea of the pleasure (or pain) that
the object, in virtue of its form, is such as to cause in its possessor or others who associate
with her; i.e., it produces an idea of the object’s utility (or disutility). The psychological
process of sympathy then converts the appraiser’s idea of another’s pleasure (or pain) into
a feeling of approval (or disapproval) toward the object.
So, as I read Hume, our judgments about the causal tendency of objects to fulfill
(or foil) particular ends have an important normative function: they can be used to
evaluative behavior and give others advice about how to act. But at least in moral cases,
such evaluations and advice will only have motivational efficacy when we are moved by
sympathy. This makes sympathy “a very powerful principle of human nature”, according
to Hume. Without the psychological process of sympathy there would be no moral
motivation. Indeed, without sympathy there would be no moral qualities; for sympathy
plays the central role in accounting for the interplay between form and feeling, and this
relation between form and feeling just is what constitutes moral qualities. Thus,

174
See EPM, section I, paragraph 9.

194
sympathy turns out to be the “chief source” and “primary origin” of moral qualities (T
574ff). This, I suggest, is why Hume stresses the role of sentiment over that of reason.
175

According to Hume, the distinction between what is useful and pernicious is, in
effect, the same as the moral distinction because the moral feelings that determine moral
distinctions arise “immediately”, via the psychological process of sympathy, in response
to our reasoned utility judgments (i.e., our “reflecting on the tendency of characters and
mental qualities”). As Hume puts it: “Wherever an object has a tendency to produce
pleasure in the possessor, or in other words, is the proper cause of pleasure, it is sure to
please the spectator, by a delicate sympathy with the possessor” (T 576-7, his emphasis).
Thus, whenever we observe or contemplate the form of an object, the relevant moral
feeling is sure to be present (assuming the conditions of the moral view are satisfied).
Form and feeling are, in this way, tied together.
This also helps explain Hume’s claim that, even when some virtues are
“abstracted from any consideration of utility or beneficial tendencies” (E 261), we still
approve of them. As we’ve just discussed, many objects have moral value because,
thanks to their form, they are useful. However, other objects have moral value because,
thanks to their form, they are immediately agreeable. Unlike those instances in which we
judge (in light of the moral view) that objects are useful, causal reasoning plays no role in
determining the beneficial features of objects that are immediately agreeable. No utility

175
Sayre-McCord (1996: 293) argues that, on Hume’s theory, there is “no single overarching standard” for
comparing interpersonal values. I disagree. While there is no single utility calculus, extensive sympathy
is, for Hume, the single overarching standard for comparing moral values. However, Sayre-McCord is
right that Hume’s account rejects “the utilitarian commitment to the commensurability of consequences”
(ibid). Although extensive sympathy is the standard for comparing conflicting moral ends, there is no
guarantee it will produce a decisive result. Moral incommensurability is possible in part because, unlike a
utilitarian calculus, sympathy can vary depending on social context and other factors. (See ch. VI & VII.)

195
judgment needs to be made in such instances precisely because their beneficial effects are
felt immediately. We approve of objects that are immediately agreeable to others and/or
ourselves, not because we judge that they fulfill ends that are useful, but simply because
they fulfill ends that are immediately agreeable.
§V.7 Two final remarks
The connection between our moral feelings and objects’ causal tendency to fulfill
ends also helps explain why Hume holds that the primary objects of moral evaluation are
the agent’s intentions (T 348). Acts that are directed towards an end or a set of ends are
the only kind of act that is rightly considered an object of moral evaluation. And in order
for an act to be end-directed, the agent must intend to bring about a particular end or set
of ends by performing the act. Hence the primary objects of our moral judgments turn
out to be the agent’s intentions to bring about a particular end or set of ends, ends that
will have pleasurable or painful effects.
Of course, as Hume points out (T 349), we tend to look past any accidental
consequences of an agent’s act and hold the agent morally responsible primarily because
of her intentions to bring about a particular end or set of ends that the agent knows will
probably have pleasurable or (as the case may be) painful effects. A notable exception to
this rule is our moral evaluations of an agent’s character traits. For instance, in some
cases, an agent might act out of habit without consciously forming an intention to bring
about a particular end that she knows will likely have particular emotional effects.
Nevertheless, we often still hold the agent morally responsible for her act anyway. This
is due to the fact that, even in such cases, the agent’s repeated intentional aiming at a

196
particular end is presumably what establishes the habit, or disposition, in the first place.
And she will presumably be aware of the typical emotional effects of attaining that end.
Hence the agent is morally responsible for many of her intentional acts even in cases in
which the agent does not perform them with knowledge of their emotional effects in that
particular instance.
One final point is worth reiterating. I have suggested that after writing the
Treatise Hume began to more fully appreciate the significance of the relationship of
sympathy, causal reasoning, utility judgments, and moral judgments. Consequently,
when he re-presents his moral theory in An Enquiry Concerning the Principles of Morals,
I suggest that Hume incorporated the sentiment of humanity to further emphasize, as
vital, the link between judgments of social utility and sympathy, and thereby its integral
link with moral sentiment and judgment. Indeed, this connection between social utility,
sympathy, and morality certainly seems to be much more explicit in the Enquiry than it is
in the Treatise. As we shall briefly discuss, the connection between social utility and our
feelings of moral approval is especially important in accounting for the artificial virtues,
e.g., justice. And this connection is brought out more perspicuously in the Enquiry, in
part, thanks to the Enquiry’s inclusion of the sentiment of humanity as an expository
device for his moral theory.
§V.8 Coda
Hume sets out in the Treatise and the Enquiry “to discover the circumstances on
both sides, which are common to these qualities; to observe that particular in which the
estimable qualities agree on the one hand, and the blamable on the other; and thence to

197
reach the foundation of ethics, and find those universal principles, from which all censure
or approbation is ultimately derived” (E 174). Recall that this is, in effect, our main goal
in part one of our project, namely, to explain the nature of moral properties.
Following Hume, we have discovered that we approve of all virtues (from the
moral view) because, by being such as to fulfill various ends/interests, they tend to be
(ceteris paribus) either useful or immediately agreeable to its possessor or those who
associate with her. Contemplation or observation of these causal tendencies of various
traits elicits, through sympathy subject to the general rules of the moral view, feelings of
approval toward those traits. Similarly, we have discovered that we disapprove of all
vices (from the moral view) because, by being such as to foil various ends/interests, they
tend to be (ceteris paribus) either pernicious or immediately disagreeable to its possessor
or those who have commerce with her. Contemplation or observation of these causal
tendencies of various traits elicits, through extensive sympathy, moral feelings of
disapproval toward those traits.
176

Thus, the circumstances common to all virtues are that they are useful or
immediately agreeable either to others or oneself (with certain corrective conditions in
place), and the circumstances common to all vices are that they are pernicious or
immediately disagreeable to others or oneself (with certain conditions in place).

176
As we shall discuss later in part two, this project exploits Hume’s distinction between utility
judgments—viz., judgments about the tendency of objects to fulfill (or foil) particular ends—on the one
hand, and sympathy and moral feelings, on the other. This distinction parallels a distinction to be drawn by
this project between value properties—viz., the property of an object being such as to fulfill, further, or foil
particular ends—and moral properties—viz., the property of an object being such as to elicit, via sympathy,
feelings of approval or (as the case may be) disapproval within those who satisfy the conditions of the
moral view. Among other things, this ontological distinction will play a key role in providing an analysis
of moral and evaluative language. In short, it will help this project to reinvigorate Hume’s sensibility
theory of morality.

198
What’s more, following Hume, we have found that the foundation of ethics and
the universal principles from which all censure or approbation is ultimately derived is, as
promised, sympathy and the general rules of the moral view, which of course produce the
moral motives of extensive benevolence (a.k.a., the sentiment of humanity) and a sense
of moral duty. We’ve also found that causal reasoning—in particular, reasoning about
the causal tendency of objects (e.g., acts or traits) to fulfill or foil ends/interests and
thereby cause pleasure or (as the case may be) pain—is often requisite to trigger the
sympathy process. Hence causal reasoning too is part of the foundation of ethics.
Now that we have a better grasp of Hume’s theory of sympathy as the non-moral
foundation of morality, we can turn our attention to the moral view and its regulative
principles. As we shall discuss in the next two chapters, the “general rules” that comprise
the moral point of view are explicated according to both our capacity for sympathy and
our capacity for causal reasoning. By carefully laying out these principles, we will
consequently give further insight into the nature of sympathy and its connection to moral
impressions and moral properties. We will also acquire further insight into the nature of
moral properties, for the moral point of view lays out constraints upon sympathy that are
necessary in order for sympathy to engender the sentiments that determine moral
distinctions.

199
Chapter VI: The Disinterested Moral Point of View

§VI.1 General rules & Hume’s conception of the moral view
Thus far, I have argued for a particular interpretation of Hume’s moral theory,
according to which moral judgments are descriptive, property ascribing, truth-evaluable
judgments, some of which are true; hence there are moral properties and facts, and both
moral error and disagreement are not only possible but do actually occur. Chapter II has
argued that Hume conceives of moral properties as dispositional, feeling-dependent
qualities. Hume holds the following characterization: an object—e.g., an act or trait—has
a moral property if and only if observation or contemplation of the tendency of that
object to fulfill or (as the case may be) foil particular ends elicits an attitude of approval
or (as the case may be) disapproval within those who satisfy certain conditions. Because
moral properties have this essential link with these attitudinal responses, any plausible
analysis of moral properties and moral judgments must appeal to those attitudes.
Hume holds that our moral attitudes are the product of sympathy. Sympathy is
the naturally occurring non-moral psychological process through which our beliefs about
others’ past, present, or future mental states can cause sympathetic feelings and thereby
moral feelings in ourselves. These feelings can move to act us because they have a direct
link with the passions via various principles of mental association. (See chapter III.)
However, as discussed in chapters IV and V, the extent of sympathy’s natural
variability and partiality prevents our unregulated sympathy from producing moral
motives such as the moral sentiments and extensive benevolence. Sympathy per se can

200
only produce a more limited benevolence, and sometimes it fails to do even that.
According to Hume, the original non-moral motives of self-interestedness and limited
benevolence provide the impetus for us to establish principles that regulate our natural
but limited sympathy, thereby producing a more extensive, or disinterested, sympathy.
And significantly, the byproducts of this regulated sympathy are moral motives, in
particular, moral sentiments and an extensive benevolence.
177

This explains why Hume claims that sympathy is the “chief source” of our moral
motives and thereby genuine moral distinctions. Without sympathy, in conjunction with
the disinterested moral point of view, there would be neither extensive benevolence nor
the moral sentiments, and hence no moral distinctions. “Tis only when a character is
considered in general, without reference to our particular interest, that it causes such a
feeling or sentiment, as denominates it morally good or evil” (T 472).
In short, it is not just what we feel that matters in Hume’s view but also how we
come to feel these feelings. The feelings generated by sympathy are indicative of moral
distinctions only when the conditions of the moral point of view—henceforth MPV—are
satisfied. The conditions of the MPV secure uniformity and stability in our moral
judgments insofar as all human beings who satisfy these conditions will have the same
basic feelings—in particular, feelings of moral approbation and/or disapprobation. The
MPV accomplishes this in part by securing agreement in belief about the typical
emotional effects of the act or trait in question. Significantly, our moral judgments are in

177
Hume does admit that there is a “natural sentiment of approbation and blame” (T 579), but the moral
sentiments only occur under the conditions imposed by the moral view; and, as we have discussed, that
shared perspective does not occur naturally, but instead is (at least partially) a product of artifice.

201
this way capable of empirical confirmation/validation or disconfirmation/invalidation.
Hence we have already made considerable progress toward explaining Hume’s system of
validating and invalidating moral judgments. However, we have yet to discuss in any
detail the corrective conditions, or “general rules”, that comprise the MPV.
As anticipated, the key to fully understanding the nature of moral properties and
the judgments that ascribe them lies in Hume’s account of the general rules that comprise
the MPV. Laying out these general rules is the last step in accomplishing the primary
goal of part one of the project. Unfortunately, however, Hume does not delineate many
of these rules in a lucid and/or systematic way. Hence the last and perhaps most
important explanatory step in Hume’s moral theory faces a serious problem: various
aspects of Hume’s MPV are arguably too vague for the MPV to serve as a regulatory
device for sympathy and hence moral feeling and judgment.
Based upon what Hume actually says, this criticism is not entirely without merit.
But its merits should not be overstated. Hume does at least provide a basic framework
for the MPV, and this framework is rather informative. For instance, Hume, in effect,
conceives of morality as having both a cognitive aspect—viz., the role of causal
reasoning in morality, in particular, reasoning about the tendency of objects to fulfill or
foil various ends—and an emotional aspect—viz., the role the psychological process of
sympathy in morality.
As discussed in the previous chapter, there is an intimate relationship between
these two aspects of morality. The sympathetic/emotional aspect of morality is

202
responsive to the causal reasoning/cognitive aspect.
178
Nevertheless, Hume gives the
emotional aspect priority in his “system of morality” because, according to that system,
the domain of moral properties is determined by our sympathy-induced moral feelings
(when various general rules are satisfied).
This division between these two aspects is significant because Hume, in effect,
bifurcates the rules, or norms, of the MPV along this division. Hume’s MPV provides a
set of general rules that govern the cognitive thought process (i.e., causal reasoning) on
the one hand, and another set of general rules that govern the emotional process (i.e.,
sympathy) on the other.
The norms governing the cognitive side of the equation are rather perspicuous.
As we shall discuss, Hume holds that we naturally exhibit various tendencies in our
causal reasoning that produce irrational, or unreflective, prejudices, which often results in
false causal/predictive judgments. This tends to produce disagreement and other practical
problems. We apply corrective general rules routinely in our causal reasoning to
eliminate or at least mitigate those tendencies. This not only brings stability and
uniformity to our causal/predictive judgments, and hence agreement in judgment; it also
resolves many of the practical problems that result from those tendencies.
179
These
general rules appear to contribute to the successful regulation of our causal reasoning—
e.g., as demonstrated by the success of the sciences. This is not to say that all of the

178
In part two of the project, we shall briefly discuss how Hume’s division between the cognitive and
emotional aspects of morality, as well as his claim that they often work in tandem, is supported by recent
observations in the field of cognitive neuroscience. I shall also go on to argue that this division is also
revealed in the language we use to deliver our moral evaluations.

179
In the Treatise (173-6), Hume lists 8 regulative general rules. (See §VI.3.1.)

203
general rules Hume proposes on the cognitive side are uncontroversial—see §VI.2 in
which I raise doubts about some of those rules. But there is good reason to think that any
troubling ambiguity in Hume’s MPV doesn’t reside on the cognitive side of the equation.
So any significant ambiguity to be found in the MPV appears to rest on the
emotional side, that is, with the norms governing sympathy. As discussed, Hume holds
that our sympathy has natural tendencies toward partiality and indifference. These
tendencies produce disagreement in our moral judgments, as well as other practical
problems, both of which produce negative feelings within us. We are thereby motivated
to apply corrective general rules to eliminate, or at least mitigate, these tendencies. As
also discussed, we apply these regulative principles because it (i) brings stability and
uniformity to our moral judgments and (ii) resolves practical problems that arise due to
sympathy’s natural tendencies. These regulative principles accomplish this by producing
agreement in both feeling and judgment. For instance, when successfully applied, these
general rules engender an “extensive sympathy”, and thereby moral feelings and
“extensive benevolence”, which have many personal and social benefits.
We know that the MPV’s principles that regulate sympathy produce extensive
sympathy by requiring moral spectators to be disinterested. For example, Hume claims
that, as spectators, we are to overlook our own interests when issuing moral judgments.
So Hume does offer at least some insight into the general rules that regulate sympathy.
However, the textual evidence supporting the details of those rules is rather abstruse, if
not entirely absent in some cases. It is just not entirely clear what exactly Hume’s full
conception of disinterestedness amounts to. Nor is it clear precisely how Hume attempts

204
to regulate sympathy’s natural tendencies toward partiality and indifference. For
instance, do these general rules of the MPV aim to exclude sympathy’s partiality entirely,
or do they aim to preserve what are arguably important aspects of sympathy’s partiality?
This and the following chapter aim to answer these and other important questions
by offering a particular interpretation of Hume’s MPV. We shall begin by discussing the
general rules that Hume claims regulate causal reasoning/judgment. We will then turn
our attention to the general rules, or principles, that Hume claims regulate sympathy. If I
am right, then any ambiguity found in Hume’s MPV is not as problematic as it may
initially appear.
§VI.2 The influences of general rules
Hume claims that general rules have two primary kinds of influence. Some
general rules have a corrective influence and others have an error-prone influence.
180
All
general rules are established from custom and experience. And we are easily influenced
by the customs of the society in which we live, as well as the habits we’ve already
formed from past experience. The error-prone influence of some general rules is due to
the fact that these are general rules “which we rashly form to ourselves, and which are the
source of what we properly call PREJUDICE” (T 146, Hume’s emphasis), or what Hume
elsewhere refers to as a “very unphilosophical species of probability” (T 150). Because
general rules are products of the imagination informed by custom and habit, any
prejudices we might have can cause us to form general rules that reinforce those
prejudices and thereby easily lead us astray. For instance, we can mistakenly generalize

180
T 147-151. Kathleen Wallace (2002) also discusses these two influences of general rules.

205
over what are, in fact, accidental or irrelevant considerations and/or overlook non-
accidental and relevant considerations. Hume uses the following famous example to
illustrate our error-prone generalizing tendencies.
An Irishman cannot have wit, and a Frenchman cannot have solidity; for
which reason, tho’ the conversation of the former in any instance be
visibly very agreeable, and of the latter very judicious, we have entertain’d
such a prejudice against them, but they must be dunces or fops in spite of
sense and reason. Human nature is very subject to errors of this kind… (T
146-7)

In short, custom and habit can contribute to the establishment of false general
rules that give rise to jingoism, racism, sexism, and other systemic forms of bigotry and
prejudice.
181
This is significant because false prejudices impact our judgments, including
judgments involving our own experiences. For instance, as described in Hume’s example
above, our false prejudices can become so ingrained within us that we overlook or ignore
observations that are contrary to some general rules, thereby preserving our error-prone
general rules and perpetuating false prejudices and future miscalculations. This not only
contributes to disagreement; it can also lead to various other practical problems. Hume
warns that we must therefore be careful in both forming and applying general rules.
Fortunately, sometimes we do recognize our error-prone generalizing tendencies,
e.g., our tendency to over generalize. In order to regulate these tendencies and the error-
prone general rules that already exist and that we readily apply out of custom and habit,
Hume holds that we must use the understanding/reason to form and follow other general
rules. So, although general rules can be the product of various forms of prejudice, it is

181
This tendency of the mind is ultimately due to various principles of mental association—e.g.
resemblance and other forms of comparison—that underlie our habits and customs.

206
“only by following them that we can correct this, and all other unphilosophical
probabilities” (T 150). Thus, while some general rules tend to have an error-prone
influence, we form others that have a corrective influence and counter-act our
“unphilosophical”, error-prone rules and tendencies. For instance, Hume writes:
We shall… take notice of some general rules, by which we ought to
regulate our judgment concerning causes and effects; and these rules are
form’d on the nature of our understanding, and on our experience of its
operations in the judgments we form concerning objects. By them we
learn to distinguish the accidental circumstances from the efficacious
causes; and when we find that an effect can be produc’d without the
concurrence of any particular circumstance, we conclude that that
circumstance makes not a part of the efficacious cause, however
frequently conjoin’d with it. (T 149)

The corrective influence of such general rules is important because it helps secure
stability and uniformity of judgment, thereby avoiding interpersonal disagreement and
other negative consequences. However, Hume also observes that the corrective rules that
we put in place do not always successfully eradicate error-prone general rules or the
negative influences some rules have on our beliefs and feelings. Quite often corrective
principles, even when applied successfully, only palliate the influence of our error-prone
generalizing tendencies. For instance, they may only mitigate, rather than eliminate, our
prejudicial beliefs and feelings formed and/or justified on the basis of general rules.
182

[C]onsider the case of a man, who being hung out from a high tower in a
cage of iron cannot forbear trembling, when he surveys the precipice
below him, tho’ he knows himself to be perfectly secure from falling, by
his experience of the solidity of the iron, which supports him; and tho’ the
ideas of fall and descent, and harm and death, be deriv’d solely from
custom and experience. The same custom goes beyond the instances,
from which it is deriv’d, and to which it perfectly corresponds; and
influences his ideas of such objects as are in some respects resembling, but

182
K. Wallace (2002) also stresses this point.

207
fall not precisely under the same rule. The circumstances of depth and
descent strike so strongly upon him, that their influence cannot be
destroy’d by the contrary circumstances of support and solidity, which
ought to give him a perfect security.
183
(T 148)

§VI.3 General rules & the regulation of causal reasoning
As Hume’s example above illustrates, we use corrective general rules to guide our
judgments of cause and effect, which are derived from prior experience, with the hopes of
extricating any tendency for miscalculation due to custom and habit. But sometimes the
opposing ideas that arise from custom and habit are too vivid to be “destroyed”. Ideas
that have such a strong vivacity will, in effect, force themselves on us without any
reflection whatsoever. The best we can hope for in such cases is that our corrective
general rules are sufficient to alleviate that force and vivacity, paving the way for other
opposing ideas and beliefs to take hold of us. The opposing ideas and beliefs will then
create internal conflict. And this conflict typically engenders further reflection, which
usually reinforces the corrective general rules and any ensuing ideas and beliefs.
This reflective reinforcement of those rules is due to the fact that our corrective
tendencies, and the remedial rules they generate, are in part a product of the
understanding, in particular, our understanding of the nature of causes and effects, which
guides the imagination. Returning to Hume’s example, recall that the imagination is
what gives vividness to certain ideas, such as fall, descent, harm, and death, based upon
one’s habits of association, experiences, and customs. In Hume’s view, the
understanding and the corrective principles governing causal reasoning remedy our error-

183
Saul Traiger (2005) offers an informative discussion of this example.

208
prone tendencies by guiding the imagination so as to reduce the vividness of various
ideas and thereby affect our judgments of cause and effect. If successfully applied, these
regulative rules about cause and effect will at least sufficiently mitigate the vividness of
those ideas even if it cannot completely eradicate those ideas from our present conscious
thinking.
As we’ve discussed in the previous chapter, causal reasoning plays an important
role in morality because it informs us about the tendency of objects (e.g., traits or acts) to
fulfill or foil various ends and thereby cause pain or pleasure. But causal reasoning also
predicts the severity of any anticipated pain or pleasure. This is significant because, in
Hume’s view, (1) the extent to which pleasurable or painful feelings are deemed probable
directly corresponds to the extent to which the imagination focuses on the ideas of
pleasure or pain, or other related ideas (e.g., fall, descent, harm, and death); and (2) the
imagination enlivens those ideas according to the anticipated severity of the
corresponding feelings.
184
This, in turn, impacts our motivation to act via various mental
processes, including sympathy and the association of ideas and impressions.
Thus, by diminishing the initial influence of the imagination upon certain ideas,
the corrective general rules regulating causal reasoning allow other ideas (e.g., solidity,
security, and support) to take hold in us via our understanding and imagination. The
result is that we will have different thoughts and feelings than we otherwise would.
Significantly, this means that the general rules that regulate causal reasoning have the

184
See also Traiger (2005).

209
potential to drastically affect our motivations to act in various ways, a fact that clearly has
important implications for moral motivation.
Of course, as we have already discussed, there is no guarantee that the corrective
general rules will have sufficient influence on the imagination to overcome its tendency
to mislead. So, for instance, a coward will often have fearful feelings even when he
knows from the application of corrective general rules and reflection that there is actually
a very low probability that he will be caused harm. The coward will often have such
feelings because he “readily assents to every account of danger” (T 120).
For example, standing on the edge of a precipice, the coward acquires the idea of
its height and is predisposed to associate that idea with other ideas such as fall, descent,
harm, and danger. Even supposing he knows that there is little chance of falling, his
imagination gets the best of him and enlivens his idea of danger. Such predispositions of
a coward’s imagination make him hypersensitive to various beliefs, such as the belief that
he will fall, even though his understanding tells him that it is highly unlikely. The
enlivening of these ideas and beliefs causes strong feelings of fear. We can chalk the
coward’s predisposition up to what we might call an ‘irrational’ habit of association.
Moreover, being in the grip of fear, the coward is much more susceptible to
overestimating the probability and severity of harm in the majority of cases. In this way,
custom and habit often give rise to misevaluations by (mis)leading us to either over or
underestimate the probability of a certain outcome and/or by over or underestimating the
severity of that outcome. This proneness to inaccurately assess the situation can also

210
create a sort of feedback loop by further triggering the imagination and thus further
enlivening the idea of harm/pain, thereby motivating cowardly action.
In any case, the important point is that reflective awareness of our ‘irrational’
habits can sometimes overcome these habits by manipulating our experiences in order to
create new habits that seem more ‘rational’. For instance, as Hume observes: “Custom
soon reconciles us to heights and precipices, and wears off these false and delusive
terrors” (Enquiry). Or to illustrate this with another example, one way to overcome
aviatophobia, or what is sometimes called the ‘irrational’ fear of flying, is to embark on
frequent flights. By doing so, we may be able to overcome our fear both by breaking
various mental associations/beliefs we have about flying and by building new
associations/beliefs, such as the experienced-based thought that flying is a safe and
efficient way to travel. We may also be able to overcome our fear by reflecting on
various “matters of fact,” such as the fact that we are statistically much more likely to be
killed in the vehicle that takes us to and from the airport than we are in the airplane that
takes us to our destination. It is the truth of such modal judgments that leads us to call
the fear of flying ‘irrational.’
Of course, these steps might neither defeat nor assuage our fear of flying. Our
fear might be much too intense for these ratiocinations to have either an immediate or
long-term impact. Hume would no doubt say that our habits and/or imagination have
gotten the best of us, and consequently we are caught in the grip of certain emotions and
preconceived ideas. But we can also begin to reflect on why our ‘irrational’ fear has us in
its grasp, why we are having such difficulty managing it. Perhaps we have lost a loved-

211
one in a plane crash; perhaps we have issues with relinquishing control; or perhaps we
have grown attached to our fear.
This example also nicely illustrates how our feelings such as fear can be quite
useful. They can be useful because these feelings can call attention to salient features of
the situation and prompt change by causing us to reflect upon and reevaluate our beliefs
and feelings. We can use our reflective awareness—sometimes with the help of others,
such as a friend or therapist—to try to identify the cause of our unregulated, or
‘irrational’, beliefs and feelings in order to redress that cause.
In sum, corrective general rules are put in place precisely because, overtime,
experience shows us both that we have error-prone tendencies and that exercise of these
tendencies often has adverse consequences and creates various practical problems.
Through experience and reflection, we arrive at various regulatory principles to remedy
those tendencies and thereby resolve the practical problems that they create. In other
words, we try to incorporate into the corrective conditions our understanding of all of the
mechanisms and processes that give rise to inconsistencies in judgment and unreflective
feelings. This allows us to construct principles, or rules, that can counteract those habits
and customs that cause the inconsistencies.
Because these corrective general rules are “more extensive and constant” than the
other general rules, they not only promise to ensure uniformity and stability in our
judgments, which is important for reasons we have already discussed; they also have the
potential to create new, more consistent habits and customs that are conducive to more
beneficial outcomes. In other words, these rules are internalized; they become a matter of

212
habit and so are applied unconsciously. But whether the corrective principles are
ultimately successful at regulating one’s habits and customs depends primarily upon that
person’s character and dispositions—e.g., one’s mental habits and temperament,
including the ability to exert control over one’s passions. Of course, it also depends upon
one’s past experiences, as well as the circumstances surrounding the situation.
185

§VI.3.1 Eight general rules for regulating causal reasoning
As discussed, Hume’s view is that corrective general rules can influence our
imagination, and hence judgment and feeling, primarily by distinguishing between those
features that are accidental and those that are truly, or at least most likely to be, the actual
causes of the considered effects. The accidental features are irrelevant to our assessment
of causal connections, but because our imagination is easily led astray by experience,
habit, and custom we sometimes fail to identify and distinguish between the accidental
and efficacious features. So, for instance, we might seize upon a particular quality as the
cause of a particular effect based upon prior experience and/or custom. For example, we
might be taught that a particular quality is a non-accidental cause (such as someone’s race
or gender) but we are mistaken. Nevertheless, through habit and custom, that false causal
connection is unfortunately treated as a sort of general causal rule. In order to avoid this
proneness toward error, Hume offers 8 corrective general rules for regulating causal
reasoning. They are as follows:
(1) The cause and effect must be contiguous in space and time. (2) The
cause must be prior to the effect. (3) There must be a constant union
betwixt the cause and the effect… (4) The cause always produces the

185
See also Wallace (2002) on this point.

213
same effect, and the same effect never arises but from the same cause…
(5) Where several different objects produce the same effect, it must be by
means of some quality, which we discover to be common amongst them…
(6) The difference in the effects of two resembling objects must proceed
from that particular, in which they differ… (7) When any object encreases
[sic] or diminishes with the encrease or diminution of its cause, ‘tis to be
regarded as a compound effect, deriv’d from the union of the several
different effects, which arise from the several different parts of the
cause… (8) An object, which exists for any time in its full perfection
without any effect, is not the sole cause of that effect, but requires to be
assisted by some other principle, which may forward its influence and
operation. For as like effects necessarily follow from like causes, and in a
contiguous time and place, their separation for a moment shews [sic], that
these causes are not compleat ones. (T 174-5)

Many of these general rules appear to be rather intuitive, but some do not. For
instance, rules 4 and 5 are certainly not uncontroversial. And our current understanding
of quantum phenomena flies in the face of many of these rules. Of course, the quantum
world is filled with counter-intuitive phenomena that scientists still struggle to explain.
Fortunately, the exact content of Hume’s general rules isn’t the main issue for our
purposes. What is important is that we do indeed apply general rules (at least somewhat
similar to those Hume lists above) to govern our judgments of cause and effect. And
such rules do appear to help us overcome our prejudices and other error-prone tendencies
by distinguishing between accidental features and actual causes.
For instance, Hume was quite critical of religious views of morality that appeal to
faith in divine authority to ground moral claims because, in Hume’s view, the sort of
reasoning involved in such an appeal to authority is nothing more than prejudicial and
hence error-prone causal/probabilistic reasoning. To illustrate this, consider the Old
Testament story according to which God tested Abraham’s faith by commanding him to

214
sacrifice his son Isaac. Abraham blindly obeyed God’s command, taking it as an article
of faith that God would not command him to act in a manner that is contrary to the
dictates of morality. In Hume’s view, Abraham’s blind obedience to the divine is largely
due to Abraham’s belief that the sacrifice of his son’s life must somehow bring about a
moral end, an end that Abraham approves of via sympathy.
Hume thinks that such religious faith in an unknown future benefit of the agent’s
act is the product of purely prejudicial and unfounded probabilistic reasoning. Of course,
in the biblical story, God spares the life of Abraham’s son at the last minute, which
reaffirms Abraham’s blind faith in divine authority—indeed the “moral” of the story
appears to be reassurance that the perfectly obedient should not fear that God will allow
them to do wrong and hence that those who are truly faithful should not question blind
devotion to God’s commands. Hume is critical of such religious devotion precisely
because he believes that the sort of probabilistic reasoning associated with it is riddled
with prejudicial and error-prone tendencies. And this can, consequently, lead to
disastrous outcomes.
Hume argues that corrective general rules can help us overcome such prejudices
and error-prone tendencies, in part by distinguishing between accidental features and
actual causes. Scientific methodology and its success in explaining all kinds of causal
phenomena provide strong evidence in support of this claim.
186
Our ability to distinguish
between accidental features and actual causes has brought stability and uniformity to our

186
It is no accident that the scientific revolution served to undermine a significant amount of religious
superstition. Even today, many people view science and religion as issuing conflicting views of the world.

215
cause/effect judgments, and thereby substantial agreement in our understanding of the
physical world, at least among those whose judgments are guided by empirical data.
187

Nevertheless, there clearly still is substantial disagreement about causes and
effects, including the effects of various acts and traits. It might be objected that these
general rules therefore fail to ensure stability and uniformity in judgment, and this gives
us reason to call into question the success of such general rules. But I suggest that it does
not. The disagreement that we find with respect to our judgments of cause and effect are
all but entirely due to the lack of relevant information and/or the lack of understanding of
various causes and effects. In other words, current disputes are the result of current
epistemological limitations rather than some sort of failure with the general rules.
In fact, one could plausibly argue that disagreement about causes and effects are
partly the result of a failure to apply corrective general rules to our judgments, and partly
a consequence of the fact that acquiring the requisite experiences to clearly distinguish
between accidental and efficacious qualities in particular cases is often a slow and
arduous matter that requires social cooperation. Consider, for example, all of the
cumulated work across disciplines over the last two and a half centuries that has
contributed to scientific discoveries and progress in our understanding of the world.
We have every reason to believe that disputes about causes and effects (e.g., the
consequences of various actions) will begin to fade as relevant information becomes

187
We might therefore wish to add to Hume’s set of general rules governing causal reasoning/judgment.
As it stands, it appears to be incomplete. To complete it, we would arguably need to include certain kinds
of deliberative constraints such as those common to scientific inquiry—e.g., giving and responding to
reasons and evidence, following the evidence wherever it leads, an openness to new evidence and revision
of our judgments in light of that evidence, and in cases in which evidence is lacking, to adjudicate as best
as possible among conflicting positions; all plausible options should always remain on the table for future
consideration.

216
more available and reliable, and our understanding of all of the relevant causes and
effects increases. Indeed this is precisely what we have observed across the sciences over
the last two and a half centuries. As we acquire more knowledge, our predictive
judgments become more accurate. We gain greater agreement about the probable effects
of various objects—e.g., the effects of particular acts and traits. And we, consequently,
observe more agreement in our modal/causal judgments—e.g., the probability that an act
or trait will fulfill or foil particular ends and thereby cause particular emotional effects.
188

In sum, I suggest that any troubling ambiguity found in Hume’s MPV, or at least
any ambiguity so troubling as to undermine the MPV as a corrective artifice, does not
reside with the general rules that govern causal reasoning/belief. Let us therefore turn
our attention to the general rules that govern the sympathy process.
§VI.4 General rules, sympathy’s variability, & comparison
The corrective general principles discussed thus far act as norms governing belief-
formation about cause and effect. They tell us what ought to be believed regarding the
causal goings-on in the world. In the moral case, these general rules function as norms
governing a particular instance of causal reasoning/judgment, to wit: our
reasoning/judgments about what objects (e.g., acts or traits) fulfill or foil particular ends,
namely, moral ends. These are ends that, upon being fulfilled or foiled, have typical
emotional effects to which we sympathetically respond.

188
Of course, with respect to various causal judgments matters are more complicated because there might
be disagreement about what ends are pertinent in a given situation. Moreover, the set of pertinent ends may
vary across cultures/societies. If so, then this might give rise to an appearance of disagreement in judgment
across cultures when in fact there is no disagreement in judgment across cultures. (See part two, chapter
XI, of this project for further discussion of these matters.)

217
Hume also holds that the MPV has regulatory principles that govern the
psychological process of sympathy. So besides causal belief-guiding norms, the MPV
also has sympathy-guiding norms and hence motivation-guiding norms.
189
This second
set of norms is as equally important as the first because together these two sets of norms
account for the stability and uniformity that Hume claims the MPV brings to our moral
feelings and judgments.
As we consider Hume’s account of the general rules regulating sympathy, it is
worth repeating that the normative component governing our causal judgments/beliefs is
integral to, though distinct from, the normative component governing sympathy. As
discussed, many of our causal judgments directly impact our sympathetic feelings.
Hence the regulatory principles that ensure stability and uniformity in our causal
judgments are also indispensable for ensuring stability and uniformity in our sympathetic
and moral feelings, and thereby stability and uniformity in our motivation to act.
However, the principles that govern our causal judgments/beliefs are not
sufficient in themselves to bring stability and uniformity to our sympathetic and moral
feelings. This is significant because the success of Hume’s project and Humean projects
like it (e.g., this one) depends upon their ability to establish regulatory principles that can,
at least in principle, secure stability and uniformity in moral feeling. Otherwise there
would be no interpersonal agreement regarding our moral judgments, and hence no moral
objectivity.

189
As discussed, objects that elicit sympathetic feelings can also thereby elicit moral feelings, which have a
strong motivational impetus. So, by guiding sympathy, general rules guide our motivation to act morally.

218
In order to secure stability and uniformity of moral feeling, other principles are
needed to regulate sympathy’s natural variability, that is, its natural tendency to be overly
partial and indifferent. As discussed in chapters IV and V, our sympathetic feelings tend
to vary according to the degree of contiguity, resemblance, and causal (i.e., blood)
relations between others and ourselves. It is a matter of psychological fact, in Hume’s
view, that our sympathetic feelings are naturally stronger when we sympathize with those
we judge to be closer to us, and naturally weaker when we sympathize with those we
judge to be mere acquaintances. This variability in our sympathetic feelings is quite
significant because, as noted in chapter IV, Hume claims that “when a sympathy with
uneasiness is weak, it produces hatred and contempt” for others rather than a concern for
their wellbeing (T 385ff). In other words, sympathy’s natural variability can have a
direct impact on both our moral feelings and how we treat others.
I have also claimed in chapter IV (§3) that, according to Hume’s account,
consideration of others’ easiness—i.e., other’s pleasurable emotional states—can even
result in feelings of “hatred” and “contempt” for those individuals. Hume explains this
by appealing to our natural tendency to engage in a particular kind of comparison,
namely, comparison between our ideas of others’ situation and resulting emotional states
and our own situation and resulting emotional state. He writes:
In general we may observe, that in all kinds of comparison an object
makes us always receive from another, to which it is compar’d, a sensation
contrary to what arises from itself in its direct and immediate survey…
The direct survey of another’s pleasure naturally gives us pleasure; and
therefore produces pain, when compared with our own. His pain,
considered in itself, is painful; but augments the idea of our own
happiness, and gives us pleasure. (T 375-6)

219

When comparisons are made between ideas of our own emotional states and what
we judge to be others’ emotional states, such comparisons tend to produce feelings in us
that are the opposite of the sympathetic feelings we experience in response to the feelings
we judge others to have. Suppose, for example, that I go to the hospital and visit a
member of my own military regiment who loses both of his legs in a roadside bombing. I
sympathetically feel pain for that person and the realities that he must now face.
However, a pleasurable feeling (e.g., relief) is likely to be elicited in me upon considering
the fact that I, unlike the other person, am fortunate enough to still have both of my legs.
Thus, sympathy and comparison produce what are, in effect, contrary outcomes.
Whereas sympathy amounts to a tendency to identify with others and take their pains and
pleasures upon oneself, this sort of comparison amounts to a tendency to compare others’
feelings with one’s own; and when other’s feelings contrast with our own, this tends to
augment our sympathetic feelings and produce feelings in us that are contrary to them. In
short, this kind of comparison operates on the feelings aroused by sympathy and produces
feelings that oppose those sympathetic feelings. Hume sums up this account as follows:
We judge more of objects by comparison, than by their intrinsic worth and
value; and regard every thing as mean, when set in opposition to what is
superior of the same kind. But no comparison is more obvious than that
with ourselves; and hence it is that on all occasions it takes place, and
mixes with most of our passions. This kind of comparison is directly
contrary to sympathy in its operation… In all kinds of comparison an
object makes us always receive from another, to which it is compar’d, a
sensation contrary to what arises from itself in its direct and immediate
survey. The direct survey of another’s pleasure naturally gives us
pleasure; and therefore produces pain, when compar’d with our own. His
pain, consider’d in itself, is painful; but augments the idea of our own
happiness, and gives us pleasure. (T 593-4)

220

This explains why consideration of others’ easiness can sometimes result in
feelings of hatred and contempt toward those individuals. But can it also explain why we
are much less likely to experience pleasure at our family and/or friends’ bad fortune than
we are to experience these feelings toward others? Of course we can, and unfortunately
sometimes do, have such feelings directed toward our family and friends, but it is not as
common an occurrence as it is when the person(s) under consideration is not a
friend/family. Can Hume’s account explain why this is so?
It can. The answer lies, once again, in the force and vivacity of the relevant idea
and corresponding feeling. As noted in §IV.2, our ideas of others’ pleasure and/or pain
are enlivened by an impression of the self. Hence the force and vivacity of those ideas
tend to be proportional to our conception of others’ relations to ourselves. This is why
our sympathetic feelings tend to vary according to the degree of contiguity, resemblance,
and causal (i.e., blood) relations between others and ourselves.
190
These feelings tend to
strengthen or weaken according to others’ psychological proximity to oneself. According
to Hume, such variation in sympathy affects comparison because comparison operates on
the feelings aroused by sympathy. For instance, Hume writes: “…if the idea [of
another’s emotional state] be too feint, it has no influence by comparison; and on the
other hand, if it be too strong, it operates on us entirely by sympathy, which is the
contrary to comparison” (T 594-5).

190
Recall that we are also prone to form “unphilosophical general rules”, i.e., prejudicial causal judgments.
These prejudices can inform our judgments of resemblance by leading us to focus on accidental features,
such as the color of one’s skin, instead of more essential features, such as a shared psychology or shared
interests, ends, and goals. This, in turn, can also directly impact our sympathetic feelings.

221
So Hume’s explanation is roughly as follows. Since comparison operates on the
feelings aroused by sympathy, the idea of another’s emotional state must be sufficiently
vivacious to influence the comparison and enliven the relevant opposing feeling;
otherwise there would be no feeling on which to operate. If, on the other hand, the idea is
too vivacious, then sympathy converts that idea into such a strong feeling in oneself that
it tends to overpower any opposing feeling that may be produced by comparison. Thanks
to sympathy’s natural variability, we tend to experience stronger sympathetic feelings for
our family and friends than others. Consequently, one’s sympathetic feelings for one’s
family and friends are more likely to overpower any opposing feeling that may be
produced by comparison with oneself. In fact, if our sympathetic feelings are strong
enough, it is unlikely that we will even engage in such comparisons.
191

In this way, Hume explains why we are less likely to experience pleasure at our
family and/or friends’ bad fortune than the bad fortune of our enemies, complete
strangers, or mere acquaintances. Assuming we do engage in such comparisons between
our family and/or friends’ emotional states and our own, it highly unlikely that any
resulting feelings of hatred and contempt will be strong enough to outweigh our
sympathetic feelings for our family and/of friends. On the other hand, our sympathetic
feelings for mere acquaintances or complete strangers will tend to be much weaker.
Hence contrary feelings that result from comparing their emotional states to our own are
more likely to compete with those sympathetic feelings.

191
As discussed in§IV.2, a sufficiently strong sympathy can create an extensive concern for the other. As a
result, strong sympathetic feelings will tend to counteract any inclination one might have in this case to
compare one’s own lack of pain with another’s pain because an extensive concern for the other puts one’s
focus on the other rather than on the self.

222
§VI.4.1 Corrective general rules: a parallel approach
Our main task in this chapter is to provide an account of the general rules that
regulate sympathy’s natural variability—i.e., the normative principles that regulate
sympathy’s tendency to be overly partial and indifferent—because such fluctuation in
sympathy undermines interpersonal agreement in feeling, which is requisite for stability
and uniformity in moral judgment. Furthermore, given that various comparisons between
others and ourselves can produce feelings that run contrary to sympathy, such
comparisons also tend to undermine agreement in feeling, as well as stability and
uniformity of judgment. Hence the general rules that govern sympathy’s variability will
also need to somehow regulate such comparisons.
Hume’s general approach to addressing sympathy’s natural tendency to be
excessively partial and indifferent, as well as our tendency to engage in various
comparisons that undermine sympathy, parallels his approach to addressing our natural
error-prone tendencies associated with causal reasoning. Just as with the error-prone
tendencies associated with our causal judgments, Hume claims that we must use our
imagination and reflection/understanding to combat the natural tendencies of sympathy to
be overly partial and indifferent, as well as our tendency to engage in various
comparisons. This imaginative and reflective exercise enables us to, in effect, abstract
away from our immediate situation and correct any contrary “momentary appearances.”
As discussed, Hume claims that “such corrections are common with regard to all
the senses” (T 582). Consider, for example, our judgments of color. When judging the
color of an object, we sometimes need to abstract away from our immediate experience,

223
as for instance when lighting conditions fluctuate. When this occurs we do not say that
the color of the object changes with the fluctuations in our color impressions. We judge
that the color remains the same irrespective of such variations. For instance, using my
imagination governed by the understanding, I can judge that the strawberries in the bowl
are red even if they don’t currently appear red to me; e.g., perhaps I observe the
strawberries through a green colored glass jar.
Of course there is an important disanalogy with our color judgments insofar as the
corrections we make via our imagination and reflection can only correct our judgments;
i.e., they do not affect our impressions. In the moral case, however, our corrections often
do affect our impressions. This is due to the fact that our sympathetic and moral feelings
are responsive to our causal judgments. Put in Hume’s terms, our sympathetic and moral
feelings are impressions of reflection, whereas our color impressions are impressions of
sensation. It is precisely because our sympathetic and moral feelings are impressions of
reflection that reflection and imagination can directly affect them.
192

When this imaginative and reflective corrective exercise becomes a matter of
habit and custom, we will habitually experience an extensive sympathy, which gives rise
to moral feelings and extensive benevolence. Hence there will be no need for conscious
correction because the corrective general rules are applied unconsciously. Recall from
the previous section that this can happen in the case of our causal judgments. Corrective

192
See §III.3 for further discussion of impressions of reflection and impressions of sensation.

224
general rules can be applied habitually.
193
Habitual application of both sets of corrective
general rules—i.e., taking up the MPV as a matter of custom—is arguably one of the
most important virtues because application of those rules is the only means by which an
extensive sympathy and moral sentiments are engendered.
However, just as with our causal judgments, sometimes our natural tendencies are
not so easily eliminated. We often must make a conscious effort to exercise our
imagination and reflection to exorcise our bad habits. In order to secure stability and
uniformity (and thereby avoid various practical problems), sometimes the best we can
hope for in these cases is mitigation of our undesirable natural tendencies and the
contrary feelings associated with them. In the case of our causal judgments, we typically
need to consciously exercise our imagination and reflection to mitigate our tendency to
inadequately distinguish between accidental features and actual causes, as well as our
tendency to over or underestimate the effects. In the case of sympathy, we typically need
to consciously use our imagination and reflection to curb the excessive partiality and/or
indifference in our sympathetic feelings, as well as our tendency to consider various
relations and comparisons between others and ourselves.

193
In discussing our corrective tendencies, Hume also appeals to an analogy with distance perception in
which objects further away appear smaller in our visual field, but we do not judge that their actual size has
changed (T 603). In this case, the “corrections” we make are clearly a matter of custom or habit established
in early infancy. Hence they almost always occur unconsciously. As the distance between objects and us
increases, the objects don’t simply appear smaller in our visual field; they actually appear to be further
away from us. We can create optical illusions by taking advantage of this fact. For example, we can make
an object appear to be much larger than it is by making it appear as if the object is farther away from us
than it really is. As we will briefly discuss in part two, there is evidence that at least some of the MPV’s
corrective principles have been unconsciously selected—viz., through natural selection.

225
In sum, just as with causal reasoning, sometimes we can use our imagination and
reflection to eliminate our error-prone tendencies;
194
sometimes we can only mitigate
these tendencies; and sometimes we fail to do either. Regardless, we can still know from
reflection what sympathetic and (thereby) moral feelings we would have were we to
eliminate, or at least effectively mitigate, sympathy’s natural tendencies for excessive
partiality and indifference. Hume summarizes much of this in the following passage.
When we form our judgments of persons, merely from the tendency of
their characters to our own benefit, or to that of our friends, we find so
many contradictions to our situation, that we seek some other standard of
merit and demerit, which may not admit of so great variation. Being thus
loosen’d from our first station, we cannot afterwards fix ourselves so
commodiously by any means as by a sympathy with those, who have any
commerce with the person we consider. This is far from being as lively as
when our own interest is concern’d, or that of our particular friends; nor
has it such an influence on our love and hatred: But being equally
conformable to our calm and general principles, ‘tis said to have an equal
authority over our reason, and to command our judgment and opinion. We
blame equally a bad action, which we read of in history, with one
perform’d in our neighborhood t’other day: The meaning of which is, that
we know from reflexion, that the former action wou’d excite as strong
sentiments of disapprobation as the latter, were it plac’d in the same
position. (T583-4)
195

194
As we shall discuss in the following chapter, eliminating the error-prone tendencies of sympathy does
not amount to eliminating the totality of sympathy’s natural variability.

195
Another key passage in which Hume discusses the use of reflection to issue moral judgments in those
instances in which we fail to properly mitigate sympathy’s variability is the following:
The case is here the same as in our judgments concerning external bodies. All objects
seem to diminish by their distance: But tho’ the appearance of objects to our senses be the
original standard, by which we judge of them, yet we do not say, that they actually
diminish by the distance; but correcting the appearance by reflexion, arrive at a more
constant and establish’d judgment concerning them. In like manner, tho’ sympathy be
much fainter than our concern for ourselves, and a sympathy with persons remote from us
much fainter than that with persons near and contiguous; yet we neglect all these
differences in our calm judgments concerning the characters of men. Besides, that we
ourselves often change our situation in this particular, we every day meet with persons,
who are in a different situation from ourselves, and who cou’d never converse with us on
any reasonable terms, were we to remain constantly in that situation and point of view,
which is peculiar to us. The intercourse of sentiments, therefore, in society and

226
The questions now under consideration are: (1) What exactly are these “calm and
general principles” that govern sympathy and comparison? And (2) do those principles
regulate sympathy and comparison in such a way so as to produce general motivational
agreement among those who satisfy them? That is, what reason do we have to think that
moral agents will share the same basic moral attitudes of approval or (as the case may be)
disapproval towards a given object (e.g., an act or trait) upon satisfying the principles
stipulated by Hume? Hume’s account clearly must secure such interpersonal agreement
in feeling, for without this interpersonal agreement, we cannot attain the desired stability
and uniformity in our moral judgments. Moral objectivity will fall by the wayside.
Hence providing an answer to both of the above questions is crucial for any defense of
Hume’s moral ontology and its foundations. We shall now address these and other
important questions in the following chapter.

conversation, makes us form some general inalterable standard, by which we may
approve or disapprove of characters and manners. (T 603)

227
Chapter VII: The disinterested moral point of view (continued)

§VII.1 General rules & the regulation of sympathy and comparison

As we have seen, Hume is well aware that how we approach a situation, how we
think about it and evaluate it, will affect our feeling-responses to that situation.
Moreover, he also acknowledges that the feelings we experience prior to a situation can
affect how we think about it, evaluate it, and then respond to it. And how we think about
a situation, how we evaluate it, and respond to it can either promote or prevent stability
and uniformity of judgment and feeling, and this in turn has various practical
consequences. For this reason, Hume argues that we seek out, construct, and take up
“steady and general points of view”.
The MPV is one instance of such steady and general points of view. The general
rules that constitute the MPV are put in place to eliminate, or at least effectively palliate,
the prejudicial tendencies in our causal reasoning, sympathy’s tendency to be indifferent
or overly partial, and our tendency to consider various relations and comparisons between
others and ourselves. The extent to which we successfully apply these corrective general
rules will be the extent to which our feelings reliably indicate the moral facts. In other
words, our feelings can be more or less reliable because satisfaction of the MPV can be a
matter of degree.
196
If we fail to regulate these tendencies, then our moral feelings, which
arise due to our sympathetic feelings, will not reliably indicate moral distinctions.
Moreover, our moral (and sympathetic) feelings will tend to be too weak to counteract

196
Others have also noted this point; see, for instance, Radcliffe (1994), Cohon (1997), Abramson (1999),
and Wallace (2002).

228
any contravening feelings we might experience, and in some cases we may not
experience any moral (or sympathetic) feelings at all when the property is present.
Hume’s general approach for regulating sympathy’s natural variability, as well as
various comparisons between others and ourselves, is clear. According to Hume, we
regulate sympathy’s natural variability by imposing principles that regulate our
tendencies to focus on ourselves and on various relations between others and ourselves.
Not only do these tendencies produce partiality and indifference in our sympathetic
feelings. Our tendency to focus on ourselves and our own interests and ends often gives
rise to various comparisons between others and ourselves, and this frequently produces
feelings that are contrary to sympathy. Consequently, when sympathizing and judging
from the MPV, Hume claims that, as moral appraisers, we are to “consider not whether
the persons affected…be our acquaintance or strangers, countrymen or foreigners. Nay,
we over-look our own interest in those general judgments” (T 582). In short, the MPV
ensures disinterestedness; for it is “only when a character is considered in general,
without reference to our particular interest, that it causes such a feeling or sentiment, as
denominates it morally good or evil” (T 472). Hume stresses this point once again:
Now, in judging of characters, the only interest or pleasure, which appears
the same to every spectator, is that of the person himself, whose character
is examin’d; or that of persons, who have a connexion with him. And tho’
such interests and pleasures touch us more faintly than our own, yet being
more constant and universal, they counter-balance the latter even in
practice, and are alone admitted in speculation as the standard of virtue
and morality. They alone produce that particular feeling or sentiment, on
which moral distinctions depend. (T 590-1)

229
Unfortunately, unlike the eight general rules that govern causal reasoning, Hume
does not explicitly list specific general rules that govern sympathy (and comparison). For
this very reason, the charge that Hume’s MPV is too vague or ambiguous to be helpful as
a corrective device for our moral feelings and judgments is not easily dismissed. In an
attempt to answer this charge, our present aim is to further unpack Hume’s conception of
the general rules that govern sympathy.
I suggest that the general rules regulating sympathy ensure uniformity and
stability of moral feelings and moral judgment in two fundamental ways. First, the MPV
places restrictions upon moral appraisers with respect to the group with whom they are to
sympathize. Second, it places restrictions on moral appraisers regarding how they are to
sympathize with that group. Consider, for instance, the passage below in which Hume
reveals important clues about the principles that are supposed to regulate sympathy.
Concerning those with whom we are to sympathize from the moral point of view,
Hume writes:
When experience has once given us a competent knowledge of
human affairs, and has taught us the proportion they bear to human
passion, we perceive, that the generosity of men is very limited, and that it
seldom extends beyond their friends and family, or, at most, beyond their
native country. Being thus acquainted with the nature of man, we expect
not any impossibilities from him; but confine our view to that narrow
circle, in which any person moves, in order to form a judgment of his
moral character. When the natural tendency of his passions leads him to
be serviceable and useful within his sphere, we approve of his character,
and love his person, by a sympathy with the sentiments of those, who have
a more particular connexion with him. (T 602-3)

And concerning how we are to sympathize with that group, Hume immediately
goes on to say:

230
We are quickly oblig’d to forget our own interest in our judgments of this
kind, by reason of the perpetual contradictions, we meet with in society
and conversation, from persons that are not plac’d in the same situation,
and have not the same interest with ourselves. The only point of view, in
which our sentiments concur with those of others, is, when we consider
the tendency of any passion to the advantage or harm of those, who have
any immediate connexion or intercourse with the person possess’d of it.
And tho’ this advantage or harm be often very remote from ourselves, yet
sometimes ‘tis very near us, and interests us strongly by sympathy. This
concern we readily extend to other cases, that are resembling; and when
these are very remote, our sympathy is proportionably weaker, and our
praise or blame fainter and more doubtful. (Ibid)

Thus, in taking up the MPV, we are to “confine our view” to the agent’s “narrow
circle…in order to form a judgment of his moral character.” We are also to ignore our
own interests and disinterestedly sympathize “with the sentiments of those…who have
any immediate connexion or intercourse with the person” being evaluated. Hume also
says that we are to take into account not simply the actual emotional effects on that group
but the tendency of the agent’s act or trait to cause particular emotional effects in that
group. More precisely, we are to determine the agent’s act or trait’s causal tendency to
fulfill or foil ends that tend to produce emotional effects on those within the agent’s
“sphere”. When the emotional effects are pleasurable, we deem the agent’s act or trait
useful or immediately agreeable and approve of it through extensive sympathy. When
the effects are painful, we deem the act or trait pernicious or immediately disagreeable
and disapprove of it through extensive sympathy.
§VII.1.1 A problem with Hume’s restriction to the narrow circle
Hume’s restriction to the agent’s “narrow circle” is supposed to help regulate
sympathy’s natural partiality and indifference. Consequently, a key issue concerns how

231
Hume conceives of the agent’s “narrow circle”? For instance, how narrow is Hume’s
conception of the narrow circle? Does it include only those who are in close
psychological proximity to the agent, or does it also include others who are further
away?
197
Hume does, after all, speak of sympathizing with those typically affected by the
agent’s act or trait, and presumably those typically affected will at least sometimes, if not
often, include those who are neither “causally related” (i.e., blood related) nor
“contiguous” to the agent (i.e., those that have a well established close relationship with
the agent), just as it might also include those who lack various “resemblances” to the
agent. There are also other important issues concerning how Hume conceives of the
narrow circle. For instance, does he conceive of the narrow circle as the agent’s actual
narrow circle or an idealized conception of that group?
198

Regardless, it is not at all obvious how Hume’s narrow circle, no matter how he
conceives of it, can accomplish its aim. As Kate Abramson (1999: 338) correctly points
out,
199
none of the competing conceptions of Hume’s “narrow circle” appear to resolve
the problem because no matter how narrowly or widely one conceives of the agent’s
narrow circle, different appraisers will still lie at varying psychological distances to those
with whom they are to sympathize.
200
In other words, the extent to which the appraiser

197
For instance, contrast Brown (1988), Darwall (1994), and Radcliffe (1994) with Baier (1991), Brown
(1994), Sayre-McCord (1994), and Loeb (2004a).

198
Contrast Brown (1994) with Baier (1991), Radcliffe (1994), Abramson (1999), and Loeb (2004a).

199
Abramson claims that her observation is merely an extrapolation of Darwall’s (1994, n. 37) claim.

200
Among the accounts that appear to fall prey to her claim are: Stewart (1963), Ardal (1966), Mercer
(1972), Mackie (1980), Capaldi (1989), Baier (1991), Brown (1994), Darwall (1994), Radcliffe (1994),
Sayre-McCord (1994), Cohon (1997), and Wallace (2002). Loeb (2004a) is a notable exception.

232
deems the members of the agent’s narrow circle to be less “causally related”,
“contiguous”, and/or less “resembling” to the appraiser will presumably parallel the
extent to which the appraiser’s sympathetic feelings will be excessively partial and/or
indifferent to the members of that group. Hence it’s far from clear how Hume’s appeal to
the agent’s narrow circle can serve to prevent the appraiser’s sympathy from exhibiting
excessive partiality and/or indifference with respect to that circle.
Since the variability of an appraiser’s sympathy will be proportional to his or her
psychological proximity to the agent’s narrow circle, the appraiser’s moral feelings of
approval or (as the case may be) disapproval towards the agent’s act or trait will also vary
in proportion to the appraiser’s psychological distance from the agent’s narrow circle.
And since different appraisers will be positioned at different distances from the agent’s
narrow circle—i.e., since they will have different contiguity, resemblance, and causal
(blood) relations to the members of the narrow circle—appraisers will therefore not
experience the same moral feelings towards the agent.
201
Hence they will presumably not
issue the same moral judgments about the agent as other appraisers.
Recall that, on Hume’s view, appraisers must experience roughly the same degree
of approval or (as the case may be) disapproval if they are to judge that the acts or traits
of two agents equally merit praise or (as the case may be) blame. “We blame equally a
bad action, which we read of in history, with one perform’d in our neighborhood t’other
day: The meaning of which is, that we know from reflexion, that the former action wou’d

201
Erin Kelly (2004) also discusses the same basic problem for Hume’s view, although she raises it
specifically as a criticism of Louis Loeb’s (2002) proposal that Hume holds a stability account of
normativity.

233
excite as strong sentiments of disapprobation as the latter, were it plac’d in the same
position” (T584). It follows from this that different appraisers, when sympathizing from
the disinterested moral point of view, must experience the moral sentiments to roughly
the same degree if they are to agree in their moral judgments of a given agent’s act or
trait. Hume observes that we not only “give the same approbation to the same moral
qualities in China as in England”, those qualities also “recommend themselves equally to
the judicious spectator” (T 581). In other words, in Hume’s view, the MPV must
guarantee that every “judicious spectator”—i.e., everyone who meets all of the MPV’s
conditions—will not only experience the same kind of moral sentiment; they will all also
experience it to roughly the same general degree. Hence the seemingly inevitable
extreme variability among appraisers’ sympathetic feelings toward the members of an
agent’s narrow circle presents a serious problem.
202

An interlocutor might suggest that we can easily resolve this problem by
idealizing—i.e., abstracting away from—the agent’s narrow circle such that we conceive
of the “narrow circle” as anyone’s narrow circle. We can then effectively eliminate the
variability among appraisers’ sympathy by removing all of the contexts—e.g., the agent’s
social-context and the identity of the members of the agent’s narrow circle—by which
appraisers lie at varying proximities from the narrow circle. Unfortunately, there are two
insurmountable obstacles that prevent us from adopting this solution as Hume’s. First,

202
Abramson (1999: 339-40) also discusses a related problem, to wit: differences in the degree of felt
sympathy can amount to differences in the kind of sentiments produced. For instance, as we’ve already
discussed in chapter IV, Hume claims that the degree, or vivacity, of feeling can make the difference
between feeling anger towards a person rather than pity; this, in turn, can amount to the difference between
experiencing hate rather than love towards the person, and thus disapproval rather than approval (T 381-9).

234
Hume explicitly states that we are to sympathize with “the agent’s narrow circle” rather
than just anyone’s narrow circle. This is in part because, as Abramson points out, “Hume
recognizes that some traits have different typical effects, depending on the particular
social setting in which the agent acts on that trait, and he claims that we do sometimes
properly take account of such differences in making moral evaluations” (1999: 338-9).
203

Second, Abramson’s point can also be used to demonstrate that the interlocutor’s
conception of the narrow circle would actually give faulty results in many cases.
Because some traits tend to produce different effects in different social settings, there
may be no fact about what effects those traits would tend to produce independent of their
being placed within a particular social context. Consequently, if appraisers were simply
to conceive of the agent’s narrow circle as just anyone’s narrow circle, then they would
lack information about the typical effects of some traits that is requisite to produce moral
feelings towards agents who possessed those traits. According to this conception of the
narrow circle, there would be no moral facts to judge in many instances in which there
arguably are moral facts.
How, then, can Hume’s restriction to the agent’s narrow circle help ensure
agreement in both the kind and degree of appraisers’ moral feelings toward the agent’s
act or trait? This is requisite if the MPV is to serve as a shareable point of view that
produces stability and uniformity of moral feeling and judgment. Perhaps unsurprisingly,
I suggest that the only way to resolve the problem is to appeal both to a particular

203
In support of her claim, Abramson cites the following passages in the Enquiry Concerning the
Principles of Morals: E.336.29-338.5; E.241.29-242; E.248.21-249; and E.262.21-262.31.

235
conception of the agent’s narrow circle—i.e., those with whom we are to sympathize—
and a particular account of how appraisers are to sympathize with that group.
In order to explain how this approach can resolve the problem we need to give a
more detailed account of these restrictions. Abramson seems to agree with this approach.
However, she offers an interpretation of Hume’s account of both those with whom we are
to sympathize and how we are to sympathize with that group that is quite different from
the interpretation to be proposed by this project. In the proceeding section, we will
consider her interpretation of these restrictions, and then in the section that follows it I
shall critique her account and use it as an expository device to motivate the interpretation
of Hume’s restrictions that I favor.
§VII.2 Abramson’s interpretation
Both Abramson and I agree that extensive sympathy is a two-stage imaginative
exercise. In the first stage, appraisers assess the situation. This amounts to using causal
reasoning to determine the typical emotional effects of a given agent’s act or trait, in
particular the typical emotional effects on those with whom appraisers are to sympathize
from the MPV. In the second stage, appraisers’ ideas of the emotional states of those
with whom they are to sympathize from the MPV are converted via extensive sympathy
into like impressions of pleasure or (as the case may be) pain within the appraisers.
204

On Abramson’s reading, those with whom appraisers are to sympathize from the
MPV are “those who are typically direct recipients of that trait’s usual effects” (343).

204
Others have offered an alternative four-stage reading of Hume in which there are two sympathetic
stages. The first stage is unregulated, whereas the second stage is regulated. See, for instance, Stewart
(1963), Mackie (1980), Baier (1991), Radcliffe (1994), and Sayre-McCord (1994).

236
She goes on to claim that the range of persons picked out by this restriction is determined
by two factors: (i) the particular trait under consideration and (ii) various features of the
social world in which the trait is expressed (343-4).
205
Depending upon the particular
trait in question, the MPV can be more or less inclusive. In principle, the affected group
may be conceived more narrowly up to the point of including only the agent’s immediate
family and/or friends, or it may be conceived more widely perhaps up to the point of
including all of humanity. Abramson’s examples are the character traits of loyalty and
charity: “loyalty is typically expressed to a relatively more narrow group than…charity”
(343). The other factor determining the range of the particular group with which we are
to sympathize from the MPV is the particular social context in which the particular trait is
expressed. Some character traits, when manifested in different social/cultural settings,
will affect a wider or narrower range of people (and/or other sentient beings).
In short, although the range of those with whom we are to sympathize from the
MPV can vary on Abramson’s reading, Hume’s restriction to the agent’s “sphere” of
“connection”—i.e., the “narrow circle in which [the agent] moves”—nevertheless
illustrates the fact that, once a particular trait and specific social circumstances are
identified, the MPV sets “definite parameters for our sympathetic survey” (343).
206
We
are only to sympathize with those who are typically direct recipients of a particular trait’s

205
Sayre-McCord makes a similar suggestion when he notes that: “Which group comes within our view
depends on the context and character of the person being judged” (1994, fn 24).

206
Following Baier (1991) and others, Abramson appears to view the agent’s “narrow circle” as including
only the agent’s close companions. But Hume’s text seems to suggest that the agent’s “narrow circle” and
the agent’s “sphere of connection” are meant to identify the same group of individuals, regardless of how
narrowly or widely that group is conceived. Since this identification has no bearing on Abramson’s claims,
I shall treat the agent’s narrow circle as picking out the same group as those who have commerce with the
agent, the agent’s sphere of connection, etc.

237
usual effects within a specific social context. It is noteworthy that the specific social
context is not only crucial for determining those who are typically affected by a particular
trait. It also plays a key role in determining the usual effects of a trait. As both
Abramson and Hume observe, when traits are expressed in different social/cultural
settings, their typical effects may differ.
207

Abramson readily admits that, as things stand, this account does not resolve all of
the problems due to sympathy’s variability. Although we must consider the particular
trait to be evaluated, its typical effects, the particular social context in which the trait is
expressed, and those typically affected, it is not clear how this sufficiently regulates
sympathy’s variability due to: (a) certain peculiarities in the agent’s circumstances; (b)
peculiarities in those typically affected by the trait; (c) causal (i.e., blood) relations
between appraisers and those typically affected; (d) the contiguity between appraisers and
those typically affected; (e) relations of resemblance between appraisers and those
typically affected; and (f) the relative specificity of appraisers’ ideas.
Abramson holds that certain elements of “idealization” are needed to sufficiently
regulate these sources of variability so as to produce the requisite interpersonal
agreement.
208
To ensure uniformity of moral feeling and judgment, the MPV requires
that we idealize our conception of: (1) the overall group typically affected by the trait, (2)

207
As we shall discuss in the next section, traits might also have different typical effects depending upon
the contiguity, resemblance, and causal relation (i.e., blood relation) that those affected bear to the agent.
Whether the trait under evaluation does or not, and if so to what extent it does, may also vary across social
contexts. Later I shall argue that this fact spells trouble for Abramson’s interpretation.

208
With respect to the present discussion, idealization, for Abramson, appears to be a matter of using our
imagination to regulate variability.

238
the individuals that compose the group, (3) the usual effects of a particular trait, and (4)
the character trait itself.
Consider first sympathy’s variability due to (a) certain peculiarities in the agent’s
circumstances. The relevant peculiarities pertain to those instances in which it just so
happens that no one is around to be affected by the agent’s trait. Recall from §IV.7
Hume’s point that “virtue in rags is still virtue.” Of course, as just discussed, the MPV
requires that appraisers are to conceive of those typically affected by the agent’s trait
rather than those, if any, who are actually affected. In other words, appraisers must use
their imagination to conceive of the group to whom the agent would typically express the
trait (given the particular social context), and by doing so appraisers thereby exclude the
aforementioned peculiarities in an agent’s circumstances. Hence the MPV requires
idealization of (1) the group.
209

With respect to our idealization of (2) the individuals that compose the group,
matters are a bit more complicated. According to Abramson, we need to use our
imagination to exclude (b) various peculiarities of an individual typically affected by an
agent’s character trait such as, for example, an individual’s abnormal response to the
trait.

Abramson has in mind cases such as, for example, when the agent is witty “but has
rather dull friends, so the effects of his wit are typically painful” (344, her emphasis). In
this case, although the effects are typically painful for the agent’s actual friends, we still
do not disapprove of his wit. We might disapprove of the agent’s insensitivity to his
friends’ feelings, but we do not disapprove of his trait.

209
Abramson credits Brown (1994) for pointing out the necessity of idealizing the group in this way.

239
Abramson claims that we do not disapprove of the agent’s trait in this case
because we recognize that the effects on his friends’ feelings are atypical insofar as his
actual friends’ reaction to the trait is atypical. In other words, we recognize that those
individuals actually affected are atypical. The MPV addresses this problem by requiring
appraisers to idealize the individuals typically affected by the agent’s character trait, and
conceive of those individuals as psychologically normal (i.e., typical) even if they in fact
happen to be psychologically abnormal (i.e., atypical). This idealization thereby excludes
various peculiarities of any individuals typically affected by an agent’s character trait.
Abramson claims that we must also idealize our conception of the individual in
order to correct for variability of sympathy that results from appraisers’ awareness of (c)
any causal (i.e., blood) relations between individuals typically affected and appraisers.
According to Abramson, the MPV accomplishes this by requiring appraisers to use their
imagination to exclude individuals’ particular location within their social world (345).
Abramson offers the following example an as explanation of this restriction.
Suppose, for instance, that I wonder about the value of my own parental
affection. To determine the value of that affection, I must conceive of
those directly affected by that trait as standing in a particular relationship
to one with that trait, namely, the child-parent relationship. But in so
doing, I do not conceive of them as my children, for that would be to give
their particular location within their social context. Rather, I conceive of
them just as someone’s children (in a specific social context). This
controls for the variability of sympathy which arises from the issue of
whether or not one is blood-related to those with whom one sympathizes,
because in doing so, we do not conceive of those who are directly affected
by any agent’s trait as standing in any particular relationship to us as
appraisers. (345, Abramson’s emphasis)

240
This brings us to sympathy’s variability due to (d) any contiguity between
appraisers and those typically affected by the agent’s trait. Abramson claims that the
key to excluding the variability in sympathy due to relative contiguity “involves how we
conceive of the relevant usual effects in relation to those who have commerce with the
agent” (345, her emphasis). In other words, it involves idealization of appraisers’
conception of (3) a particular trait’s usual effects. Abramson offers the following
illustration. Suppose that misery is the usual effect of a particular trait on those to whom
the agent typically expresses that trait. She considers two general ways in which we
might conceive of this misery: (i) the misery of those with whom the agent has
commerce, or (ii) the misery usually produced by this trait.
Abramson dismisses the first option for two reasons. First, according to
Abramson, that approach does not resolve the variability problem because the relation
between the appraiser and those with whom the agent has commerce will be variable.
Consequently, the appraiser’s sympathetic feelings will vary accordingly. Second, she
claims that the resulting sentiments could not count as moral sentiments because these
feelings are inevitably directed towards those who have commerce with the agent rather
than the agent himself or his trait (346). Her reasoning is as follows: “if the idea one has
is ‘the misery of those in the agent’s narrow circle,’ then the object of that idea is those in
the narrow circle, the subject is the circumstances of those in the narrow circle that are
miserable, and the quality is that misery. The sentiment that results when that idea is
enlivened into an impression will take the narrow circle as its object as well” (358, n 46).
The problem, according to Abramson, is that moral sentiments must take either the agent

241
or the agent’s act or trait as its object, not the agent’s narrow circle. (I shall challenge
Abramson’s conclusion in the following section. See also §V.5)
Both of these reasons lead Abramson to the second option, namely: conceiving of
the misery as “the misery usually produced by the trait”. Interestingly, she admits that
the second option implies the first option, to wit: conceiving of it as “the misery of those
with whom the agent has commerce” (346). This is due to the fact that we acquire our
conception of the misery usually produced by the trait by observing or contemplating the
misery produced in those individuals with whom the agent has commerce, where the
individuals are, of course, conceived of as psychologically normal. However, she goes
on to suggest that there is also a sense in which the first option is not implied. This is this
sense in which, when we conceive of the misery as the typical effect of the trait, we are
conceiving of it as “misery that can be anyone’s, including mine, were I in her [viz., the
agent’s] social circumstances and directly affected by her trait” (ibid, emphasis added).
In other words, on Abramson’s interpretation, both the agent and the agent’s
narrow circle are only important insofar as they help us conceive of the typical effects of
a particular trait. Although we are to idealize both the group and the individuals that
compose the group, once we have acquired the conception of the trait’s typical effects,
that conception includes neither the agent nor the group. Consequently, when
sympathizing with the typical effects of the trait from the MPV, the appraiser’s relation
of contiguity with both the agent and those individuals typically affected by the agent’s
trait is excluded. In short, Abramson holds that by requiring that appraisers idealize the

242
idea of a trait’s typical effects, the MPV restricts appraisers’ purview merely to the
typical effects of a particular trait (within a particular cultural context).
It is noteworthy that, according to Abramson, this is the point of Hume’s claim
that the MPV requires that we “approach equally near” the agent (T 582). By
approaching equally near the agent, we in effect conceive of the typical effects of the trait
as the typical effects upon anyone, that is, the typical effects that the trait would have on
anyone were they to approach equally near anyone who possesses that trait (in a
particular social context). In this way, the idealization of the idea a particular trait’s
typical effects excludes the specific identity of the agent and the agent’s narrow circle.
Abramson uses this approach to help eliminate as a source of sympathy’s
variability (e) resemblance between appraisers and those typically affected. Since the
MPV disallows appraisers from considering the identity of either the agent or those
typically affected by the agent’s trait, Abramson claims that the only sort of resemblance
remaining is the appraiser’s resemblance to the trait itself (347). Although it may
initially seem odd to think of oneself as resembling a trait, Abramson claims that Hume’s
view entails this resemblance. She writes: “In a variety of ways for Hume, to think of
‘myself’ is essentially to think in terms of my character” (347). Consequently, some
traits can more closely resemble my character than others, and some traits can be totally
foreign to my character. Abramson holds that variability due to resemblance is therefore
still a problem. If I more closely identify with a particular trait than other appraisers, then
I’m more likely to experience more intense sympathetic feelings and hence moral

243
feelings than other appraisers. Consequently, we will issue different moral judgments
from the MPV.
Abramson attempts to redress this aspect of sympathy’s variability by restricting
how appraisers conceive of the trait’s relation to themselves. In other words, appraisers
are to idealize their conception of (4) the trait itself. On Abramson’s interpretation, we
once again “approach equally near” the agent and conceive of the trait as one to which we
are “all subject simply in virtue of our humanity… In this way, we can understand the
traits that are the objects of our moral evaluations as ones which we all, as appraisers,
equally ‘resemble’” (347). In short, by idealizing the trait in this way, appraisers in effect
abstract away from their own identity since all appraisers equally resemble all traits from
within the MPV. The idealization required by the MPV thereby also eliminates all
relations of resemblance as a source of sympathy’s variability.
This approach also helps Abramson resolve the final source of sympathy’s
variability, namely, (f) relative variability in the specificity or generality of the
appraisers’ ideas of a particular trait’s usual effects. The problem is that different
appraisers might have a more or less specific/general idea of the trait’s typical effects.
Since Hume holds that more specific ideas are more vivid than less specific (i.e., more
general) ideas, this variation in the vivacity of appraisers’ ideas will lead to variation in
appraisers’ sympathetic feelings and hence moral feelings.
For example, Abramson (348) points out that one appraiser might conceive of the
usual effects of loyalty to one’s lover, while another appraiser might conceive of the
usual effects of loyalty to one’s colleagues. The salutary effects of loyalty to one’s lover

244
will typically be greater than the salutary of effects of loyalty to one’s colleagues.
Consequently, when our respective ideas of those effects are enlivened into impressions
via the sympathy process, the resulting sympathetic and moral feelings will be more
intense in the first appraiser than in the second. Hence the MPV must ensure a uniform
level of generality/specificity in appraisers’ ideas of a trait’s typical effects in order to
ensure that those ideas are uniformly vivid (348). Only then can the MPV ensure
stability and uniformity of moral feeling and judgment.
Abramson has already argued that the MPV requires appraisers to “approach
equally near” the agent and idealize the agent’s trait such that it is no longer conceived as
the agent’s trait, but simply as a trait to which we are all subject in virtue of our shared
humanity. She suggests that the MPV also requires appraisers to once again idealize their
conception of a trait’s typical effects such that we conceive of those effects as “the usual
effects of a trait on any and all members of those to whom the agent typically expresses
that trait” (349, her emphasis). By preventing appraisers from conceiving of any
variability in the typical effects of a given trait upon any typically affected individual,
Abramson holds that Hume’s MPV thereby prevents any variability of sympathy that
“traces to variations in the specificity of our ideas” (ibid). In short, according to
Abramson, “to restrict ourselves to the general point of view is not only to limit our
purview to those to whom an agent typically expresses that trait, but also to form ideas of
the typical effects of traits that are equally applicable to all members of that group” (349,
her emphasis).

245
So, to sum up Abramson’s general approach, she conceives of Hume’s MPV as a
device for eliminating sympathy’s variability by abstracting away from various sources
of variability, namely: (a) certain peculiarities in the agent’s circumstances; (b)
peculiarities in those typically affected by the trait; (c) blood-relations between appraisers
and those typically affected; (d) the contiguity between appraisers and those typically
affected; (e) relations of resemblance between appraisers and those typically affected, as
well as relations of resemblance between appraisers and the trait under evaluation; and (f)
the relative specificity/generality of appraisers’ ideas. In order to secure the requisite
stability and uniformity in moral feeling and thereby judgment, Abramson claims that
Hume’s MPV requires appraisers to conceive of a given trait as one to which we are all
subject in virtue of our humanity and focus only on that trait’s usual effects on any and all
members of those to whom anyone within a specific social context would typically
expresses that trait. Appraisers achieve this focus by idealizing their ideas of: (1) the
trait; (2) the trait’s typical effects; (3) each individual typically affected by the trait; and
(4) the overall group typically affected.
§VII.3 The proffered interpretation
Let’s now consider the interpretation of Hume’s general rules regulating
sympathy that this project favors. In order to motivate this interpretation, we shall
critique Abramson’s proposal, which provides a remarkably different explanation of
Hume’s conception of, and focus on, the agent’s “narrow circle”. First, however, it is
noteworthy that Abramson and I do agree that the range of those with whom appraisers
are to sympathize from the MPV (i.e., the agent’s narrow circle) can vary based on two

246
factors, namely: (i) the particular trait under consideration and (ii) various features of the
social context in which the trait is expressed. We also agree that once a particular trait
and specific social circumstances are identified, the MPV sets “definite parameters for
our sympathetic survey”. The social circumstances also play a key role in determining
the usual effects of a trait. Hume’s view is that others’ emotional states that are elicited
by the causal tendency of an agent’s act or trait to fulfill or foil particular ends, can vary
depending on the social circumstances. Finally, we also agree that, as things stand, this
account does not resolve all of the problems due to sympathy’s variability.
Disagreement between Abramson and I arises when it comes to our interpretation
of how Hume’s MPV resolves the remaining problems. As discussed, Abramson claims
that Hume’s MPV requires appraisers to conceive of a given trait as one to which we are
all subject in virtue of our humanity and focus merely on that trait’s usual effects on any
and all members of those to whom anyone within a specific social context would
typically expresses that trait. In her view, appraisers achieve this by idealizing their ideas
of: (1) the overall group affected by the trait; (2) the particular individuals that compose
that group; (3) the trait’s typical effects; and (4) the trait itself.
In what follows, I argue that Hume’s MPV neither requires appraisers to conceive
of a given trait as one to which we are all subject in virtue of our humanity; nor does it
require appraisers to focus merely on the trait’s usual effects on any and all members of
those to whom anyone within a specific social context would typically expresses that
trait. Moreover, apart from the proposed idealization of the affected group, the MPV
does not necessitate any of the specific idealizations that Abramson claims, or so I argue.

247
Consider first the idealization of the individuals affected by the trait. According
to Abramson, we need to use our imagination to exclude various peculiarities of an
individual typically affected by an agent’s character trait.

Abramson’s example is an
agent who is witty but has dull friends. Consequently, the effects of the agent’s wit on
his friends are typically disagreeable. Nevertheless, we presumably do not disapprove of
the agent’s wit. Abramson claims that we do not disapprove of the agent’s trait in this
case because we recognize that the effects on his actual friends’ feelings are atypical.
And the effects are atypical because the agent’s actual friends are atypical; they are
psychologically abnormal.
According to Abramson, the MPV addresses this problem by requiring appraisers
to idealize the individuals typically affected by the agent’s character trait, and conceive of
those individuals as psychologically normal (i.e., typical) even if they in fact happen to
be psychologically abnormal (i.e., atypical). This idealization thereby excludes various
psychological peculiarities of any individuals typically affected by an agent’s character
trait. Although I see the point of Abramson’s idealization, caution is needed here.
Sometimes the affected individuals’ psychological peculiarities are morally relevant.
Consider, for instance, the following example.
Suppose that I want to evaluate whether a particular trait of mine is an
occupational virtue, that is, whether the trait is useful or immediately agreeable among
those to whom my trait is typically expressed at work. Presumably my particular
occupation is morally relevant to this evaluation because it helps decipher who it is at my
work that are typically effected by my trait. Suppose, for instance, that my occupation

248
primarily involves working with individuals who have certain psychological peculiarities
such as autism. In this case, a trait of which we might typically approve or (as the case
may be) disapprove may be disapproved or (as the case may be) approved of when
expressed in this particular set of social circumstances C, namely, the circumstances that
include the peculiarities of those individuals with whom the agent typically has contact
given the agent’s particular occupation.
How would Abramson’s interpretation handle this sort of case? As a start, it is
noteworthy that her interpretation does allow for role-specific moral evaluations. For
instance, Abramson writes: “we must conceive of individuals [typically affected by an
agent’s trait] as standing in specified relationships to the agent…because doing so will
always be necessary when our evaluation is of a role-specific virtue or vice (e.g., if we
are asking whether a trait is a virtue among friends)” (344).
210
Given her allowance of
such role-specific virtues and vices, perhaps the peculiarities her idealization excludes are
only those features that are atypical within the particular group of individuals typically
affected by that particular agent’s trait. In other words, in the case above, those typically
affected by the agent’s occupational trait are predominantly individuals with autism.
Hence their autism doesn’t qualify as the sort of peculiarity that the idealization excludes.
However, it isn’t entirely clear whether this sort of move is available to
Abramson. The problem is that it is unclear just how specific the roles between the agent
and those typically affected can be on her view. Indeed, when Abramson speaks of “role-
specific” virtues and vices, she only appears to be referring to such roles as parent-and-

210
She offers a second reason why she makes this claim, namely: “because doing so will sometimes be part
and parcel of determining just who is directly affected by that agent’s trait” (ibid).

249
child, doctor-and-patient, and friendship roles. But can these roles be even more specific
so as to include, for instance, the parent-and-child-with-cerebral-palsy role, the doctor-
and-patients-with-autism role, the friend-and-dull-friends role, etc?
Abramson doesn’t appear to allow these more specific roles to be included in our
moral evaluations. For instance, she seems to suggest that, on her view, we can evaluate
a particular virtue among friends only if the virtue is specific to that role, that is, only if
the virtue only typically affects that group (e.g., the agent’s friends). But it’s hard to
think of many virtues that play such a specific role. Loyalty, for example, isn’t specific
to friends. We are typically loyal to our spouse, our friends, our pets, our colleagues, our
country, etc. So the worry is that these more specific roles are incompatible with the
proposed idealizations. In other words, the worry is that the restrictions she proffers
preclude various features of those typically affected by the agent’s trait or act (e.g., the
autism of the agent’s patients or the dull disposition of his friends) in our moral
evaluations. Yet these features appear to be morally relevant in some cases.
Abramson’s exclusion of these features appears to be a symptom of a larger
problem: Abramson’s proposed idealizations do not appear to take into account the fact
that, or at least the extent to which, the range of the individuals to be included in our
moral evaluations of the typical effects of an act or trait can vary. For instance, the range
may include all individuals typically affected by a particular trait; e.g., we may want to
evaluate whether the trait is, generally speaking, a virtue. But we sometimes we may
want to only evaluate whether that trait is a virtue among a narrower range of individuals,
for example, whether it is a virtue among friends. Abramson does not appear to allow for

250
these range changes. And she certainly does not appear to permit even narrower ranges
of moral evaluation. For instance, she does not appear to recognize that appraisers can
further narrow the range of their moral judgments and evaluate the typical effects of
loyalty solely among the agent’s actual friends that are typically affected.
In such narrow-range moral evaluations, various peculiarities of one’s actual
friends (e.g., their dullness) may indeed be morally relevant depending on the trait under
evaluation. Yet Abramson’s interpretation of Hume’s MPV excludes these peculiarities.
She claims that inclusion of such peculiarities within our moral evaluations produces the
wrong results. While this is true of wide-ranged moral evaluations (e.g., the evaluation of
loyalty per se) as well as narrower-ranged evaluations (e.g., evaluation of loyalty among
friends), such peculiarities sometimes are morally relevant in those evaluations that take
an even narrower range (e.g., evaluation of one’s loyalty to disloyal friends).
In short, depending on the range of the individuals under consideration, the
morally relevant features to be taken into account by our moral evaluations will
sometimes include various peculiarities of the individuals typically affected by the trait
(or act). Consequently, idealizations of the individuals typically affected by the agent’s
trait that preclude various peculiarities of those individuals would sometimes produce the
wrong results. Hence there is a problem with Abramson’s approach.
Moreover, because Abramson’s proposed idealization excludes moral evaluations
that have a rather narrow-range, it is unclear how appraisers can morally evaluate a
particular act or trait in a particular situation. For instance, suppose I want to inquire into
what particular traits I morally ought to express given my particular occupation and the

251
peculiarities of those with whom I will typically come into contact. Or suppose I want to
know what I morally ought to do in a particular situation. How would one arrive at an
answer to such inquiries on Abramson’s view given the idealizations that she proffers?
Abramson’s account tells us what traits are virtuous and vicious generally
speaking, but that is all it tells us. It doesn’t tell us what traits we ought to express in a
given situation; nor does it tell us what to do in particular instances. While it’s true that
Hume tends to focus on traits rather than acts, it would be a serious shortcoming of
Hume’s MPV if it could not tell us what acts are morally right or wrong, good or bad, in
a given scenario. Surely any plausible moral theory must be able to tell us, at least in
principle, what we morally ought or ought not to do in a given situation. As we shall
discuss in the following section, according to the interpretation of Hume’s MPV offered
by this project, it does allow us to morally evaluate particular acts.
It is noteworthy that Abramson could adjust her approach and take into account
the fact that the range of those typically affected by an act or trait can and does vary in
our moral evaluations. As far as I can tell, she can do so in a way that is generally
compatible with her proposed idealization of the individual, while also allowing for
individuals’ various peculiarities to be included in appraisers’ moral evaluations when
they are morally relevant. To explain how this can be done, we need only look to
Abramson’s approach to addressing other causes of variability in sympathy. Recall that
Abramson already requires appraisers to approach as “equally near” the agent as the
agent’s narrow circle so that appraisers can accurately determine the typical effects of the
agent’s trait (or act) on that group. Appraisers are then to abstract away from the agent’s

252
narrow circle, and simply conceive of the typical effects of the trait (or act) upon anyone,
i.e., the typical effects that the trait would have on anyone were they to approach equally
near anyone who possesses that trait.
My suggestion is that Abramson could make a similar move so that appraisers can
determine whether various peculiarities of those typically affected are morally relevant or
irrelevant. Given the range, in a particular evaluation, of the individuals typically
affected, appraisers could focus in on that group (e.g., the agent’s actual friends) and
determine whether the group possesses any peculiar features that are morally relevant;
e.g., perhaps we are evaluating the typical effects of the agent’s wit on his dull friends.
Appraisers would determine the moral relevance of any such peculiarities by determining
whether those peculiarities impact the act or trait’s typical effects on the group. If those
features do have an impact, then they are morally relevant; if they do not, then they are
irrelevant. Appraisers would then abstract away from the actual individuals who
comprise the group and conceive of them as just anyone who possessed those features
and belonged to the group that comprises the range of individuals under evaluation.
To be sure, Abramson’s requirement that appraisers abstract away from the actual
individuals typically affected would still prevent moral evaluation of an agent’s actual
friends. But the whole point of the appeal to the agent’s actual friends in the first place is
to pick out any morally relevant peculiarities they might possess (e.g., their dullness).
Significantly, the proposed alteration to Abramson’s approach takes those peculiarities
into account.

253
In short, the alteration limits Abramson’s idealization of the individuals who fall
within a particular range of those typically affected by the agent’s act or trait. The
idealization would still serve to preclude the various peculiarities of those individuals
typically affected in all cases in which those peculiarities are morally irrelevant, but it
would not preclude them when they are morally relevant. Thus, Abramson can, in
principle, preserve her idealization while avoiding the problem I have raised against it.
However, there are other problems with Abramson’s idealizations. Consider the
second way in which Abramson claims moral appraisers are to idealize the individuals
typically affected by a given trait. Abramson holds that Hume’s MPV requires appraisers
to idealize those individuals such that appraisers exclude those individuals’ particular
location within their social world. Recall that this idealization is meant to correct for any
variation in appraisers’ sympathy due to their awareness of any blood relations between
the individuals typically affected by the agent’s trait and the appraisers. While I agree
with Abramson that Hume’s MPV does exclude the particular relationship that affected
individuals have to us as appraisers, I shall argue that this does not entail excluding
altogether the affected individuals’ particular location within their social world. In many
cases, those individuals’ particular location with respect to the agent often is morally
relevant. For instance, it is often morally relevant whether an affected individual is either
the agent’s child, or lover, or friend, or a total stranger.
Abramson disagrees. She argues that if appraisers were to include within their
moral evaluations the individuals’ particular location in their social world, then this
would “not only threaten the possibility of intersubjective agreement, it would also make

254
Hume’s moral perspective personal… It would do so because my [i.e., the appraiser’s]
moral evaluations would then depend on my knowledge of the relationship between the
object of evaluation and myself ” (357, n 45). Consider Abramson’s own example.
Suppose that I want to assess the moral value of a particular agent’s parental affection
(e.g., my own). Abramson claims that the knowledge that I am morally evaluating my
children would make my moral judgment personal. And such personal judgments are
negatively influenced by the natural variability of sympathy.
Hume is well aware of this fact, and according to Abramson this is why Hume
makes the MPV impersonal—i.e., “[Hume’s MPV] is not essentially informed by a
person’s awareness of the relationship between the object of evaluation and herself as
such” (334). This is why Abramson concludes that, when I am evaluating my own
parental affection, I cannot conceive of the affected individuals as my children. In order
to be impersonal, she thinks that Hume’s MPV must require me to conceive of my
children just as “someone’s” children (in a specific social context). In other words, we
are to conceive of them as just any individuals who could be “any agent’s” children
within a specific social context (345).
Abramson’s conclusion is much too hasty however. I agree with Abramson that
we should not “conceive of those who are directly affected by any agent’s trait as
standing in any particular relationship to us as appraisers” (345, my emphasis).
Nevertheless, contrary to Abramson’s conclusion, when an appraiser morally evaluates
her own traits—that is, when the appraiser is also the agent being evaluated—I suggest
that she should conceive of those who are directly affected by the agent’s trait as standing

255
in a particular relationship to herself, where this reference to herself is a reference to
herself qua agent, not qua appraiser. So, to return to Abramson’s example, it is true that
I should not conceive of the individuals typically affected by my parental affection as my-
qua-appraiser’s children, for that relation is morally irrelevant. But I should conceive of
them as my-qua-agent’s children, for that relation often is morally relevant. It is often
morally relevant because different affected individuals can have different relations with
agent. As a result, different individuals are often affected differently by the agent’s trait
(or act). And this fact is relevant to appraisers’ moral evaluation of the trait (or act).
I shall say more about this in the following section, but at this point I want to
dispute Abramson’s contention that this proposal—namely, that we should include the
relation between myself qua agent and each typically affected individual—would make
Hume’s MPV personal. I have claimed that moral evaluation of my own parental
affections must include knowledge of how my own trait affects my children because that
knowledge is morally relevant. Hence it is true that, on my proposal, some of my moral
evaluations “depend on my knowledge of the relationship between the object of
evaluation and myself”. This, however, does not make Hume’s MPV personal.
The MPV is not personal in such instances because the relationship is only
relevant when I am the agent under evaluation. The reference to myself in this case is
myself qua agent, not qua appraiser. Hence, when I morally evaluate any other agent’s
trait, my knowledge of the relationship between the narrow circle and myself is

256
excluded.
211
Since none of my moral evaluations depend upon my knowledge of the
relationship between the narrow circle and myself qua appraiser, I contend that Hume’s
MPV remains impersonal; for it remains the case that the MPV is not “essentially
informed” by one’s awareness of the relationship between those typically affected and
oneself. The MPV is never informed by one’s awareness of the relationship between the
agent’s narrow circle and oneself qua appraiser, although it is sometimes informed by
one’s awareness of the relationship between the narrow circle and oneself qua agent,
namely, in those cases in which the appraiser is also the agent under evaluation.
So, in opposition to Abramson’s interpretation, I contend that Hume’s MPV does
not altogether exclude the particular location of those individuals affected by the agent’s
act or trait. Hume’s MPV does not exclude those individuals’ particular location with
respect to the agent precisely because Hume recognizes that this location is morally
relevant, or so I shall argue. Nevertheless, Abramson is right that the MPV cannot allow
us to “conceive of those who are directly affected by an agent’s trait as standing in any
particular relationship to us as appraisers.” It cannot both because this relationship is
morally irrelevant and because it threatens the possibility of interpersonal agreement, just
as Abramson claims. As I read Hume, then, the MPV does not include the affected
individuals’ particular location with respect to the appraiser, but it does include their

211
It is morally irrelevant unless, of course, I am the agent’s child. Regardless, when conceiving of the
relationship between the object of evaluation and myself, I conceive of myself qua an individual who is
affected by the agent’s act or trait, and not myself qua appraiser.

257
particular location with respect to the agent.
212
Consequently, we must be careful not to
confuse or conflate the narrow circle’s relation to the appraiser with their relation to the
agent, especially when the appraiser is also the agent.
It is not difficult to see how this same basic solution—viz., excluding the
appraiser’s relation to both the agent and the individuals typically affected by the agent’s
trait—also resolves any variability in sympathy due to the relations of both contiguity and
resemblance between those typically affected by an agent’s trait and appraisers, as well
as any variability in sympathy due to relations of contiguity and resemblance between the
agent and appraisers. Consideration of contiguity and resemblance between the
appraiser and those typically affected by the trait in question, as well as the agent herself,
amounts to a failure to meet the conditions of the MPV. By excluding these
considerations, the MPV regulates key aspects of sympathy’s variability that undermine
interpersonal agreement.
As discussed, Abramson objects to this approach on the following grounds. In
her view, appraisers cannot conceive of the trait’s usual effects as the misery or, as the
case may be, pleasure of those who comprise the agent’s narrow circle (in a specific
social context). Abramson claims that appraisers cannot conceive of the trait’s usual
effects in this way because the resulting feelings would not count as moral sentiments.

212
This, however, is certainly not to say that it is easy to use our imagination to exclude or suspend our
(qua appraiser) relation to those affected by the agent’s act or trait. Hume recognizes that it can be quite
difficult in many cases to satisfy the conditions of the MPV. This is why he repeatedly talks about using
the MPV to correct our judgments even when our sentiments are more stubborn. In other words, we can
know what we would feel were we to successfully exclude both the agent’s and the affected individuals’
relation to us as appraisers. Consequently, we can issue a true moral judgment without experiencing the
moral sentiment that confirms the judgment.

258
The problem is that these resulting feelings are inevitably directed toward the agent’s
narrow circle rather than the agent himself or his trait.
To recapitulate, her reasoning is as follows: “if the idea one has is ‘the misery of
those in the agent’s narrow circle,’ then the object of that idea is those in the narrow
circle, the subject is the circumstances of those in the narrow circle that are miserable,
and the quality is that misery. The sentiment that results when that idea is enlivened into
an impression will take the narrow circle as its object as well” (358, n 46). According to
Abramson, this is problematic because the relevant moral sentiments clearly take the
agent, or at least the agent’s trait, as its object rather than the agent’s narrow circle.
Abramson would therefore conclude that the proposed account is untenable. In her view,
the MPV necessitates that appraisers idealize their conception of a given trait’s usual
effects and conceive of its usual effects merely as the misery or, as the case may be,
pleasure usually produced by the trait (in a specific social context).
I disagree. Abramson’s conclusion—viz., that a sentiment that results from those
sympathetic feelings for the agent’s narrow circle cannot be a moral sentiment—does not
follow. She is correct that the sympathetic feelings that take those in the agent’s narrow
circle as its object cannot be the relevant moral feelings for the reason she mentions.
However, recall the argument in §V.5, which has argued that the appraiser is presumably
aware that the agent (or at least the agent’s trait) is the cause of the narrow circle’s misery
and thus ultimately the cause of the appraiser’s sympathetic pain as well. Due to the
principles of mental association among ideas and impressions, this knowledge, in
conjunction with the appraiser’s sympathetic pain, will result in the appraiser having

259
feelings of disapproval toward the agent and/or his trait. Hence, in opposition to
Abramson’s claim, one’s feeling of disapproval that results from one’s extensive
sympathy with the agent’s narrow circle can and does qualify as a moral sentiment.
Significantly, however, it does follow from this account that our sympathetic feelings and
our moral feelings of approval and disapproval are distinct since our sympathetic feelings
engender these moral feelings.
213

Interestingly, in a footnote, Abramson considers the proffered suggestion as a
third option, namely, that we should conceive of the misery in question as the misery that
the trait of the agent typically causes in the agent’s narrow circle (358, n 46). But she
dismisses this option on the grounds that “conceiving of the misery in this way will not
control for variability due to contiguity or resemblance” (ibid). In Abramson’s view, it
will not correct for such variability because, by including both the agent and the agent’s
narrow circle within the moral equation, appraisers’ relations to the agent and the agent’s
narrow circle will also consequently be included. And this makes the MPV personal.
Because “appraisers will lie at varying distances from the agent [and the agent’s narrow
circle]” (ibid), there will be variability in appraisers’ sympathy. Abramson concludes
that we must therefore reject this third option. Hume’s MPV must exclude these relations
in order to ensure uniformity of moral feeling and judgment.
Contrary to Abramson’s contention, however, I have suggested that Hume’s
MPV need only exclude from moral evaluations appraisers’ relations of contiguity and
resemblance to the agent as well as those affected by the agent’s trait. By disallowing

213
See the argument in §V.5 for further discussion.

260
appraisers from considering these relations of contiguity and resemblance, the MPV
regulates the unwanted variability in appraisers’ sympathetic feelings due to those
relations. Thus, Abramson’s reasons for dismissing the proffered account as a viable
option are defeated.
I have argued, and shall continue to argue, that the relation between the agent and
the narrow circle is often a morally relevant feature. To get results that tend to agree
with our moral intuitions, the MPV often must include both the relation between the
agent and the agent’s narrow circle. However, rather than taking this third option and
conceiving of a trait’s usual effects such as misery as the misery of the agent’s narrow
circle typically caused by the agent’s trait (in a particular social context C), Abramson
chooses to conceive of it simply as the “misery that can be anyone’s, including mine,
were I in her [viz., the agent’s] social circumstances and directly affected by her trait”.
According to Abramson, this is the point of Hume’s claim that the MPV requires
that we “approach equally near” the agent (T 582). By approaching equally near the
agent, appraisers can determine the typical effects of the agent’s trait on those typically
affected. We then abstract away from the agent’s narrow circle, and simply conceive of
the typical effects of the trait upon anyone, i.e., the typical effects that the trait would
have on anyone were they to approach equally near anyone who possesses that trait. In
short, on Abramson’s account, to restrict oneself to Hume’s MPV is to restrict one’s
purview merely to the typical effects of a particular trait (in a given social context).
As discussed, Abramson also uses this approach to help eliminate resemblance as
a source of sympathy’s variability. Since the identities of both the agent and those

261
typically affected by the agent’s trait have already been excluded from the MPV, these
resemblance relations should have no impact on appraisers’ moral evaluations.
Abramson claims that the only sort of problematic resemblance relation remaining is the
appraiser’s resemblance to the trait itself (347). In Abramson’s view, this resemblance is
problematic because if one appraiser more closely identifies with a particular trait than
other appraisers, then that appraiser will supposedly be more likely to experience more
intense sympathetic and moral feelings than will other appraisers.
Unfortunately, Abramson doesn’t explain why she thinks this is the case. Given
the account of the sympathy process discussed in the previous chapter, it’s not entirely
clear why an appraisers’ familiarity with a particular trait will affect the vivacity of her
sympathetic feelings for the agent’s narrow circle. By being more familiar with a trait the
appraiser might have a better grasp of the usual effects of that trait, but it might just as
well give rise to various false prejudicial judgments about its usual effects if the trait is
expressed under a substantially different set of social circumstances than those familiar to
the agent. In either case, this problem does not appear to reside directly with the
appraiser’s resemblance to the trait. It arguably resides with the appraiser’s evaluation of
the usual effects within the specific social context. And this is a problem that the
regulative general rules governing causal reasoning/judgments must address.
Regardless, let’s assume for the sake of argument that appraisers’ resemblance to
a trait would give rise to variability in appraisers’ sympathetic and moral feelings that
undermines the requisite intersubjective agreement among appraisers. Rather than
idealize the trait itself, as Abramson suggests, I argue that Hume’s MPV can resolve this

262
problem by simply restricting appraisers from comparing their own traits with the agent’s
trait. So, to sum up the proposal thus far, according to my interpretation the MPV need
only restrict appraisers from considering: (i) their causal (blood) relations with the agent
and the agent’s narrow circle; (ii) their contiguity relations with the agent and the agent’s
narrow circle; and (iii) their resemblance relations with the agent, the agent’s narrow
circle, and the agent’s trait under evaluation.
The last remaining source of sympathy’s variability is variability due to the
relative variability in the specificity/generality of the appraisers’ ideas of a trait’s usual
effects. If different appraisers have a more or less specific/general idea of a particular
trait’s typical effects, then the resulting variation in the vivacity of appraisers’ ideas will
also lead to variation in their sympathetic and moral feelings. As mentioned, Abramson’s
example involves two appraisers who are evaluating the typical effects of loyalty. She
observes that “those to whom loyalty is expressed will typically be affected differently, in
part because of the different ways in which they have commerce with the loyal person”
(348). Consequently, when one appraiser conceives of the usual effects of loyalty to the
agent’s lover, while another appraiser conceives of the usual effects of loyalty to the
agent’s colleagues, their sympathetic and moral feelings will vary.
Abramson proposes to eliminate this variability by having appraisers idealize their
conception of a trait’s typical effects such that we conceive of those effects as “the usual
effects of a trait on any and all members of those to whom the agent typically expresses
that trait” (349, her emphasis). That is, appraisers are to exclude the different ways in
which individuals who have commerce with the agent are affected by the agent’s trait.

263
Once again, I take issue with Abramson’s solution. First, in Hume’s view, the
different ways in which typically affected individuals “have commerce” with the agent,
that is, the different spheres of connection with the agent, are morally relevant.
Consequently, the MPV takes into account these differences. In the following section, I
argue that this is the main point of Hume’s claim that appraisers are to “approach equally
near the agent”, as well as Hume’s focus on the agent’s narrow circle.
Second, Abramson’s solution to the above variability problem is superfluous
because she already appears to have a constraint in place to address it, namely: the fact
that appraisers are to consider whether a given trait is role-specific or not. Consider once
again Abramson’s example of two appraisers, one who conceives of loyalty’s usual
effects on the agent’s lover, and the other who conceives of loyalty’s usual effects on the
agent’s colleagues. The variability in feeling between the appraisers occurs because they
are not evaluating the same thing. Each appraiser is evaluating the trait’s usual effects
with respect to different specific roles, that is, different groups. But if a trait such as
loyalty is not role-specific—i.e., if the trait does not typically affect only a particular
group—then why would appraisers be evaluating loyalty with respect to a specific
role/group? It appears that this would be a mistake given the constraints that Abramson
has already posited. Hence the solution she offers above appears to be superfluous.
Given that the range of the narrow circle can vary, the MPV requires appraisers to
take that range into account. If the proper range of the agent’s narrow circle is broad
(e.g., if it includes all those typically affected by the agent’s loyalty), then appraisers are
to consider the usual effects of loyalty on the agent’s entire circle of connections. It

264
would therefore be a mistake on the appraisers’ part to focus merely on the usual effects
of the agent’s trait on a narrower group (e.g., loyalty to the agent’s lover or the agent’s
colleagues). Similarly, if the object of evaluation has a narrower range (e.g., evaluation
of the usual effects of an agent’s trait on his actual friends), then it would be a mistake for
appraisers to focus on either the usual effects of the trait on all those who typically have
commerce with the agent, or the usual effects of the trait on the agent’s colleagues. In
this way, Hume’s MPV ensures that appraisers consider the same object of evaluation.
214

Finally, there is a third reason why I take issue with Abramson’s solution, that is,
her proposal that appraisers idealize their conception of a trait’s typical effects such that
they conceive of those effects as the usual effects that are equally applicable to any and
all of those typically affected by that trait (in C). Abramson does not explain how exactly
appraisers are supposed to determine the trait’s usual effects that are equally applicable
to all of those typically affected, and it is highly likely that, in at least some cases,
appraisers will not agree in their assessments.
It is noteworthy that, in her discussion, Abramson focuses on a trait (viz., loyalty)
that has beneficial effects for all of those typically affected by that trait. But presumably
there are traits that do not exhibit this feature. For example, being a workaholic seems to
be a character trait that typically has salutary effects on the agent’s colleagues and others
but also typically has deleterious effects on the agent’s friends and family (e.g., the

214
Those cases in which the range of the evaluation is not made entirely explicit I suggest that the default
range is the more general one, unless the conversational context indicates otherwise. In any case, each
appraiser must be as clear as possible about the range he or she has fixed upon in arriving at an evaluative
conclusion; for there can only be interpersonal agreement in both feeling and judgment when appraisers
agree about (inter alia) the specific range of the evaluation.

265
agent’s spouse and/or children). Significantly, this trait and others like it do not appear to
have usual effects that are equally applicable to any and all of those typically affected (in
C). Consequently, it is highly unlikely that appraisers would agree about such a trait’s
usual effects that are equally applicable to all those typically affected (in C).
The usual effects of some traits or acts are not equally applicable to any and all of
those typically affected (in C) precisely because the usual effects of those traits or acts
vary depending on the affected individuals’ particular social location with respect to the
agent. Indeed, there would presumably be no reason to narrow the range of our moral
evaluations if an agent’s act or trait had roughly the same effect on all typically affected
individuals irrespective of their particular location with respect to the agent. I suggest
that any plausible interpretation of Hume’s MPV needs to take these facts into account.
Unfortunately Abramson’s proposed idealizations, which require appraisers to exclude
various features of the relationship between the agent and the agent’s narrow circle,
prevent appraisers from taking into account these and other morally relevant facts.
This project has offered an alternative to Abramson’s approach. In sum, it argues
that Hume’s MPV requires appraisers to exclude as morally irrelevant (i) their causal
(blood) relations with the agent and the agent’s narrow circle; (ii) their contiguity
relations with the agent and the agent’s narrow circle; and (iii) their resemblance relations
with the agent, the agent’s narrow circle, and the agent’s trait under evaluation. This
corrects for unwanted variability in sympathy. This project also argues that the MPV
requires appraisers to include as morally relevant: (a) the particular act or trait under
evaluation; (b) the specific social context C in which the act or trait is expressed; (c) the

266
affected individuals’ particular social location, or “sphere,” with respect to agent; (d) the
range of the moral evaluation; and finally, depending on the range of the evaluation and
the particular act or trait under evaluation, (e) various peculiarities of those who fall
within the given range.
§VII.3.1 Disinterestedly sympathizing from the agent’s position
The proffered interpretation of Hume’s MPV does agree with at least two of
Abramson’s observations. First, it agrees that Hume’s MPV can be more or less
inclusive; depending on the agent’s social context, the trait in question, and the range of
the moral evaluation, the typically affected group may be conceived more narrowly up to
the point of including only a member of the agent’s immediate family, or it may be
conceived more widely up to the point of including all of humanity and all other sentient
beings. Second, it agrees that Hume’s MPV must require uniformity in our ideas about
the usual effects of the trait in question; otherwise the MPV could not ensure uniformity
and stability of feeling and judgment.
However, contrary to Abramson’s interpretation, I argue that Hume’s MPV
requires appraisers to consider and weigh the various spheres of connection that the
affected individuals have with the agent. As I read Hume, appraisers are required to
conceive of a trait’s typical effects as the typical effects that are equally applicable to any
and all individuals typically affected relative to those individuals’ particular sphere of
connection with the agent (in C). In this way, Hume in effect builds partiality into the
MPV. This allows variability in appraisers’ sympathy with respect to different spheres,
but it does so in a way that also ensures stability and uniformity in appraisers’ sympathy.

267
This, I suggest, is the real import of Hume’s claim that appraisers are to
“approach equally near” the agent when sympathizing with the agent’s “narrow circle”.
First, the focus on the agent’s narrow circle helps ensure interpersonal agreement by
indicating that we are to “over-look our own interest in those general judgments” (T 582)
and be disinterested appraisers. As we’ve already discussed, the MPV forbids moral
appraisers from referencing themselves, qua appraisers, when issuing moral judgments.
Hence we are neither to consider our own interests/ends nor “whether the persons
affected…be our [qua appraisers] acquaintance or strangers, countrymen or foreigners”
(T 582). The only morally relevant interests/ends and relations are those pertaining to the
agent and the group affected by the agent’s act or trait. Thus, the only instances in which
we are allowed to reference ourselves are those instances in which we are either the agent
being evaluated—i.e., we are evaluating ourselves—or a member of the group with
which we are sympathizing—i.e., we are among those affected by the agent’s act or trait.
Second, Hume’s stipulation that we are to “approach equally near” the agent
helps preserve variability in our sympathetic feelings by indicating that we are to
consider whether those affected by the agent’s act or trait are the agent’s acquaintance or
strangers, countrymen or foreigners. As I read Hume, we are to use our imagination to
mirror the “particular connection”—i.e., the contiguity, resemblance, and causal (blood)
relations—between the agent and those who have commerce with him. In other words,
we approach equally near the agent by imagining and reflecting upon the various
aforementioned relations that the agent stands in with respect to those typically affected
by his act or trait. We then sympathize with those individuals in a disinterested manner

268
that reflects their particular connection to the agent. This will result in the intensity of
our sympathetic feelings for the individuals typically affected by the agent’s act or trait
not only being proportional to the usual intensity of their pain or pleasure (in C); those
feelings will also be proportional to, and vary with, their propinquity to the agent (in C).
So, as I read Hume, appraisers are not simply to consider the interests/ends of
those typically affected rather than their own. Appraisers are also to divide those with
whom the agent typically has commerce into various concentric spheres of connection
with the agent such that those who are closer in psychological proximity to the agent fall
within a narrower circle. Appraisers are then to approach equally near the agent as each
concentric circle of connection the pleasure/pain of those individuals typically affected
by the agent’s act or trait according to the distance that they lie from the agent. The
pleasure/pain of those who are closer in psychological proximity to the agent will thereby
be given more weight than the pleasure/pain of those who are farther away, and the
weight decreases proportionally as the distance from the agent increases.
215

Textual evidence in support of the idea that Hume builds such variability and
hence partiality into the MPV can be found in the following passages:
…[W]e always consider the natural and usual force of the passions, when
we determine concerning vice and virtue; and if the passions depart very

215
Louis Loeb suggests a similar approach: “Suppose Hume’s thought is that the judicious spectator adopts
a point of view in which the pleasures and pains that register on him most strongly are those of the agent’s
narrow[est] circle…In order for the pleasures and pains that register most strongly to be those of the agent’s
narrow[est] circle, a moral judge must sympathize with its members from close-up, either from nearby or
from within the narrow[est] circle itself.” (2004a: 344). Loeb suggests the latter option and goes on to state
that: “Virtue and vice is a matter of the sentiment that would be produced from the point of view of a
member of that circle” (345, my emphasis). I disagree. Hume clearly states that the relevant point of view
is one in which the appraiser approaches equally near the agent rather than a member of the agent’s narrow
circle. Interestingly, in a separate article, Loeb (2004b) suggests that the MPV should be identified with
that of the agent. This later approach is therefore, in this respect, more similar to the view proposed here.

269
much from the common measures on either side, they are always
disapprov’d as vicious. A man naturally loves his children better than his
nephews, his nephews better than his cousins, his cousins better than
strangers, where every thing else is equal. Hence arise our common
measures of duty, in preferring one to the other. Our sense of duty always
follows the common and natural course of our passions. (T 483-4)
216

And as every immorality is deriv’d from some defect or unsoundness of
the passions, and as this defect must be judg’d of, in a great measure, from
the ordinary course of nature in the constitution of the mind; ‘twill be easy
to know, whether we be guilty of any immorality, with regard to others, by
considering the natural, and usual force of those several affections, which
are directed towards them. Now it appears, that in the original frame of
our mind, our strongest attention is confin’d to ourselves; our next is
extended to our relations and acquaintance; and ‘tis only the weakest
which reaches to strangers and indifferent persons. This partiality, then,
and unequal affection, must not only have an influence on our behaviour
and conduct in society, but even on our ideas of vice and virtue; so as to
make us regard any remarkable transgression of such a degree of
partiality, either by too great an enlargement, or contraction of the
affections, as vicious and immoral. This we may observe in our common
judgments concerning actions, where we blame a person, who either
centers all his affections in his family, or is so regardless of them, as, in
any opposition of interest, to give the preference to a stranger, or mere
chance acquaintance. (T 488-9)

So, on the proffered interpretation of Hume’s MPV, we are to sympathize most
intensely with the pleasures and pains of those who are closest to the agent, and our
sympathy with those typically affected becomes proportionately less intense as their
propinquity to the agent decreases. However, the priority given to those closest to the
agent does not negate the fact that our sympathetic feelings will also be proportional to

216
C.D. Broad’s (1971) “Self-referential altruism” has much affinity with this aspect of Hume’s account,
though Hume’s account is perhaps best construed as an agent-referential altruism. According to Broad,
self-referential altruism “holds that each of us has specially urgent obligations to benefit certain individuals
and groups which stand in certain special relations to himself, e.g., his parents, his children, his follow-
countrymen, etc. And it holds that these special relationships are the ultimate and sufficient ground for
these specially urgent claims on one’s beneficence” (1971: 280, Broad’s emphasis). Of course, in Hume’s
view, the ultimate ground is sympathy; for it is the natural variability of sympathy that explains why those
special relationships are especially urgent, namely, because they register more strongly on our sympathy.

270
the intensity of the pain or pleasure felt by any and all of those typically affected. The
priority merely suggests that pains or pleasures of equal intensity should register more on
an appraiser if it is had by the agent’s those who have a strong connection with the agent
(e.g., the agent’s family and friends) than if experienced by someone who has a weak
connection (e.g., a mere acquaintance of the agent).
Consider, for instance, Abramson’s example about the effects of loyalty on an
agent’s lover and colleagues (348). Rather than just considering the overall usual effects
of loyalty that apply equally to anyone typically affected, as Abramson proposes, I
suggest that the usual effects of an agent’s loyalty to his lover should ceteris paribus
register more strongly in appraisers than the usual effects of his loyalty to his colleagues.
We are still to generalize the usual effects, but we are to do so within each “sphere” or
“particular connection” to the agent (T 602). For example, the usual effects of the
agent’s loyalty are weighted the same for any and all of the agent’s lovers, just as they are
weighted the same for any and all of his colleagues.
217

Hume’s system of ethics can therefore be understood as a kind of weighted
universalism.
218
It is universal because the system takes into account the tendency of an
act or trait to fulfill or (as the case may be) foil ends that subsequently produce
pleasurable or (as the case may be) painful affects on any and all individuals within every
sphere of connection typically affected by an agent’s act or trait. But Hume’s moral

217
As we shall discuss in the following section, appraisers’ evaluation of so-called “artificial” virtues and
vices will not be weighted in this way. Regarding artificial virtues and vices, appraisers are to be impartial.
Consequently, appraisers will need to be careful to distinguish between natural and artificial virtues and
vices before evaluating an agent’s trait.

218
I borrow this terminology from Loeb (2004a: 346).

271
theory is also weighted, or partial, because the system does not count equally each
individual’s pains and/or pleasures. The system is weighted towards those who fall
within narrower and narrower circles of connection to agent.
This is the real import of Hume’s MPV having appraisers “approach equally near”
the agent and focus on the agent’s “narrow circle”, namely, to signal the importance of
having a universal but also a weighted system. By requiring appraisers to approach
equally near the agent and sympathize with those typically affected from that position—
that is, by requiring appraisers to sympathize from the agent’s position or point of view—
Hume’s MPV standardizes sympathy while at the same time preserving some variability
in sympathetic feeling. All appraisers are to give priority to those closest to the agent
(e.g., the agent’s spouse and children). This priority pays respect to the natural variability
of sympathy, and it reflects the sort of partiality that we often expect the agent’s conduct
to exhibit.
It is worth noting that in sympathizing from the agent’s position or point of view,
our feelings are not to simply mirror the actual feelings of the agent. Not just any
variability is permissible. The agent might, for instance, lack sympathetic feelings for
particular individuals affected by his act or trait, perhaps due to various comparisons
between himself and those individuals or to a psychology abnormality. In those instances
in which the agent does experience sympathetic feelings, they might be overly partial to
his family and/or friends, or they might not be partial to them at all. Thus, approaching
equally near the agent and sympathizing from the agent’s position does not simply
amount to sharing the agent’s actual sympathetic feelings, or lack thereof.

272
Although we are to sympathize from the agent’s position, we are also to
sympathize in a disinterested manner. The disinterestedness required by Hume’s MPV
effectively mitigates any tendency of sympathy to be overly partial or indifferent to those
individuals affected by the agent’s act or trait, just as it eliminates any contravening
feelings due to various comparisons. As we’ve discussed, the MPV’s disinterestedness is
achieved by requiring appraisers to exclude various features, and also to include others.
This general approach is promising because, on the one hand, it eliminates the sort of
comparisons between oneself and others that tend to undermine sympathy. On the other
hand, it does not exclude various forms of comparison that are morally relevant. It also
thereby helps ensure that appraisers consider various relevant “matter of fact” (i.e.,
causal) relations that might otherwise be overlooked or ignored.
In sum, by restricting both how and with whom we are to sympathize, the MPV
ensures stability and uniformity of moral feeling and thereby judgment. Interpersonal
agreement is thus secured. But unlike Abramson’s interpretation, the proffered account
of Hume’s MPV does not eliminate all sources of sympathy’s variability. Abramson
holds that appraisers are to use their imagination to abstract away from the identity and
relations of both the agent and the agent’s narrow circle. I, on the other hand, suggest
that appraisers are to use their imagination to abstract away only from their own relations
to both the agent and the agent’s narrow circle. In this way, Hume’s MPV (1) eliminates
the appraiser’s partiality and/or indifference to the agent and those affected by the agent,
as well as interference due to certain comparisons, and it also (2) incorporates sympathy’s
natural variability in a way that ensures uniformity and stability of feeling and judgment.

273
§VII.3.2 Impartiality & the artificial virtues
As discussed in chapter IV and once again earlier in this chapter, Hume argues
that our extensive social concern is derived from non-moral motivations, in particular
self-interestedness and limited benevolence. In Hume’s view, self-interestedness and
sympathy’s natural partiality for one’s family and friends (i.e., limited benevolence) give
rise to a willingness to subject one’s own self-interestedness and one’s own partiality
towards one’s family and friends to the regulatory principles of the MPV. These
normative principles produce an extensive concern for others.
I have argued that the principles of Hume’s MPV mitigate but do not eliminate
sympathy’s natural partiality. We’ve just seen examples in which Hume clearly suggests
that we do inculcate at least some of sympathy’s natural partiality within the MPV and
hence our moral judgments. My interpretation of the regulatory principles of Hume’s
MPV is consistent with this finding: the MPV requires that we sympathize most intensely
with those most closely associated with the agent. In short, the MPV’s requirement that
we approach equally near the agent when we disinterestedly sympathize with the agent’s
narrow circle appears to accurately reflect how we actually do tend to sympathize with
others when we issue moral judgments.
However, as Hume goes on to point out immediately after the last longer passage
quoted in the previous section, sometimes even this sort of partiality must be “remedied”
(T 489). There are instances in which our moral judgments are not supposed to reflect
any such partiality. Our moral judgments about issues of justice (e.g., the distribution of
rights and property) are a case in point. We do not think that justice admits of the sort of

274
partiality that many of the other moral virtues do. Justice is impartial; equal
consideration should be given to all, irrespective of any individual’s distance from the
agent, the agent’s act or trait, etc. Hume writes: “As our first and most natural sentiment
of morals is founded on the nature of our passions, and gives the preference to ourselves
and friends, above strangers; ‘tis impossible there can be naturally any such things as a
fix’d right or property, while the opposite passions of men impel them in contrary
directions, and are not restrain’d by any convention or agreement” (T 491). In other
words, Hume thinks that there are many instances in which society, our own family and
friends, and our own selves would benefit from totally eliminating sympathy’s partiality.
In Hume’s view, we create and accept various social conventions such as the rule
of law and its system of justice because the impartiality that they establish furthers our
natural concern for our own interests, our family and friends, as well as our more
extensive concern for others. Hume writes: “Instead of departing from our own interest,
or from that of our nearest friends, by abstaining from the possessions of others, we
cannot better consult both these interests, than by such a convention; because it is by that
means we maintain society, which is necessary to their well-being and subsistence, as
well as to our own” (T 489). Thus, self-interest and sympathy’s natural partiality for
one’s family and friends (i.e., limited benevolence) not only give rise to a willingness to
subject one’s self-interest and one’s own partiality towards one’s family and friends to
the MPV. According to Hume, these features sometimes give rise to a willingness to
implement various artifices that restrict the sort of partiality typically associated with the

275
MPV, thereby generating an impartial perspective by which we evaluate a particular act
or trait.
Hume refers to the virtues of justice, as well as our duty to the rule of law, as
artificial virtues precisely because they are products of such artifices. He contrasts these
artificial virtues with the natural virtues, which are not products of these artifices (T
473ff). Consequently, the aforementioned sort of permissible partiality is characteristic
of the natural virtues, whereas the artificial virtues tend to embody impartiality. Indeed,
impartiality itself is perhaps an artificial virtue in Hume’s view!
Furthermore, unlike the natural virtues, the artificial virtues often vary widely
across social contexts since different societies and cultures tend to construct different
rules of law that determine rights, property ownership, and other societal obligations.
219

This helps explain why the usual effects of some traits—in particular, the artificial virtues
and vices—tend to vary across different social settings. It also helps explains why, in
some cases, satisfaction of the MPV’s regulative principles not only requires that we have
knowledge of those with whom the agent has commerce, knowledge of their pleasures
and/or pains, and knowledge of their propinquity to the agent; it also requires knowledge
of the social context in which the agent’s trait is expressed. Different societies/cultures
conceive differently many of the social obligations that we have towards one another.
And many of these social obligations are relevant to our moral evaluations of the
tendency of acts or traits to fulfill or foil various ends that subsequently produce
pleasures and/or pains within those individuals typically affected by those acts or traits.

219
We might even expect the notion of impartiality itself, as it pertains to the artificial virtues, to vary
somewhat across cultures.

276
This does not mean that a society/culture’s system of justice is beyond moral
criticism however. Any system of justice is subject to moral evaluation.
220
That is, from
the MPV, we conceive of a social system’s usual effects upon those typically affected by
that system, and then we sympathize with those individuals. Observe that in this case we,
in effect, treat the system itself as the agent under evaluation. So those with whom we
are to sympathize most intensely will be the society to which the system belongs. And,
ceteris paribus, our sympathy with others typically affected decreases in proportion to the
distance they lie from that society—e.g., neighboring societies, nation states, etc.
§VII.3.3 Equanimity
Finally, there is one other way in which Hume’s MPV regulates any tendency that
we might have to be overly partial or indifferent in our sympathetic feelings, or to
entertain comparisons that contravene those feelings. Hume’s MPV requires that
appraisers evaluate the agent’s act or trait with equanimity, that is, with mindfulness and
a cool, calm demeanor. For instance, Hume speaks of the “calm judgments concerning
the characters of men”. Elsewhere he states that judicious spectators’ “joint verdict is the
true standard” because they judge “most coolly” from the disinterested sympathetic point
of view.
Because moral impressions are impressions of reflection, one’s prior emotional
states can dramatically impact one’s sympathetic and moral feelings. (See §III.2ff.)
Consequently, moral appraisers must be of calm temperament when morally evaluating
an act, trait, or person. They must also be mindful to include all relevant features in their

220
This would also include its particular conception of impartiality.

277
moral evaluations. In short, equanimity contributes to and fosters appraisers’ ability to be
both disinterested and sympathetic in the manner required by Hume’s MPV.
221

§VII.4 The MPV: a summary
We are now in a position to sum up chapters VI and VII and our discussion of
Hume’s MPV. Hume explicates the regulatory principles of the MPV according to our
capacity for causal/evaluative reasoning—i.e., reasoning about the causal tendency of
objects (e.g., acts or traits) to fulfill or foil particular ends that thereby has pleasurable or
painful effects—and our capacity for sympathy, which responds to those effects. He
divides those principles into two distinct but interrelated sets, namely: (1) the principles
that govern causal/evaluative reasoning, and (2) the principles that govern sympathy.
The principles that regulate our causal/evaluative reasoning are derived from our
understanding of the nature of cause and effect. The principles that regulate sympathy
are derived from our understanding of the nature of sympathy. By governing how we are
to reason about cause and effect, and how and with whom we are to sympathize, Hume
holds that the MPV secures stability and uniformity of moral feeling and judgment.
Regarding sympathy, Hume’s MPV requires appraisers to sympathize after a
certain manner with those causally affected by the agent’s act or trait. The manner of
which Hume speaks is both equanimity and disinterestedness. Appraisers must
disassociate themselves from their current situations and imaginatively place themselves
next to the agent and judge calmly and disinterestedly from “steady and general points of

221
In part two, we will consider empirical evidence that supports Hume’s contention that our prior
psychological states can have a serious impact upon our sympathetic and moral feelings. We will also
improve upon Hume’s account by providing more specific details and instructions about the sort of
disinterestedness and equanimity required by the MPV.

278
view.” We are to ignore our own (qua appraiser) ends/interests and only consider the
ends at which the agent’s act or trait is directed, as well as the ends/interests of those
typically affected by the agent’s act or trait—in particular, their end/interest in
experiencing pleasure and not experiencing pain, which are ends/interests that all sentient
beings share.
Hume is admittedly not entirely explicit about how the MPV’s general rules
govern sympathy. However, I have argued that Hume should be read as offering the
following account. In order to regulate sympathy’s natural variability, Hume’s MPV
requires appraisers to exclude as morally irrelevant: (i) their causal (blood) relations with
the agent and the agent’s narrow circle; (ii) their contiguity relations with the agent and
the agent’s narrow circle; and (iii) their resemblance relations with the agent, the agent’s
narrow circle, and the agent’s trait under evaluation. It also requires appraisers to include
as morally relevant: (a) the particular act or trait under evaluation; (b) the specific social
context C in which the act or trait is expressed; (c) the affected individuals’ particular
social location, or “sphere,” with respect to agent; (d) the range of the moral evaluation;
and finally, depending on the range of the evaluation and the particular act or trait under
evaluation, (e) various peculiarities of those who fall within the given range.
As I read Hume, this sphere of connection between the agent and those typically
affected by the agent’s trait or act is quite important. We are to consider the
resemblance, contiguity, and causal relations between the agent and all of those
individuals (typically) affected by his trait or act (or at least all of those who fall within
the given range of the evaluation). We are then to use our imagination to “approach

279
equally near” the agent so that the feelings of those who have the most intimate
connection with the agent registers most strongly on our sympathy. The result is that
those closest in psychological proximity to the agent have the most weight and the weight
decreases as the aforementioned relations decrease. But appraisers are also to sympathize
with those individuals typically affected such that the intensity of their sympathetic
feelings is proportional to the intensity of the pleasure and/or pain of those individuals.
In these respects, Hume’s system of ethics can be conceived of as a kind of weighted
universalism.
Although sympathetic partiality is, in this way, built into Hume’s MPV, the MPV
uses its requisite disinterestedness and equanimity to regulate sympathy’s natural
variability so as to eliminate indifference and excessive partiality. For instance, the
MPV’s disinterestedness serves to prohibit any comparisons that interfere with such
sympathy, including those that give rise to indifference or excessive partiality. It also
necessitates comparisons that contribute to such sympathy.
222

However, Hume also claims that there are many instances in which we create
artifices to eliminate sympathy’s natural variability and partiality altogether and to assess
the agent’s act or trait from an impartial—rather than a merely disinterested—moral
point of view. Issues concerning justice and other societal obligations such as property
rights are paradigmatic examples. In such cases, equal consideration should be given to
all those typically affected by the agent’s trait or act, irrespective of their particular
sphere of connection with—i.e., their psychological proximity to—the agent.

222
An insightful summary of Hume’s account can be found in T 602-3.

280
According to Hume, we create these artifices and encourage impartiality within
certain contexts for much the same reasons that we encourage a general concern for
others and enforce corrective principles that regulate sympathy’s natural variability.
Recall that, according to Hume, we tend to encourage acts and traits that exhibit a general
concern for others (e.g., through praise) and discourage acts and traits that don’t (e.g.,
through punishment) because overtime we come to realize that possessing and acting
upon a more extensive concern for others has general positive effects for both our loved
ones and ourselves.
223
We also come to realize that failing to possess and act upon this
extensive concern has general negative effects for our loved ones and ourselves.
Consequently, we establish and enforce regulative principles that ensure, as best as
possible, that persons will both possess and act upon this extensive concern.
Similarly, overtime we come to realize that, in certain contexts, acting in impartial
ways toward others tends to have beneficial affects. We also come to realize that, within
these contexts, failing to act in impartial ways toward others tends to have deleterious
affects. Consequently, society establishes various artifices to encourage traits and acts
that exhibit impartiality in such contexts and discourage those acts and traits that fail to
exhibit impartiality.
Nevertheless, as Hume observes, we also do not expect “impossibilities” from
humankind (T 602-3). Just as we don’t expect people to act upon a general concern for

223
Evidence that a sympathetic concern for others is generally beneficial can be observed from the fact that
a sympathetic capacity has evolved through natural selection, and that a concern for others is exhibited and
encouraged in those who are presumably not entirely aware of its beneficial affects, namely, the great apes.
Thus, it appears that there were already natural and social mechanisms in place to encourage a concern for
others before humans began to consciously do so. This is discussed in more detail in part two of the
project.

281
others in each and every case and ignore their own competing interests and the competing
interests of their family and friends, we do not expect people to act impartially in each
and every case and ignore their own competing interests and the competing interests of
their family and friends. In fact, as discussed, we sometimes morally condemn others for
failing to behave in a somewhat partial manner. In establishing the regulative principles
of the MPV, we therefore take these important facts about human psychology into
account and inculcate into the MPV at least some of sympathy’s natural partiality in
many instances—in particular, those pertaining to the “natural” virtues.
§VII.5 The conclusion of part one
Hume’s presentation of the regulative principles governing sympathy is not as
lucid or systematic as his presentation of the principles governing causal reasoning. But I
suggest that a careful reading of Hume is sufficient to dismiss the claim that his
conception of the MPV is too vague to serve as a regulatory device for sympathy and
hence moral feeling and judgment. However, even though the MPV is arguably
sufficient to serve as a regulatory device for sympathy, it remains to be seen whether the
regulatory principles of the MPV are sufficient to ensure in each and every case the
requisite level of interpersonal agreement among those who successfully govern their
feelings and judgments according to those principles. Clearly the success of Hume’s
moral theory rests on the success of the MPV to produce interpersonal agreement among

282
appraisers; for without at least a certain level of interpersonal agreement among
appraisers, Hume’s moral ontology is untenable.
224

Hume undoubtedly maintains that the regulative principles of the MPV are
sufficient to ensure the requisite agreement. He appears to be confident that this is the
case for the following reasons. First, the MPV’s regulative principles serve as norms
governing appraisers’ sympathetic and moral feelings and thereby their moral judgments.
And both our sympathetic and moral feelings and thus our moral judgments are products
of our psychology. Hence the MPV provides norms that in effect govern aspects of our
psychology. Second, Hume believes that human beings share the same basic psychology.
From this, Hume concludes that, when successfully applied, the norms of the MPV will
produce roughly the same sympathetic and moral feelings, and hence validate the same
moral judgments in all appraisers. Stability and uniformity, as well as interpersonal
agreement among appraisers, is therefore to be expected.
But is Hume’s confidence that his proffered regulative principles will produce
interpersonal agreement justified? I am somewhat skeptical. While the principles of
Hume’s MPV may be sufficient to secure interpersonal agreement in some cases, it is far
from clear that, as they stand, Hume’s general rules are sufficient to guarantee the sort of
uniformity of feeling and interpersonal agreement in belief that his moral theory needs.
Part of the problem is that, as Williams James might have put it, Hume’s regulatory
principles don’t seem to give sufficient practical instructions for bringing about the

224
Just as we do not need complete agreement among appraisers regarding each and every shade of color in
order to uphold a dispositional, impression-dependency conception of color qualities, we do not need total
agreement among appraisers in order to uphold a dispositional, impression-dependency conception of
moral qualities. We shall discuss this point in more detail in part two.

283
requisite goal of interpersonal agreement in feeling and judgment.
225
It is one thing to
observe that sympathy, disinterestedness and equanimity are necessary conditions for
bringing about the goal of interpersonal agreement in moral feeling and judgment. But it
is quite another thing to provide the sort of practical instructions that directs moral
appraisers how to sympathize in a disinterested and equanimous manner that also yields
interpersonal agreement. I am skeptical that Hume has accomplished this.
This is not to say that Hume’s moral theory should be totally abandoned however.
Although Hume’s MPV, as he presents it, is unsuccessful insofar as it (arguably) fails to
deliver the requisite level of interpersonal agreement, this does not entail that no set of
regulative principles and practical instructions, when properly applied, can produce that
level of interpersonal agreement among appraisers. There is reason to think that there are
principles and practical instructions that, when properly applied, can regulate our
psychology so as to produce a general concern for others. Whether we also have reason
to think that these principles and practical instructions, when properly applied, can
regulate our psychology in a way that yields interpersonal moral agreement remains to be
seen. Assuming that the nature of morality is rooted in our psychology and that human

225
William James writes:
The faculty of voluntarily bringing back a wandering attention, over and over again, is
the very root of judgment, character, and will. No one is compos sui if he have it not. An
education which should improve this faculty would be the education par excellence. But
it is easier to define this ideal than to give practical instructions for bringing it about.
(Principles of Psychology, 1890, author’s emphasis)
Of course, Hume’s MPV requires more than that appraisers simply maintain a focused attention.
It also requires appraisers to focus their attention on morally relevant matters in an equanimous and
disinterested manner, and sympathize with a particular group in a particular way. As Hume would have it,
an education that improves upon all of these faculties would indeed be the moral education par excellence.
But the problem is that Hume’s MPV arguably lacks the sort of practical instructions needed for such a
moral education to take place.

284
psychology is at least somewhat uniform, the level of our understanding of the causal
mechanisms underlying the various psychological processes that affect our sympathetic
and moral feelings will be a crucial determinant of our success in developing such
regulative principles and practical instructions. As our knowledge of these psychological
processes and causal mechanisms increase so will our prospects of success.
Of course, the assumption that the nature of morality is rooted in human
psychology is a controversial one. Consequently, Hume’s moral ontology and his
conception of the MPV inherit this controversy. As we shall now discuss in part two of
this project, there is some scientific and philosophical evidence in support of the claim
that the nature of morality is rooted in human psychology, as well as the claim that
ceteris paribus humans share the same basic psychology.
226
The scientific evidence
presented in part two is gathered from such fields as primatology, social psychology, and
cognitive neuroscience.
The philosophical evidence presented in part two is gathered from its
reconstruction of Hume’s moral theory. Part two, in effect, uses this reconstruction to
lend plausibility to Hume’s moral ontology by demonstrating its potential to explain the
intimate and rather peculiar connection between moral judgment and human conduct in
such a way that resolves the accommodation problem.
227
After laying out its basic
Humean framework, part two reconsiders whether we have reason to think that we can
develop various principles and practical instructions that, when properly applied, can

226
To recapitulate, the ceteris paribus caveat is due to the fact that some human beings are psychologically
abnormal, e.g., sociopaths or those who are severely mentally impaired.

227
See chapter I for discussion of the accommodation problem.

285
regulate our psychology in a way that yields the requisite level of interpersonal moral
agreement. I suggest that there is reason for cautious optimism. We shall now take up
this reconstructive exercise in part two.

286

PART TWO: Reconstructing Hume’s Moral Theory

287

Chapter VIII: Hume’s Moral Ontology

§VIII.1 Introduction

Our main goal in part two is to lend support to Hume’s moral ontology and its
foundations—in particular, the general roles of sympathy, causal reasoning, and the
moral point of view (MPV) in producing the moral feelings that determine moral
properties—by demonstrating their potential to explain the intimate and somewhat
peculiar relationship between moral thought/judgment and human conduct. As discussed
in chapter I, this amounts to accommodating various presuppositions of moral practice
and discourse. Recall that among the commonsense assumptions to be accommodated
are the following: (1) moral evaluation/thought is descriptive—i.e., moral evaluations/
thoughts are straightforward beliefs; hence (2) our moral judgments/evaluations ascribe
moral properties to objects; (3) at least some moral judgments/beliefs are true; therefore,
(4) there are moral properties and moral facts; (5) both error and moral disagreement are
not only real possibilities but plausibly often do occur—i.e., we can be mistaken and
genuinely disagree about the moral facts, as well as the truth-value of a given moral
judgment/belief; and finally (6) moral thought/judgment has a direct link with motivation
and action, as well as (7) a special kind of normative authority (or reason-giving force),
namely, one that applies to individuals independently of their particular attitudes.
Each of these commonsense presuppositions (1-7) involves positive ontological
commitment. For instance, presuppositions 1-3 entail positive commitments concerning
the existence and nature of moral expressions’ putative truth-value; presuppositions 4 and

288
5 involve positive commitments with respect to the existence and nature of the moral
objects—e.g., properties, facts, and/or reasons—putatively denoted by moral expressions;
presupposition 6 involves positive commitment regarding the existence and nature of the
connection between those putative moral objects and motivation; and 7 entails positive
commitments concerning the existence and nature of moral expressions’ putative
normative character (or reason-giving force).
228

Recall that the challenge of accommodating these positive commitments is that
they raise certain burdens of explanation, and these explanatory burdens prima facie
appear to subvert at least some of those positive commitments and thereby threaten the
commonplace view of morality. For instance, one significant burden is to explain how an
epistemologically accessible moral property can possess the sort of practicality—i.e., the
rather peculiar connection with normativity and motivation—that the commonplace view
of morality suggests.
229
In short, we face a kind of expository struggle. On the one hand,
accommodation of the commonplace view of morality necessitates that we take on certain
positive ontological commitments. On the other hand, accommodation of the explanatory
burdens raised by these commitments is thought to entail that we must give up at least
some of those very commitments. I have called this expository struggle—i.e., the tug
upon our commitments in opposing directions—the accommodation problem.
230

228
Whether or not such positive ontological commitments amount to a commitment to, and presumption of,
moral realism remains to be seen. I shall postpone taking up this issue until the end our project.

229
See chapter I for a brief discussion of this peculiar connection with normativity and motivation.

230
I borrow this terminology from Mark Timmons (1999). See chapter I footnote 7 for a discussion of the
difference in Timmons’ conception of the accommodation as compared to my own.

289
For purposes of clarification and exposition, this project divides the
accommodation problem into three distinct but interrelated steps: ontological, semantic,
and normative accommodation. The first step in resolving the accommodation problem
is to provide ontological accommodation, namely: to offer an account of the nature of
moral properties and facts that can accommodate the various features of our moral
practice and discourse. The two remaining steps are to demonstrate how this moral
ontology can indeed accommodate the semantic and normative features of moral
discourse that our moral practice appears to presuppose. Semantic accommodation, as
such, is basically a matter of explaining how moral thought and language express moral
properties and facts. Normative accommodation primarily involves explaining how these
moral properties and facts can have a normative authority or reason-giving force that is
independent of individuals’ actual motivations.
I have suggested that Hume’s moral ontology has significant potential to provide
ontological accommodation. Hume conceives of moral qualities as a “relation…between
the form [of an object] and the sentiment [in the judge]” (SOT, paragraph 10). The
interpretive part of the project—viz., part one—has argued that this claim in effect
amounts to the view that moral qualities are intersubjective, dispositional, feeling-
dependent properties. To recapitulate, Hume holds the following general characterization
to be true: an object x has a moral property if and only if observation or contemplation of
the typical emotional effects of x (in C) causes, via sympathy, feelings of approval or (as
the case may be) disapproval toward x within those who satisfy the regulative principles

290
of the MPV. In short, in Hume’s view, the moral domain is determined by particular
conative feelings had under certain conditions.
Of course, the question remains whether Hume’s moral ontology can
accommodate the semantic and normative features of moral discourse that our moral
practice appears to presuppose. To this end, part one has argued that, in Hume’s view,
our moral judgments ascribe these intersubjective, dispositional, sentiment-dependent
properties to objects. Hence our moral evaluations are descriptive, property ascribing,
truth-evaluable judgments, some of which are true. There are moral properties and moral
facts, and both moral error and disagreement, including legitimate disputes over a given
moral judgment/belief’s truth-value, are not only possible but often actually occur. Thus,
as I read Hume, he offers an account that is generally congruous with presuppositions 1-5
listed above.
231
Moreover, Hume’s account of the nature of moral properties—i.e.,
properties that are determined by particular conative attitudes had under certain
conditions—can account for an intimate relationship between moral judgment/thought on
the one hand and motivation and action on the other. (See presupposition 6.)
This is not sufficient to resolve the accommodation problem however. First, as
things stand, we do not yet have a semantic account that tells us how our moral
judgments pick out these attitude-dependent properties. For instance, which particular
terms denote these properties? Semantic accommodation requires an answer to this
question. Second, it is not yet clear whether the sort of intimate relationship that falls out
of Hume’s moral ontology is sufficient to account for the extent to which our moral

231
However, see chapter III for a critique of Hume’s account of the relation of thought and semantic
content.

291
judgments are practical. For instance, we’ve not yet addressed whether such an account
can accommodate the putative normative authority, or reason-giving force, of moral
considerations on an agent irrespective of that agent’s desires, beliefs, or intentions.
232

(See presupposition 7.) Normative accommodation requires that the account
accommodate this presupposition. Third, we have also yet to defend a crucial underlying
presumption of Hume’s account, namely, the presumption that anyone who satisfies the
conditions of the MPV will have a feeling of approval or (as the case may be) disapproval
toward the object under evaluation, and have that feeling to roughly the same degree.
Indeed the prevalence and extent of moral disagreement raises some doubt about the
plausibility of this presumption. And this, in turn, raises doubts about the plausibility of
Hume’s moral ontology. Hence ontological accommodation requires a defense of this
underlying presumption.
In the remainder of this chapter, we will begin a contemporary defense of Hume’s
moral ontology and its foundations by briefly considering some empirical evidence that
lends at least some support to that account. The following chapter offers a theory of the
semantics of moral evaluative language consistent with both Hume’s moral ontology and
its foundations, as well as the current scientific evidence regarding moral cognition that
will be presented in this chapter. The general framework of this semantic account
appears to have all the features needed to provide semantic accommodation. This

232
Part one (§V.6) did suggest that our judgments about the causal tendency of objects to fulfill (or foil)
particular ends have an important normative function, namely: they can be used to evaluative behavior and
give others advice about how to act. But we have yet to discuss whether, and if so how, this provides the
sort of normative authority/force that our moral judgments appear to presuppose. (See chapters I and X for
further discussion of the sort of normative authority/force that our moral judgments appear to presuppose.)

292
framework also helps the project address the problem of moral disagreement: it offers a
means by which we can assess the extent of moral disagreement and mount a defense of
the presumption of interpersonal agreement among those who satisfy the conditions of
the MPV. Finally, this framework also plays an important role in providing normative
accommodation. The semantic account offers insight into the nature of normativity; and
as we shall see, when this account is combined with Hume’s (or at least a Humean) moral
ontology, we appear to have all the features needed to accommodate the peculiarities of
moral normativity. If successful, this project will also have thereby demonstrated how to
provide ontological accommodation and resolve the accommodation problem.
§VIII.2 Empirical evidence in support of Hume’s account
In the remainder of this chapter, we shall consider some of the empirical evidence
that facilitates Hume’s moral ontology and its foundations. As we shall see, there are
some important respects in which the proposed account differs from Hume’s, but this
section still insists that the proffered moral ontology and the foundations underlying that
ontology are nonetheless, at least generally speaking, Hume’s.
Much of Hume’s account is, I believe, correctly viewed as a phenomenalistic
approach. For instance, his fundamental distinction between impressions and ideas, as
well as his rather intricate account of their relations, is basically an attempt to distinguish
and explain the contents of our minds according their phenomenal “vivacity” or strength.
This explains why Hume primarily appeals to introspective “experiments” to support his
account. Nevertheless, Hume does think that various physiological, causal processes
underlie the mental phenomena to which he appeals. In other words, Hume is not purely

293
a phenomenalist. He believed that science would some day provide physiological
confirmation of his introspective experiments.
I shall argue that Hume’s belief is, at least generally speaking, currently in the
process of being vindicated. I say “at least generally speaking” because, even though he
does not speculate about the physiological processes involved,
233
many of Hume’s
introspective experiments commit him to various substantive claims—consider, for
example, his views (discussed in chapter III) concerning the indispensable role of the
“passions” in accounting for the motivational impetus of moral considerations. It is
highly unlikely that many of his more detailed substantive commitments will be born out
by the empirical evidence. Nevertheless, regarding his more general claims about the
nature of morality, Hume is basically correct. For instance, as we shall see, the empirical
evidence suggests that Hume is correct about the crucial role played by our
considerations of the tendency of objects to fulfill (or foil) various ends or goals when we
issue moral judgments. Hume also appears to be generally correct about the crucial role
played by our moral feelings, as well as the role of our psychological capacity for
“sympathy” in evoking those feelings.
As discussed in part one, in the Treatise Hume conceives of “sympathy” as a
psychological process that converts one’s ideas of others’ pains and pleasures into like
feelings within oneself. In the Enquiry, on the other hand, Hume does not discuss the
distinction between impressions and ideas; yet he still holds that our sympathy is the
foundation of our moral feelings. This, to my mind, suggests that that distinction is not

233
See, for instance, T 275-6.

294
crucial to the psychological process of sympathy. Sympathy can be understood simply as
a psychological process (with physiological underpinnings) that engenders what Hume
refers to as a “fellow-feeling” (EPM 219). According to Hume, it is sympathy that
“actuates our bosom, and interests us in the felicity of the person whom we contemplate”
(EPM 231). In short, sympathy is what generates our feelings of concern for others, or
what Hume calls “limited benevolence”. And according to Hume, when sympathy is
subjected to various regulatory principles, the resulting “extensive” sympathy is
responsible for “extensive benevolence” or “the sentiment of humanity”; i.e., it generates
our interest in the general “public good”.
234

Current literature in cognitive neuroscience and social psychology offers similar
accounts, although they use the term “empathy” rather than “sympathy”. For instance,
empathy has been defined as:
• An affective response that stems from the apprehension or
comprehension of another’s emotional state or condition, and which is
similar to what the other person is feeling or would be expected to feel
in the given situation
235

• An other-oriented emotional response congruent with the other's
perceived welfare
236

• An affective response more appropriate to someone else’s situation
than to one’s own
237

234
See chapters IV and V.

235
Eisenberg, N. (2000).

236
Batson, C.D., Sager, K., Garst, E., Kang, M., Rubchinsky, K., and Dawson, K. (1997).

237
Hoffman, M.L. (1982).

295
• The ability to experience and understand what others feel without
confusion between oneself and others.
238

• The ability to put oneself into the mental shoes of another person to
understand his or her emotions
239

Thus, according to current literature: “‘Empathy’ denotes, at a phenomenological
level of description, a sense of similarity between the feelings one experiences and those
expressed by others” (Decety and Lamm, 2006: 1146). It is noteworthy that all of the
above definitions of empathy share the following features with Hume’s account. First,
like Hume, they all appeal to a phenomenological level of description. Second, they all
point to an agent’s emotional response to others’ perceived or anticipated (or imagined)
emotional states in a given situation. And, like Hume, they all thereby involve self-other
differentiation linked with motivation.
240
This latter feature is important because it serves
to distinguish empathy from mere emotional contagion or personal distress. Empathy is
not merely a matter of sharing feelings; it is an “other-oriented” social emotion.
241

While this account of empathy clearly has a great deal in common with Hume’s
account of sympathy, it is not at all obvious that it in any way abets Hume’s moral
ontology. Nor is it clear whether empathy can produce a more general concern for others
and pro-social behavior, even supposing that it can somehow be regulated. In order to

238
Decety, J. and Lamm, C. (2006).

239
Goldman, A. (1993).

240
This is most explicit in the Treatise. Recall from part one (§III.3 and §IV.2) that sympathy, being an
instance of the association of ideas and impressions includes: (i) conceptions of self—the source of
vivacity—and other—the “object” of sympathy; ii) the idea of pain (or pleasure)—the “cause” of
sympathy; and finally, (iii) an impression of pain (or pleasure)—the “quality” of sympathy.

241
Decety and Lamm (2006).

296
make the case that it lends support to Hume’s moral ontology, and that we can regulate
empathy to yield a concern for others and thereby pro-social behavior, we need to take a
closer look at empathy and our current understanding of its physiological underpinnings.
§VIII.2.1 The origins & workings of empathy
As a start, it is noteworthy that empathy is not an all-or-nothing phenomenon.
There is compelling evidence that there are many intermediate forms of empathy between
the extremes of mere agitation at the observed distress of another, on the one hand, and a
genuine concern for another’s overall wellbeing that results from a full understanding of
the individual’s predicament.
242
The full gamut of the various forms of empathy can be
observed in humans, while we tend to find only the more rudimentary forms of empathy
in non-human mammals.
243
However, it is noteworthy that quite advanced forms of
empathy have been observed in non-human primates.
For example, Frans de Waal (2006) discusses a case in which a particularly
difficult chimpanzee refused to come down from the top of her cage. One of the handlers
discovered that she could coax her down by pretending to cry. Upon hearing the cry, the
chimp climbed down, approached the handler, and assessed the situation, attempting to
determine why the handler was crying. Failing to make this determination, the chimp
consoled the handler, gently touching her face. In the same manuscript, de Waal also
describes a situation in which a chimp finds an injured bird, picks it up, and climbs to the
top of a tree. Clutching onto the tree with her feet, the chimp used both hands to spread

242
See, for instance, Preston, S.D. & de Waal, F.B.M. (2002).

243
See, for instance, de Waal (1996).

297
the bird’s wings and tried to help it fly away. In both cases, these primates showed
empathic concern not merely for non-kin but individuals belonging to different species.
Evolutionary biologists suggest that empathic helping behavior has evolved
precisely because of its contribution to genetic fitness (kin selection). In humans and
other mammals, an empathic impulse to care for offspring is almost certainly genetically
hardwired. The examples above suggest that even more complex forms of empathy are
also predominantly hardwired, although various social pressures have also undoubtedly
shaped them. This squares with Hume’s claim that “sympathy” is innate but malleable.
When viewed from an evolutionary perspective, these results are not entirely
surprising. Evolutionary processes start with quite simplistic systems and operations, and
then begin to construct layers of increasing complexity. For instance, empathy itself has
more rudimentary components that precede it, such as emotional contagion and somatic
mimicry—i.e., “the tendency to automatically mimic and synchronize facial expressions,
vocalizations, postures, and movements with those of another person, and consequently
to converge emotionally.”
244
Rudimentary forms of motor mimicry can be found in most
mammals. Such lower levels of organization form the building blocks upon which higher
levels of organization, such as empathy, depend.
For example, research demonstrates that viewing others’ facial expressions
triggers similar expressions on one’s own face, even in the absence of conscious

244
Decety and Lamm (2006: 1149). See also Hatfield, E., Cacioppo, J.T., and Rapson, R.L. (1993).

298
recognition of the stimulus.
245
It is argued that such automatic behavioral mimicry may
have had survival value by helping animals to communicate with each other. This
proposal is consistent with social psychology experiments showing that humans tend to
non-consciously mimic the behaviors of others,
246
which leads to smoother interactions
and increased liking and affiliation, serving to foster relationships with others.
247
It is
noteworthy that people who exhibit this so-called “chameleon effect” (ibid) to a greater
extent also have higher levels of dispositional empathy.
Emotional contagion is of a piece with motor mimicry, both of which function as
automatic, non-conscious perception-reaction mechanisms. Accumulating evidence
suggests that similar neural networks are stimulated when humans experience emotions
and when they perceive others expressing emotions.
248
For example, functional
neuroimaging (fMRI) studies reveal that brain areas implicated in processing the
affective and motivational aspects of pain—i.e., aspects of pain that pertain to desires,
urges, or impulses to avoid or terminate a painful experience—are also active when
observing pain in others.
249
In other words, merely observing another individual in a

245
Dimberg, U., Thunberg, M., and Elmehed, K. (2000). See also Rizzolatti, G. and Craighero, L. (2004)
for review of the literature on the mirror-neuron system.

246
Chartrand, T.L. and Bargh, J.A. (1999).

247
Lakin, J. L., Jefferis, V. E., Chen, C. M., and Chartrand, T.L. (2003).

248
Wicker, B., Keysers, C., Plailly, J., Royet, J. P., Gallese, V., and Rizzolatti, G. (2003); Carr, L.,
Iacoboni, M., Dubeau, M. C., Mazziotta, J. C., and Lenzi, G.L. (2003).

249
See, for instance, Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R. J., and Frith, C.D. (2004)
and Morrison, I., Lloyd, D., di Pellegrino, G., and Roberts, N. (2004).

299
painful situation automatically yields pain-related responses in the brain regions/neural
networks associated with the motivational-affective dimension of pain in oneself.
250

Another fMRI study showed that observing fearful body expressions not only
produces increased activity in brain areas associated with emotional processes, but also in
areas linked with representation of action and movement.
251
Thus, the mechanism of fear
contagion automatically prepares the brain for action. Another study
252
strongly suggests
that the brain regions associated with the affective-motivational part of the “pain matrix”
are not specific to the sensory qualities of pain, but are also associated with more general
survival mechanisms such as aversion and withdrawal. Clearly such an automatic
perception-action mechanism has adaptive value for the survival of individuals.
In sum, current evidence suggests that “bottom-up information processing”
253

components such as motor mimicry and emotional contagion, which are automatic and
often occur unconsciously, function as an important but rudimentary aspect of empathy.
By coupling perceptions and actions together, these components appear to contribute to
our ability to perceive and understand what others feel, as well as others’ associated

250
While there is a strong overlap between first-hand experiences of pain and the perception of pain in
others, there is not complete congruence between self and other phenomenal experience. Clearly, we do
not feel the pain of others in the same way that we feel our own pain. Current anatomical studies related to
pain experience and perception suggest that this disparity is due to the fact that first-hand experience has a
different organization of activity in the affected brain regions than perception of pain. See, e.g., Jackson, P.
L., Brunet, E., Meltzoff, A. N., and Decety, J. (2006); Jackson, P. L., Rainville, P., and Decety, J. (2006).

251
De Gelder, B., Snyder, J., Greve, D., Gerard, G., and Hadijkhani, N. (2004).

252
Isomura, Y. and Takada, M. (2004).

253
I borrow this terminology from Decety and Lamm (2006).

300
intentions.
254
They also prepare us to act in response to our perceptions or anticipation of
others’ feelings and intentions.
255

Significantly, this understanding of others’ feelings aids in the synchronization of
representations of self and other,
256
which is key to more complex forms of empathy. For
instance, this yoking together of perceptions of self and the other enable us to engage in
role taking, which has been linked to the development of empathy, moral reasoning, and
more generally, prosocial behavior.
257
By means of our imagination, we can adopt the
perspective of others and come to experience sensations, which are generally similar to,
although typically weaker than, those of the other person
258
—just as Hume claims.
However, the development of self and other mental-state understanding is also
functionally linked to higher-level “top-down information processing”
259
components.
260

For example, unlike the automatic aspects of empathy (motor mimicry and emotional
contagion), the ability to consciously take the perspective of others “heavily draws on the
maturation of executive resources (i.e., the processes that serve to monitor and control
thought and actions, including self-regulation, planning, cognitive flexibility, response

254
Grosbras, M.H. & Paus, T. (2006); Singer et al (2004).

255
Jackson, P.L., Meltzoff, A.N., and Decety, J. (2005).

256
Decety and Lamm (2006).

257
Ibid.

258
Ibid.

259
I have borrowed this terminology from Decety and Lamm (2006).

260
See, for instance, Russell, J. (1996) and Zelazo, P.D. (2004).

301
inhibition, and resistance to interference) of the prefrontal cortex, which continues to
mature from birth to adolescence” (Decety & Lamm, 2006: 1151).
261

These executive resources (i.e., cognitive top-down mechanisms) involve much
higher levels of complexity than bottom-up mechanisms such as motor mimicry and
emotional contagion. This is evidenced by the fact that these top-down mechanisms are
associated with the most recently evolved area of the brain, namely, the prefrontal cortex.
While executive resources such as voluntary inhibitory control are exhibited to varying
degrees in other animals, they are most fully expressed in humans. This is quite
significant, for our unrivalled capacity for executive control—i.e., self-governance—is
largely responsible for the fact that we, unlike the rest of the animal kingdom, are moral
creatures, or so I shall argue.
262
Executive control contributes to empathy, moral
reasoning, and empathetic concern, which tends to produce prosocial—i.e., other-
regarding—behavior. It does so in part by regulating empathy’s natural variability.

261
Carlson & Moses (2001) have documented the contribution of the prefrontal cortex’s (PFC) executive
functions, especially inhibitory control, in both the emergence and expression of mental state attribution in
children. There is also clear evidence [Perner & Lang (1999)] of a developmental link between theory-of-
mind development and improved self-control at about the age of 4. Using a brain-imaging technique (viz.,
positron emission tomography), Raine et al (1994) found that murderers, as compared to normal subjects,
exhibited decreased activity in the PFC. Raine et al (2000) also found that individuals with antisocial
personality disorder have, on average, eleven percent less prefrontal gray matter than control subjects.

262
One could argue that, in this respect, our capacity for self-governance represents a sea change. (See,
e.g., the claims of C. Korsgaard and P. Singer in de Waals [2006].) While this may be in some sense true,
it is also noteworthy that empathetic behavior of other animals is a precursor of human morality. Just as
human bottom-up mechanisms are more sophisticated evolutionary offspring of our genetic ancestors’
bottom-up mechanisms, so too are human top-down mechanisms (i.e., our executive resources) offspring of
our ancestors’ top-down mechanisms. Small changes can have dramatic effects!

302
§VIII.2.2 Empathy’s variability
We can observe the distress and predicament of others on a daily basis, and yet
we do not always respond in prosocial ways. This is due to the fact that empathy is
variable—just as Hume claims. Several intra- and interpersonal factors have been
identified that facilitate or inhibit the occurrence and degree of our empathic responses.
An important interpersonal factor is the type of relationship (including our conception of
that relationship) and associated attachment one has with others. Important intrapersonal
factors include: one’s prior emotional states, one’s prior experience with (and knowledge
of) a situation, and one’s ability to cope with the distress of others. We shall address
these factors in the order just presented.
First, the type of relationship and the associated attachment (e.g., parents and their
offspring) often affects how we emotionally respond to others. Researchers have found
that the closer the perceived association between others and ourselves, the stronger our
affective attitudinal responses or dispositions—just as Hume claims. It is hypothesized
that this is the result of the fact that empathy has evolved as an in-group phenomenon.
263

Several studies have also shown that our own affective dispositions, or attitudes,
vary depending on whether the other is conceived as a cooperator or a competitor. And
this, in turn, influences whether we react with a congruent or a non-congruent emotion to
another’s affect. For instance, Lanzetta and colleagues investigated the psycho-
physiological, behavioral, and psychological effects of attitudes on interpersonal

263
Preston, S.D. and de Waal, F.B.M. (2002).

303
interaction.
264
Their main finding is that, while cooperative settings result in empathetic
responses, competitive relationships lead to counter-empathic responses. For example,
psycho-physiological measures indicate that participants reacted to a painful shock of
competitors with little arousal and distress, but got distressed when seeing them in joy.
The reverse pattern was obtained for cooperators.
These findings are quite consistent with—indeed they are predicted by—Hume’s
account. The above findings also reflect an important and often ignored aspect of
empathy noted by Decety and Lamm (2006). Due to empathy’s interpersonal variability,
our empathetic ability can also be used in malicious ways, such as when knowledge about
the emotional or cognitive state of competitors is used to harm them (e.g., when enemy
soldiers are attacked after emotionally exhausting them by continuous artillery barrage).
These interpersonal factors do not necessarily occur in isolation with the
intrapersonal factors, but are likely to interact with them. The first noteworthy
intrapersonal factor is one’s prior emotional state. It has been shown that experimentally
induced emotional states affect the ability of observers to recognize the facial display of
emotion.
265
Such “emotional priming” enhances early visual processing in normal
persons.
266
It is hypothesized that this might, in turn, affect the visual perception of
emotional cues, such as the subtle changes in the facial expression of emotion. Other

264
Lanzetta, J.T. and Englis, B.G. (1989); Englis, B.G., Vaughan, K.B., and Lanzetta, J.T. (1982).

265
Niedenthal, P.M., Halberstadt, J.B., Margolin, J., and Innes-Ker, A.H. (2000).

266
Phelps, E.A., Ling, S., and Carrasco, M. (2006).

304
emotional states such as anger and depressive mood are also known to affect our
perception of the world, including our perception of the expression of emotions by others.
Of particular interest to our project, is research that demonstrates that our prior
emotional states can affect our moral judgments by affecting our empathic response.
267

Before being presented with a hypothetical moral dilemma,
268
half of a group of subjects
watched five minutes of Saturday Night Live, and the other half watched an uninspiring
documentary about a Spanish village. As compared to those who watched the dull
documentary and a control group in which no prior emotional state was induced, those
enlivened by the comedy show were more likely to say that the right response in the
moral dilemma is to push a man off a bridge to his death in order to save the lives of five
people on a runaway trolley. Researchers concluded that the negative empathic emotional
responses evoked by contemplating the “up-close-and-personal” homicide were offset by
the positive emotions evoked by the entertaining TV show. Presumably, in a real life
situation, our empathetic response would not be so easily offset. Nonetheless, the study
does demonstrate that our prior emotional states can have a dramatic impact on our
empathetic emotional responses and thereby our moral judgments—just as Hume claims.
One’s prior experience is another important intrapersonal factor that affects our
emotional resonance with others. For example, promoting knowledge about the specific
needs and problems of elderly people in a game situation resulted in significant increases

267
Valdesolo, P. and DeSteno, D. (2006).

268
The moral dilemma was taken from Judith Jarvis Thompson’s (1986) now famous “trolley-problems”.
We shall discuss these trolley problems later in the chapter. Experiments performed using these moral
dilemmas reveal that our empathetic emotions play an important role in guiding our moral judgments.

305
in empathy and caring for elderly patients.
269
The game promoted this knowledge by
including various exercises that simulate problems the elderly typically face, e.g.,
wearing a pair of heavy rubber gloves to simulate decreased manual dexterity, and
wearing goggles with a film of clear tape over the lens to simulate cataracts.
Finally, one’s ability to cope with others’ distress affects our empathetic response.
People who cannot control their emotions are more prone to experience empathic over-
arousal in the form of personal distress (i.e., an aversive, self-focused emotional reaction
to the apprehension or contemplation of another’s emotional state or situation).
270
In
contrast, previous research has shown that those who can exert control over their
emotions are more prone to experience empathetic concern (i.e., an other-oriented
emotional response congruous with the perceived or contemplated emotional state or
situation of the other person) rather than personal distress.
271

§VIII.2.3 Regulating empathy
As we have seen, several features have been identified as contributing factors to
the variability in our empathetic emotional responses to others. There are also several
cognitive top-down mechanisms and strategies that exert cognitive control over our
emotions, allowing us to regulate the occurrence and the degree of our empathic
responses.
272
Emotion regulation is defined as “the process of initiating, avoiding,

269
Varkey, P., Chutka, D.S., and Lesnick, T.G. (2006).

270
Eisenberg, N., Shea, C.L., Carlo, G., and Knight, G. (1991).

271
Derryberry, D. and Rothbart, M.K. (1988).

272
Gross, J.J. (1998).

306
inhibiting, maintaining, or modulating the occurrence, form, intensity, or duration of
internal feeling states, emotion-related physiological processes, emotion-related goals,
and/or behavioral concomitants of emotion, generally in the service of accomplishing
one’s goal” (Decety and Lamm, 2006: 1156).
273

The capacity to regulate one’s own emotions therefore has a clear adaptive
function for social interaction, both for the individual, the group, and the species as a
whole.
274
Indeed emotional regulation is a key component of empathy—just as Hume
claims. As mentioned in the previous section, individuals who can regulate their
emotions are more likely to experience both empathy and feelings of concern for others.
Consequently, those who can appropriately regulate their emotions also tend to act in
morally desirable ways.
275
On the other hand, people who cannot control their emotions
tend to experience their emotions more intensely, especially negative emotions. They are
consequently much more prone to experience personal distress in response to another's
emotional state or condition.
276
Personal distress tends to shift priorities toward the
immediate present and therefore promotes a short-term egocentric focus such as, for
instance, the goal of feeling better and getting relief from a painful situation, which is
almost always in conflict with a motivation to feel for the other. In the case of empathy,
then, the best response to others’ emotional distress will typically not be distress, but

273
See also Eisenberg, N., Smith, C.L., Sadovsky, A., and Spinrad, T.L. (2004).

274
Ibid.

275
Eisenberg, N., Fabes, R.A., Murphy, B., Karbon, M., Maszk, P., Smith, M., O'Boyle, C., and Suh, K.
(1994).

276
Eisenberg, N., Shea, C.L., Carlo, G., and Knight, G. (1991).

307
efforts to relieve their distress, which typically entails exerting executive control over
one’s own emotions.
277

Executive control (i.e., cognitive top-down) mechanisms can either function to
“down-regulate” (i.e., reduce) or “up-regulate” (i.e., promote) emotions.
278
Although
down-regulation seems to be more prevalent in the case of empathy, both processes play
an important role when responding to others in need. In the case of perception of others
in pain, the ability for down-regulation of emotion is prominent if the distress of the
target becomes overwhelming for the observer. For example, a mother alarmed by her
baby’s cries has to cope with her own aversive reaction in order to provide appropriate
care for her distressed offspring.
279
Up-regulation, on the other hand, becomes important
if an observer wants to promote his or her empathic response towards others, as in the
attempt to relate to people to whom we would regularly react with indifference or even
avoidance, such as members of an out-group.
It is noteworthy, however, that not all top-down mechanisms and strategies
contribute to empathic concern. As we have seen, empathy does not come without costs
despite the obvious advantages of understanding another person’s thoughts and
feelings.
280
Some of our top-down mechanisms and strategies can function to mitigate
these associated costs. For instance, in everyday situations, distractive strategies—i.e.

277
Decety and Lamm (2006).

278
Ibid.

279
Interestingly, women but not men, and independent of their parenthood, tend to demonstrate automatic
down-regulation in their emotional response to infant crying and laughing. See Seifritz et al (2003).

280
Hodges, S.D. and Klein, K.J.K. (2001).

308
not paying attention to another’s emotional state—can play an important role if the cost
of empathizing is too high. It is now well established that distraction can effectively
reduce and even eliminate the personal reaction to aversive painful stimulation.
281

Research has shown that selectively focusing one’s attention on specific sensory cues
(e.g., facial expression or language prosody) conveying the emotional state of another can
trigger different emotional responses in the observer.
282
For example, according to one
neuro-imaging study, focusing on the unpleasantness of noxious stimuli results in a
stronger activation of the pain matrix.
283
Thus, by selectively attending to certain aspects
of a painful experience, one can increase (i.e., up-regulate) its aversiveness.
In the case of empathy, the selective direction of attention appears to be one of the
strategies yielding different empathetic emotional responses. A parallel strategy that
contributes to different empathetic responses is to deny the relevance of an emotion-
eliciting stimulus. These and other variations in our empathic responses can be
eliminated, or at least mitigated, by means of top-down control and perspective-taking
instructions that force the participants to focus on particular aspects of the emotional
response of others, or of themselves. For instance, taking a “detached observer position”
by consciously or unconsciously generating an image of the observing self is known to

281
See, for instance, Villemure, C. and Bushnell, M.C. (2002); Wiech, K., Seymour, B., Kalish, R.,
Stephan, K.E., Koltzenburg, M., Driver, J., and Dolan, R.J. (2005); and Kalisch, R., Wiech, K., Critchley,
H.D., and Dolan, R.J. (2006).

282
Williams, M.A., McGlone, F., Abbott, D.F., and Mattingley, J.B. (2005).

283
Kulkarni, B., Bentley, D.E., Elliott, R., Youel, P., Watson, A., Derbyshire, S.W.G., Frackowiak, R.S.J.,
Friston, K.J., and Jones, A.K.P. (2005).

309
reduce the subjective experience of anxiety and pain reactivity.
284
Or to take another
example, adopting an egocentric perspective results in behavioral and neural responses
that are more self-centered, whereas attending to the other in a detached way results in an
amplification of other-oriented emotional responses.
285

Significantly, our ability to adopt the perspective of others enables us to overcome
our egocentrism in part by directing our attention to others. This in turn allows us to
tailor our behavior to others’ expectations, thus making possible satisfying interpersonal
relations.
286
Cognitive neuroscience research demonstrates that when individuals are
asked to adopt the perspective of others, common neural networks are activated for
representations of both the self and the other, including those neural circuits involved in
pain processing of the self and the other.
287
These findings reveal important similarities
in neural networks representing first- and third-person information, and this is consistent
with the hypothesis that such shared representations are largely responsible for social
interaction.
288

There is not a complete merging between representations of self and other
however. There are important differences between the self- and other-perspectives.
Interestingly, unlike taking an egocentric perspective, taking the perspective of an other
results in specific activation of parts of the frontal cortex that are implicated in executive

284
Decety and Lamm (2006).

285
Ibid.

286
Davis, M.H. (1994).

287
Jackson, P.L., Brunet, E., Meltzoff, A.N., and Decety, J. (2006).

288
Decety, J. and Grèzes, J. (2006).

310
control. It has been hypothesized that the role of the frontal lobes could be to hold
separate perspectives, or to resist interference from one’s own perspective.
289
These
important aspects—namely, self-other awareness and self-control—are likely to be play a
role in distinguishing emotional contagion, which heavily relies on the automatic link
between perception of the emotions expressed by the other and one’s own experience of
the same emotion, from empathy, which necessitates a more detached relation. It is also
noteworthy that the egocentric perspective yields higher pain ratings and involves the
pain matrix more extensively. This result suggests that the frontal lobes may also reduce,
or “down-regulate”, the pain matrix when taking the perspective of another.
290

Thus, perspective-taking turns out to be an important part of empathy because of
its contribution to executive (top-down) control and thereby to emotional regulation,
which is a key component of empathy. Of special interest to our project are findings
from social psychology that document a distinction between imagine oneself and imagine
the other perspective-taking.
291
These studies show that when participants are asked to
put themselves into the shoes of individuals experiencing pain, this “imagine self”
perspective tends to induce personal distress. On the other hand, when the participants
were asked to focus their attention on the feelings and reactions of the affected
individuals, this “imagine other” perspective tends to evoke empathic concern. In short,
projecting oneself into an aversive situation leads to higher personal distress and lower

289
Decety, J. and Jackson, P.L. (2004).

290
Ibid.

291
See, for instance, Lamm, C., Batson, C.D., and Decety, J. (2007), and Batson, C.D., Lishner, D.A.,
Carpenter, A., Dulin, L., Harjusola-Webb, S., Stocks, E.L., Gale, S., Hassan, O., and Sampat, B. (2003).

311
empathic concern, while focusing on the emotional and behavioral reactions of another’s
plight is accompanied by higher empathic concern and lower personal distress.
This observation helps to explain why empathy need not yield prosocial behavior.
If one approaches another person’s emotionally or physically painful circumstance using
a first-person perspective, then this will tend to elicit personal distress. Consequently, the
observer will tend not to attend fully to the other's experience and as a result lack
sympathetic (prosocial) behaviors. To avoid this result, the observer needs to carefully
distinguish between the self and other, as well as between one’s own feelings and the
other’s feelings. Only then can the observer successfully take up the perspective of the
other and exert control over one’s own feelings while focusing attention on the other’s
feelings.
292
As Decety and Lamm point out (2006: 1154), this distinction between self
and other is “a crucial aspect that enables a selfless regard for the other rather than a
selfish desire to escape aversive arousal.”
Once again, these results corroborate Hume’s account. Just as Hume claims,
awareness of a distinction between the experiences of the self and others constitutes a
crucial aspect of empathy, or “sympathy”.
293
Rather than taking up an egocentric
perspective, we are to adopt an equanimous and disinterested (i.e., detached) perspective.
Hume suggests that appraisers are to ignore their own desires, interests, and ends, and
calmly focus their attention on the agent and those typically affected by the agent’s
act/trait. In this way, appraisers are much more likely to experience empathetic concern

292
Decety, J. and Jackson, P.L. (2004), and Decety and Lamm (2006).

293
See part one, especially chapters IV and V.

312
and thereby be motivated to engage in prosocial behavior. Hume does claim that we are
to use our imagination to “approach equally near” the agent and sympathize from that
perspective, which in effect amounts to putting ourselves in the agent’s shoes. But we
are not to put ourselves in the shoes of those with whom we are to sympathize. We do
not sympathize with the agent’s pleasure or pain; rather, we are to sympathize with the
pleasure or pain of those typically affected by the agent’s act/trait. According to Hume’s
approach, the point of putting oneself in the agent’s shoes—i.e., “approaching equally
near” the agent—is to take into account the psychological proximity of the agent to those
typically affected by the agent, and thereby regulate sympathy’s natural variability.
In sum, Hume’s MPV offers perspective-taking instructions that regulate the
natural variability of our innate “sympathy” in order to ensure uniformity of feeling and
judgment. This basic approach is congruous with current scientific evidence. As
discussed, empathy partly relies on “bottom-up information processing” such as motor
mimicry and emotional contagion, both of which are automatic and unconscious (i.e.,
innate).
294
But empathy also partly relies on “top-down information processing,” which
exerts executive control using various mechanisms and strategies. We have seen how
various perspective-taking instructions—e.g., imagine other vs. imagined self—can
regulate some of empathy’s natural variability such that empathy produces empathetic
concern and prosocial behavior rather than personal distress and egocentric behavior.
While we have seen how taking a disinterested and equanimous perspective can
elicit empathetic concern and prosocial behavior, we have not confirmed that the

294
It is noteworthy that Hume takes emotional contagion to be an important component of empathy.

313
particular perspective-taking instructions part one has argued is imposed by Hume’s
MPV succeed in mitigating variability so as to attain interpersonal agreement. Given the
similarities between Hume’s account of empathy’s variability and the empirical evidence
presented here, it wouldn’t be surprising if the sort of perspective-taking instructions
proposed by Hume eliminated, or at least successfully mitigated, empathy’s variability.
Nevertheless, this is merely conjecture at this point. More tailored studies need to be
done in order to gauge their success.
It also remains to be seen whether these perspective-taking instructions can
produce the sort of stability and uniformity in moral feeling and judgment that is required
by Hume’s general approach and the proffered moral ontology. Empirical evidence does
suggest that the various mechanisms and strategies involved in empathy are basically the
same across human individuals—barring any psycho-physiological abnormalities of
course. It also therefore suggests that the various perspective-taking strategies deployed
will produce similar results across persons.
295
As we shall see, this in turn raises at least
some expectation of interpersonal agreement in moral judgment.
We will return to the issue of the connection between empathic emotion and
moral judgment in a moment. And we will address the issue of moral disagreement later.
First, however, we have yet to consider the role of reason in moral cognition. (As we
shall see, this role contributes to moral disagreement.) Part one has argued that, in
Hume’s view, our moral judgments typically involve “causal reasoning”, in particular,

295
Interestingly, numerous studies show that those who engage in meditation practice show increased
empathetic concern. And those who have practiced a particular form of meditation for several decades
show unprecedented activity in certain neural networks associated with empathetic concern.

314
reasoning about the causal tendency of objects to fulfill or foil ends or goals. In the
moral case, the relevant ends/goals are those that, when fulfilled or foiled, tend to cause
pain or pleasure either in the possessor of the object or those who typically associate with
her. We have already seen that perception or comprehension of the painful and
pleasurable states of others affect us through empathy. There is also evidence that our
ability to both predict outcomes and evaluate short and long-term goals/ends is an
important factor in moral cognition.
§VIII.2.4 The foundations of moral cognition
Evidence strongly indicates that the neural mechanisms associated with moral
cognition, which includes our moral/empathic emotions, are not restricted to the regions
of the brain associated with more rudimentary emotions.
296
Various studies are
beginning to unravel the connections between the brain regions involved in moral
cognition.
297
Consequently, a new model of the architecture of moral cognitive and
behavioral phenomena is beginning to emerge. According to this model, moral cognitive
and behavioral phenomena arise from the integration of three main components, namely:
structured event knowledge, which occurs in the prefrontal cortex, social perceptual and
functional features, both of which occur in the superior temporal sulcus and anterior
temporal cortex region, and finally central motive and emotional states (e.g., aggression,

296
See Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005) for review.

297
Ibid.

315
anxiety, attachment, hunger), occurring in the limbic/paralimbic regions.
298
Thus, these
components correspond to three different brain regions activated in moral cognition.
See figure (a) below.
299

Figure (a):

Before we discuss each of these three components, it may be helpful to begin by
illustrating this integrative approach with a simple example. Consider, for instance, the
moral/empathic emotion of compassion for an orphan child. The evocation of this moral
emotion requires the integration of specific contextual event representations (e.g., the fact
that the orphan child’s parents died in a car accident and the chances of adoption at her
age are low), context-independent social perceptual features (e.g., a sad facial expression
of the orphan child) and social functional features (e.g., abstract conceptual knowledge
pointing to the features of the orphan’s ‘helplessness’), and finally central motive states

298
This model is proposed and defended in Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and
Grafman, J. (2005).

299
Figure (a) is taken from Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005:
805). (The anatomic image used in this panel was adapted by Moll et al, with permission, from REF. 148
© (1996) Appleton & Lange.)

316
(e.g., one’s own sadness, anxiety and attachment).
300
These components engender a
“gestalt” moral experience (e.g., compassion) by way of temporal synchronization.
301

See figure (b) below.
302

Figure (b):

Significantly, in addition to our moral/empathic emotions such as compassion,
such “gestalt” moral experiences also tend to include other important moral feelings,
namely, feelings of moral approval and disapproval. For example, the experience of
compassion for the orphan will presumably prompt such moral feelings as approval
directed at acts of kindness toward the orphan, and/or disapproval directed at those who
ignore and/or are unconcerned about the orphan’s plight. Because these feelings of moral
approval and disapproval are typically associated with our feelings of moral/empathic

300
Ibid.

301
Singer, W. (2004).

302
Figure (b) is taken from Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005:
805). (The anatomic image used in this panel was also adapted by Moll et al., with permission, from REF.
148 © (1996) Appleton & Lange.)

317
concern (e.g., compassion), they too are to be included as part of our overall gestalt moral
experience, or so I suggest.
Part one has argued that Hume basically holds the same position. Many of
Hume’s claims suggest that he conceives of our empathic feelings as moral feelings—
recall part one’s discussion of the ambiguity in Hume’s use of the term “sympathy” in
which he sometimes uses it to refer to feelings of moral/empathic concern.
303
Hume is
also committed to a distinction between these other-oriented feelings and moral feelings
of approval and disapproval. As part one (§V.5) has argued, our empathic feelings
cannot be identical to our moral feelings of approval and disapproval because, according
to Hume’s account: (i) our empathic feelings evoke those particular moral feelings, and
(ii) the object of our empathic feelings in a particular case can be, and typically is,
distinct from the object of our moral feelings of approval or disapproval. For example, in
the case above, the object of our compassion is the orphan child, whereas the object of
our approval and/or disapproval is particular acts (e.g., acts of kindness toward the
orphan) and/or agents who perform those acts (e.g., someone who blatantly ignores the
orphan’s plight).
Structured event knowledge:
Morality is a real-world business. It is about people navigating, interacting, and
making choices in an ever-changing world. To compensate for this fact, humans
integrate extensive contextual elements when evaluating the behavior of others, and/or
one’s own actions in a given situation. The prefrontal cortex (PFC) is crucially important

303
In the Enquiry, Hume clearly treats “the sentiment of humanity” as a moral feeling, and the sentiment of
humanity is closely associated with sympathy. It is an other-oriented sentiment. (See part one, chapter V.)

318
in structuring the context-dependent social and non-social knowledge into representations
of what J. Grafman calls “event sequence complexes”.
304
These “structured-event-
complex” (SEC) representations are “long-term memories of event sequences that guide
the perception and execution of goal-oriented activities, such as going to a concert or
giving a dinner party”.
305
SEC representations include situational knowledge abstracted
across events (e.g., knowledge gathered from having given several dinner parties) and the
temporal organization of events (e.g., sending out invitations, going grocery shopping,
setting the dinner table, dressing up, etc.). According to Grafman and his colleagues,
activated SECs “sequentially bind” representations of objects, actions, and spatial maps
stored in PFC brain regions.
306

It is hypothesized that executive functions performed by the PFC, such as those
involved in empathy, are based on stored SEC knowledge.
307
This is supported by recent
research. For instance, one recent study investigated the effects of cognitive appraisal on
the experience of empathy.
308
While in a MRI scanner, participants watched video clips
from two groups of neurological patients expressing pain due to a new experimental
treatment using dissonant sounds. Cognitive appraisal was manipulated by providing

304
Grafman, J. (1995); Koechlin, E., Basso, G., Pietrini, P., Panzer, S. & Grafman, J. (1999); Ruchkin,
D.S., Grafman, J., Cameron, K., & Berndt, R.S. (2003); and Wood, J.N. & Grafman, J. (2003).

305
Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005: 803).

306
The SEC framework predicts that different subdivisions of the PFC store different types of event
sequence knowledge. For example, novel or multi-tasking events, routine tasks, multi-stage events and
long-term goals, and social and emotional events are stored in different areas of the PFC. See, for instance,
Wood, J.N. & Grafman, J. (2003); Wood, J. N., Romero, S. G., Makale, M. & Grafman, J. (2003); and
Wood, J. N., Romero, S. G., Knutson, K. M. & Grafman, J. (2005).

307
Ibid.

308
Lamm, C., Batson, C. D., and Decety, J. (2006).

319
information about the outcome of the treatment: while health and quality of life improved
in one group, members of the other group did not benefit from the treatment. As
anticipated, witnessing another person suffering and knowing that his or her treatment
had not been effective increased the negative emotional response in the observer.
Conversely, knowing that a treatment had been beneficial for the patient decreased (i.e.,
down-regulated) the negative response in the observer. In short, our evaluation of ends,
goals, and outcomes has a direct impact on our empathic emotional responses—just as
Hume claims.
Social perceptual and functional features:
The human brain deals with oodles of perceptual signs of social significance. Its
ability to manage this burden of information relies on complex patterns of aspectual and
semantic knowledge.
309
Making implicit or explicit moral evaluations in a given social
situation requires the ability to efficiently extract social perceptual and functional aspects
from the environment. Social perceptual features are extracted from facial expression,
gaze, prosody, body postures, and gestures.
310
Social “functional” features code for
context-independent semantic properties (e.g., naming particular human actions and
features such as the orphan’s ‘helplessness’) that are extracted from different social
situations.
311
There is substantial evidence that semantic impairment and dementia are

309
McClelland, J. L. & Rogers, T. T. (2003); Martin, A. & Chao, L. L. (2001); Caramazza, A. & Mahon, B.
Z. (2003).

310
Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005).

311
Ibid.

320
associated with severe behavioral changes.
312
For example, psychopathic individuals
exhibit semantic impairment.
313

Central motive states:
Moral cognition would be stripped of its practicality without its
motivational/emotional aspects. The postulation of “central motive states” has served as
an influential account of the basic mechanisms of motivation.
314
Central motive states
are conceived as “undirected” emotional states such as sexual arousal, social attachment,
hunger, aggression, and happiness. These states are to be distinguished from “basic”
emotions (e.g., fear and disgust) and moral emotions (e.g., compassion, guilt, contempt,
and approval).
315
“Basic” emotions emerge from the integration of perceptual and
contextual representations (e.g., perceiving the feared object or situation) with central
motives states (e.g., undirected anxiety). Moral emotions share this integrative feature
with “basic” emotions but, unlike basic emotions, moral emotions are linked with the
interest or welfare of other individuals.
316
Moral emotions spring from empathy. They
embody empathic concern.

312
See, for instance, Mendez, M. F., Chow, T., Ringman, J., Twitchell, G. & Hinkin, C. H. (2000); Bozeat,
S., Gregory, C. A., Ralph, M. A. & Hodges, J. R. (2000).

313
Kiehl, K. A. et al. (2001); Kiehl, K. A. et al. (2004).

314
Wood, J. N., Knutson, K. M. & Grafman, J. (2004).

315
Ibid.

316
Haidt, J. (2003).

321
§VIII.2.5 Event-feature-emotion complex framework
The proposed “event-feature-emotion complex” model contrasts with the
commonly held view that “rational” cognitive mechanisms compete with emotional ones.
According to this integrative framework, central motive states play a key role in moral
cognition and behavior due to the integration of emotional with “rational” components.
317

For example, perceptual and contextual representations allow you to notice that someone
is hurt, whereas central motive states elicit anxiety and attachment. Out of the integration
of these components emerges a moral/empathic emotion. Of course, in some cases one
may not feel moral/empathic concern; e.g., one might instead feel personal distress due to
empathic over-arousal. Hence various perspective-taking instructions may need to be
followed in order to engender a moral/empathic emotion in a given social context. This
moral/empathic emotional state motivates you to help the suffering person. As we shall
discuss momentarily, moral emotional states also directly influence implicit and explicit
moral evaluations.
Another key aspect of moral cognition, on the integrative model, is the
representation of short- and long-term goals and the prediction of the utility of outcomes
in social settings.
318
Pursuing goals and/or foreseeing possible consequences of one’s
decisions within a particular social context requires the ability to estimate the likelihood
of outcomes and the desirability of those outcomes given certain ends. Functional

317
This integration is a byproduct of the integration of brain function occurring in the cordical and limbic
regions.

318
Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., and Grafman, J. (2005). See also Schultz, W.
(2002).

322
integration of different regions of the brain (e.g., the anterior PFC, which represents long-
term goals and outcomes, and limbic structures, which code for the reward-value of
behavioral choices) is key to our ability to weigh the motivational and moral relevance of
different behavioral choices in social situations.
319
Thus, according to this integrative
approach, our ability to represent and evaluate short- and long-term goals, ends, and
outcomes is intimately linked to motivational salience
320
—just as Hume claims.
The proposed event-feature-emotion complex model also predicts various
cognitive and behavioral abnormalities within individuals who suffer from defects (e.g.,
lesions) in the three brain regions that correspond with the three main cognitive
components that give rise to moral cognitive and behavioral phenomena. Lesions in any
one of these three brain regions tend to have specific and predictable negative effects on
one’s ability to empathize and experience empathic/moral concern, along with moral
feelings of approval and disapproval. As we shall discuss in the following section, these
negative effects also, consequently, have adverse affects on both behavior and judgment.
§VIII.3 A Humean moral ontology: Whither now?
Current neurological and social psychological evidence bolsters many of Hume’s
claims regarding sympathy and its role in explaining moral cognitive and behavioral
phenomena. It even supports Hume’s contention that our moral/empathic emotions are
directly responsive to our judgments about the tendency of objects (e.g., acts or traits) to

319
Ibid.

320
This account is supported by the aforementioned study [Lamm, C., Batson, C. D., and Decety, J.
(2006)], which investigated the effects of cognitive appraisal on the experience of empathy.

323
fulfill or foil particular ends or goals.
321
According to current models of moral cognition,
our reasoning about short- and long-term ends, goals, and outcomes is an integral part of
our moral/empathic emotions. However, neurological evidence suggests that we are not
always conscious of our engagement in such cognitive evaluations.
322
Consequently, our
moral judgments often appear to be more “intuitive” than “rational”.
323

Anecdotal evidence for this later claim comes from Haidt, Bjorklund, and Murphy
(2000), who offers the following scenario:
Julie and Mark are brother and sister. They are traveling together in
France on summer vacation from college. One night they are staying
alone in a cabin near the beach. They decided that it would be interesting
and fun if they tried making love. At the very least it would be a new
experience for each of them. Julie was already taking birth control pills,
but Mark uses a condom too, just to be safe. They both enjoy making love
but decide not to do it again. They keep that night as a special secret
between them, which makes them feel even closer to each other. What do
you think about that, was it OK for them to make love?

Later, Haidt (2001: 814) describes people’s responses to this story as follows:
Most people who hear the above story immediately say that it was wrong
for the siblings to make love, and they then set about searching for
reasons. They point out the dangers of inbreeding, only to remember that
Julie and Mark used two forms of birth control. They next try to argue
that Julie and Mark could be hurt, even though the story makes it clear that
no harm befell them.
324
Eventually many people say something like “I
don’t know, I can’t explain it, I just know it’s wrong.”

321
See part one, chapter V.

322
It is unclear whether this view is incompatible with Hume’s account. This is due to the fact that it is
unclear (at least to me) whether, on Hume’s account, one must be conscious of the mind’s use of various
“ideas”—e.g., ideas of goals and actions, and whether certain actions will fulfill or foil particular short- and
long-term goals.

323
Haidt, J. (2001: 814ff). See also the discussion below.

324
One might take issue with this claim, but it does nevertheless seem to be the case that subjects’
intuitions about the wrongness of Julie and Mark’s actions are more immediate and confident than

324

Consider two other examples taken from Judith Jarvis Thomson’s classic paper,
“Trolley Problem” (1986) in which she presents a series of moral dilemmas involving
trolleys. In one dilemma, a runaway trolley is headed for five people who will be killed
if it proceeds on its present course. The only way to save these people is to hit a switch
that turns the trolley onto an alternate set of tracks where it will run over and kill one
person instead of the five. Thomson then queries her audience: Is it morally acceptable to
hit the switch and turn the trolley in order to save five people at the expense of the one?
Most people say immediately that it is morally acceptable.
In another dilemma, a runaway trolley threatens to kill five people as before, but
this time there is no switch to hit. Instead, the agent is standing next to a large stranger
on a footbridge running over the tracks, in between the oncoming trolley and the five
people. In this scenario, the only way to save the five people is to push this stranger off
the bridge and onto the tracks below. The stranger will die as a result of doing this, but
his body will stop the trolley from reaching the others. Thomson then asks whether or
not it is morally acceptable to save the five people by pushing the stranger to his death.
Most people say immediately that it is not morally acceptable.
Thomson considers these dilemmas and the intuitions they elicit as part of a
puzzle for moral theorists: What makes it morally acceptable to sacrifice one life to save
five in the original case but not in the footbridge case? Answering this question has
proven to be surprisingly difficult. Many plausible answers have been proposed, but they

whatever judgments subjects tend to make about the possibility of harm. This is clearly demonstrated by
the fact that many of the respondents admitted that they simply couldn’t explain why the act was wrong.

325
all arguably run into trouble at some point. To illustrate this, consider the following
example taken from Thomson’s discussion. Immanuel Kant famously holds that one
ought never treat someone as a means only. With this in mind, a Kantian might argue
that what distinguishes these two cases is that in the footbridge case one literally uses
someone to save the others, whereas in the original case one doesn’t use the other person;
he just happens to be in the way. Consider, then, a case just like the original one only this
time the track with the one person on it loops around so that without that person there to
block the trolley it would come back around and kill the five people you were trying to
save. In this variation of the original case, you must use the other person to stop the train.
Nevertheless, it seems to most people that it’s still okay to hit the switch and turn the
trolley, thus demonstrating the inadequacy of this (admittedly crude) Kantian principle.
While more sophisticated principles have been proposed, as far as I know, no one
has found a set of principles that adequately captures most people’s intuitions concerning
these and other cases. Thus, if the trolley problem has a solution, then its solution is
certainly not obvious. The fact that it has no obvious solution is interesting from a
psychological point of view. This is because nearly everyone judges immediately that
it’s acceptable to act in the original case and not in the footbridge case in spite of the fact
that it’s very hard to say why this is so.
325
Even if at some point we uncover a more
sophisticated set of principles that could plausibly serve as justificatory reasons for our

325
There is an interesting parallel here with Chomsky’s observation that individuals, even young children,
can adeptly apply the rules of grammar in the absence explicit knowledge of those rules. (See Pinker also
(1994).) Hauser, Chomsky, and Fitch (2002) argue that the rules of grammar are cognitively innate,
universally “hard-wired” in the human brain, and that this must have developed as an adaptation of some
neural system possessed by our distant ancestors, perhaps one used in navigation. Similarly, M. Hauser
(2007) argues for an innate “moral grammar”, universally hard-wired into the human brain.

326
moral intuitions in these cases, they would be ex post facto reasons since most people’s
immediate moral intuitions in the two cases are clearly not guided by the application of
these principles (whatever they may be).
How then do people consistently manage to immediately give the same answers
when there is no obvious line of reasoning that would lead one to those answers?
Neurological evidence is beginning to unravel this somewhat peculiar phenomenon of
moral cognition, namely, the phenomenon of moral “intuition”. The answer appears to
be that we are quite often not consciously aware of the cognitive processes that give rise
to our moral judgments and moral emotions. Joshua Greene et al. (2001) have conducted
brain-imaging research that supports this conclusion. Their research considers people’s
intuitive moral appraisals of the two hypothetical trolley dilemmas (above) among others.
Greene and his colleagues suggest that pushing someone to his death with one’s
bare hands is more “personal” than hitting a switch that will have similar consequences.
They hypothesized that the thought of harming someone in this “personal” way is more
emotionally salient than the thought of harming someone in a relatively “impersonal”
way. Because people have a robust, negative empathic emotional response to the
“personal” violation proposed in the footbridge case, people immediately say that it’s
wrong. At the same time, people fail to have a strong negative empathic emotional
response to the relatively “impersonal” violation proposed in the original trolley case.
The researchers speculate that people therefore revert to the most obvious moral
principle, “minimize harm,” which in turn leads people to say that the action in the
original case is morally permissible.

327
Green and his colleagues predicted that subjects passing judgment on dilemmas
like the footbridge case while undergoing brain scanning would therefore show increased
activity in at least some emotion-related brain areas as compared to those passing
judgment on dilemmas like the original trolley case. They also made the following
prediction concerning an interaction between people’s responses and reaction times.
According to the above hypothesis, when people contemplate the performance of a
“personal” violation such as pushing someone in front of a trolley, they experience an
empathic emotional response that inclines them to judge that action inappropriate. Some
people, however, do judge such actions to be appropriate. Given that none of the subjects
in the study are psychopaths, the researchers hypothesized that when subjects give such
answers they do so “in spite of” their countervailing empathic emotions. Due to the
“emotional interference” involved, Greene and his colleagues predicted that such
individuals would, on average, take longer to give their judgments than those who judge
such actions to be inappropriate, and longer than those passing judgments on dilemmas
that do not elicit strong countervailing emotional responses.
Both of their predictions proved accurate. Three brain regions associated with
emotional response—viz., the medial prefrontal cortex, the posterior cingulate region,
and Brodmann’s Area 39—were more active during contemplation of the “personal”
footbridge dilemma than during contemplation of the “impersonal” original trolley
dilemma. The researches also found that reaction times for trials in which the subject
judged a “personal” violation to be appropriate were, on average, significantly longer
than those trials involving “emotional interference”. This suggests that further cognitive

328
processes were engaged in these subjects, resulting in moral judgments that diverge from
their empathic emotional reaction. This, of course, does not entail that these subjects
would act in accordance with their judgments, especially given that their felt empathic
emotions are motivational states.
326

While it appears to be the case that, in general, our empathic emotions drive our
moral judgments, we can and sometimes do issue moral judgments that diverge from
these felt empathic emotions. In the latter type of case, further cognitive processes are at
work. In either case, whether our moral judgments are more driven by our empathic
emotional responses or by other cognitive processes, we often appear to be generally
unaware of the cognitive goings-on that shape our moral judgments. This suggests that
our moral judgments are often not a product of complex conscious moral reasoning.
Rather, our moral reasons tend to be ex post facto justifications of our moral “intuitions”.
This, however, is not to say that conscious moral reasoning and the application of
sophisticated moral principles does not hold sway in some of our moral judgments.
Clearly it sometimes does. But the evidence we have reviewed thus far suggests that the
sorts of considerations taken into account by our conscious moral deliberations may not
be much, if any, different from the sorts of considerations unconsciously taken into

326
Anecdotally, in my own presentations of the two trolley dilemmas to friends and acquaintances, a
surprisingly large number of people said it would be morally appropriate to push the stranger off the
footbridge. Upon questioning, every one of these persons stated that they did not think they would actually
push the stranger however. Interestingly, in offering reasons for their moral judgment, all of these persons
gave reasons that appealed to consistency with their judgment in the first dilemma. Apparently, people
overrode their “intuitive” moral judgment out of concern of appearing to be an inconsistent judge. To test
this hypothesis, I began presenting the footbridge trolley dilemma first and the original trolley dilemma
second. I have not yet had one single person to judge that pushing the stranger off the bridge was morally
appropriate when presenting the dilemmas in this order. People don’t seem to mind the inconsistency when
moving from the second to the first dilemma. Green et al. does not address this phenomenon in their study.

329
account within the cognitive processes that shape our “intuitive” moral judgments. As
discussed, our empathic emotions integrate various contextual cues taken from the social
world we inhabit. Barring various cognitive defects, we unconsciously take into account
various short- and long-term ends, goals, and outcomes, including the present and/or
future emotional states of others, by way of social perceptual features (of face, gaze,
voice, body posture, etc.), social functional features (of social behaviors), and structured
event knowledge (i.e., knowledge of events that guide our perception and execution of
goal-oriented activities, including situational knowledge and knowledge of the temporal
organization of events). Given the cognitive similarities across subjects, this would
explain the commonality among most people’s answers to the trolley dilemmas
regardless of whether those answers were more “intuitive” than consciously deliberative.
For instance, it would explain why people, whether consciously or unconsciously, still
appear to revert to a “minimize harm” principle in the original trolley dilemma.
Unfortunately, none of the studies we have discussed thus far provide any direct
evidence that supports Hume’s moral ontology—i.e., the view that moral properties are
intersubjective, dispositional, emotion-dependent properties. Of special interest to our
project are studies concerning the effects that brain damage can have on our empathic
emotions. This is of special interest because these impairments demonstrate the effects
that our empathic emotions have on our moral appraisals. Current neurological
understanding of the relationship between our emotions and our moral judgments does in
fact provide some plausibility to Hume’s moral ontology, or at least to a moral ontology
quite similar to Hume’s.

330
The study of psychopaths proves to be quite illuminating. According to Blair and
colleagues, “The clinical and empirical picture of a psychopathic individual is of
someone who has some form of emotional deficit.”
327
Psychopaths exhibit a lower level
of electrodermal activity, which is a standard indication of emotional arousal. For
instance, psychopaths show weaker electrodermal responses to emotionally significant
stimuli than normal individuals.
328
In Blair et al.’s study, electrodermal responses of
psychopaths were compared to a control group of criminals who, like the psychopaths,
were serving life sentences for murder or manslaughter.
329
While the psychopaths were
like the other criminals in their responses to threatening stimuli (e.g. an image of a
shark’s open mouth) and neutral stimuli (e.g. an image of a book), they showed
significantly reduced responses to distress cues (e.g. an image of a crying child’s face)
relative to the control group of criminals. This finding is consistent with the observation
that psychopathic individuals appear to have a severely diminished capacity for empathic
emotions.
Blair (1995) predicted that psychopaths’ diminished capacity for empathy would
prevent them from drawing a distinction between “moral” and “conventional”
transgressions. “Moral” transgressions were defined as those having negative
consequences for the “rights and welfare of others” and included instances of one child
hitting another. “Conventional” transgressions were defined as “violations of the

327
Blair, R. et al. (1997).

328
Hare, R.D. & Quinn, M.J. (1971).

329
Blair, R. et al. (1997).

331
behavioral uniformities that structure social interactions within social systems” and
included instances of a boy wearing a skirt and a child who leaves the classroom without
permission.
330
Blair also predicted that psychopaths, as compared to other criminals,
would make fewer references to the pain or discomfort of victims in explaining why
certain harmful actions are unacceptable.
Both of these predictions were confirmed. While the “normal” subjects (i.e., the
non-psychopathic criminals) drew a general distinction between moral and conventional
transgressions, the psychopaths did not. Normal subjects also found a greater difference
in the permissibility and seriousness between moral and conventional transgressions than
did the psychopaths. In addition, the psychopathic subjects were, as predicted, less likely
to appeal to the pain and discomfort of victims and more likely to appeal to the violation
of rules in explaining why various transgressions are unacceptable.
The most intriguing finding pertained to the psychopath’s judgments concerning
“modifiability.” Each of the transgression stories were set in school, and in each case the
subjects were asked whether or not it would be permissible for the child to perform the
transgression if the teacher had said earlier that such actions were permitted. Non-
psychopathic criminals tended to say that the conventional transgressions would become
permissible if the teacher were to explicitly allow their performance and that the moral
transgressions would not be permissible in either case. Curiously, the psychopaths
treated all transgressions as impermissible regardless of what the teacher would say.
Blair predicted that the psychopaths would treat both types of transgression as normal

330
Blair, R. (1995).

332
subjects treat conventional ones. Instead, he found that the psychopaths treated both
types of transgressions as normal subjects treat moral ones, namely, as impermissible.
Blair attributes this result to the fact that these incarcerated psychopathic subjects have a
prudential interest in demonstrating to their captors that they have “learned the rules.”
In attempting to assess the role of emotion—in particular, empathic emotion—in
the production of our moral judgments and behavior, psychopaths provide crucial
information. Psychopaths lack empathic emotional capacities, and they also appear to
lack the ability to accurately make, or even fully comprehend, moral distinctions. Our
moral/empathic emotions not only affect moral motivation; these emotions also appear to
affect our capacity to make moral distinctions, thereby affecting our moral judgments.
This result is compatible with Hume’s moral ontology; indeed it is predicted by it.
Assuming that moral properties are dispositional, emotion-dependent properties, we
would expect those who lack the relevant emotional capacities to be unable to apprehend
moral distinctions, thereby affecting their moral judgments. And we would also therefore
expect them to exhibit morally abysmal behavior, especially given the connection
between empathic emotions, which psychopaths lack, and moral motivation.
Psychopaths, who lack the capacity for emotional empathy, meet these expectations.
Moral motivation and behavior, as well as the capacity to apprehend moral distinctions,
appear to be products of our empathic emotional responses.
Furthermore, assuming that moral properties are dispositional, emotion-dependent
properties as Hume contends, there would be no saying what those properties were
without ultimately appealing to the particular emotions that determine them. (See part

333
one.) It would, consequently, be extremely difficult for those who lack the requisite
emotions to fully comprehend the properties. This too is consistent with observations of
psychopaths. It is, I think, noteworthy that the psychopathic subjects did not have a
negative phenomenal (i.e., emotional) experience in reaction to the transgression of moral
rules. I suggest that, as a consequence, they view moral rules as conventional rules, even
if socially inviolable conventional rules. (See the discussion below.)
In fact, I suspect that they take the ‘impermissibility’ of the transgressions as itself
a matter of social convention, for such transgressions never phenomenally (i.e.,
emotionally) strike them as impermissible. Similarly, the visually blind, lacking the sort
of phenomenal visual experience normally associated with colors, could conceivably take
the colors of objects (e.g., a strawberry’s being “red”) to be merely a matter of social
convention. While most blind persons do know that there are facts about the colors of
things, they are, in effect, simply following rules (e.g., rules like: strawberries are “red”,
lemons are “yellow”, oranges are “orange”, grass is “green”, the sky is “blue”, etc.).
As discussed in part one, Hume’s moral ontology does not entail that one must
occurrently be in a particular (i.e., moral) emotional state when issuing a moral judgment
(e.g., X is wrong)—any more than one must occurrently be in a particular visual state
when issuing a color judgment (e.g., X is red). Hume recognizes that we sometimes fail
to meet the requisite conditions for having the sort of phenomenal experience (viz., a
particular kind of “impression”) that determines the relevant property. We can

334
nonetheless issue a correct moral judgment. We in effect judge what we would feel were
we to meet the requisite conditions.
331

As we’ve seen, it can be quite difficult for those who have always lacked the
capacity to have the requisite sort of phenomenal experience, and thus have never met the
requisite conditions, to judge what they would feel were they to satisfy those conditions.
While they can still issue a judgment, to confirm that judgment they would (apart from
extraordinary circumstances) have to defer to the judgments of those who (at some point)
satisfied the requisite conditions and had the requisite sort of phenomenal experience.
For example, a person blind from birth can judge truly that the object on the table is red,
but to confirm that judgment she would (apart from extraordinary conditions) have to
defer to the judgment of a normally sighted person who (at some point) viewed the object
under standard lighting conditions.
The fact that those who have always lacked the capacity to have the requisite sort
of phenomenal experience need to appeal to the judgments of others in order to confirm
the facts (e.g., that x is wrong, or that x is red) may help to explain why moral rules
appear as mere social conventions to psychopaths. Other people are continually telling
them what and what not to pursue, what they ought and ought not to do, what is and is
not permissible, etc., without any independent access to, or understanding of, the “action-
guidingness” of the facts or properties that others apprehend and to which those others
appeal. To the pathological, moral distinctions appear to be nothing more than an
arbitrary set of rules, imposed by society, that tell people how to behave.

331
See part one, particularly chapter II, for further discussion of Hume’s account.

335
On the other hand, those who have at some point in the past had the requisite sort
of phenomenal experience under the requisite conditions may have an easier time judging
what they would feel were they presently able to satisfy those conditions. And in many
cases, they would not have to defer to the judgments of others. For example, an adult
who had just recently lost her sight would presumably correctly judge the color of many
of her possessions without needing to appeal to others’ judgments. Memories of one’s
past phenomenal experiences play a vital role in one’s ability to correctly issue such
judgments. Interestingly, similar findings have been observed in moral cases as well.
For instance, Antonio Damasio reports on a patient named “Elliot” who, as an
adult, had acquired lesions in the ventromedial portion of his frontal lobes due to a brain
tumor.
332
Despite these lesions, Elliot maintained his ability to speak and reason about
topics such as politics and economics. He scored above average on standard intelligence
tests, including some designed to detect frontal lobe damage, and responded normally to
standard tests of personality. However, his behavior was not unaffected by his condition.
Elliot exhibited poor prudential, social, and moral decision-making, as well as anti-social
behavior.
A simple laboratory probe helped reveal the subtle but dramatic nature of Elliot’s
cognitive deficits. Damasio and his colleagues found that Elliot, as well as four other
patients with similar brain damage and similar problems with social behavior,
demonstrated a consistent failure to exhibit typical electrodermal responses when
presented with socially significant stimuli (e.g., pictures of gory accidents of people

332
Damasio, A. (1994: 34-51).

336
about to drown in floods), though these patients responded normally to non-social,
emotionally arousing stimuli.
333
Elliot confirmed this result, reporting that he felt no
emotional response to the socially significant stimuli.
334
Significantly, however, he did
comment that he remembered having strong emotional responses to such stimuli prior to
the growth of the brain tumor.
Intrigued by Elliot’s comments, Damasio and his colleagues employed a series of
tests designed to assess the effects of Elliot’s damage on his decision-making skills.
They asked him, for instance, whether or not he would steal if he needed money and to
explain why or why not. Interestingly, Elliot’s answers were like those of other people,
citing the usual reasons for why one shouldn’t commit such crimes. Damasio and Saver
followed up these tests with a series of five tests of moral/social judgment.
335
As before,
Elliot performed normally or above average in each case. It became abundantly clear that
Elliot was quite different from psychopaths. His explicit knowledge of social and moral
rules was as good or better than most people’s. Yet his personal life had nonetheless
deteriorated rapidly as a result of his condition (although he did not seem to mind).
Damasio attributes Elliot’s real-life failures not to his inability to reason properly
about the situation but to his inability to integrate emotional responses into his practical
judgments. “To know, but not to feel,” says Damasio, is the essence of Elliot’s

333
Damasio, A., Tranel, D., and Damasio, H. (1990).

334
While these findings are similar to those pertaining to psychopaths, and Elliot clearly did exhibit some
“sociopathic” (i.e., severe anti-social) behavior as a result of his injuries, he is not deemed a psychopath.

335
Saver, J.L. & Damasio, A. (1991).

337
predicament.
336
Once again, these findings are consistent with the account of Hume’s
view presented in part one. That account predicts that those who have at some point in
their past had the requisite sort of emotional experience under the requisite conditions
will tend to have an easier time accurately drawing moral distinctions—understanding
what they would feel were they to satisfy certain conditions—than those who have never
had the requisite sort of experiences. It also predicts that in many cases the former group,
unlike the latter, would not have to defer to the judgments of others.
The study of Elliot and others confirm these predictions. In those instances in
which one does not or cannot meet the requisite conditions, memories of one’s past
phenomenal moral experiences appear to play a vital role in one’s ability to correctly
issue moral judgments and make moral distinctions. However, as predicted, Elliot’s
social and moral knowledge, which was acquired earlier in life, does not have an impact
on his behavior. Elliot is not moved by the moral distinctions he draws. In this respect,
he fails to apprehend moral distinctions—just as the person blinded later in life can
accurately draw the color distinctions among her possessions even though she fails to
apprehend their colors.
Contrast this outcome with those who have never had the requisite sort of
phenomenal experience. As predicted, such persons do not appear to be able to
comprehend moral properties/distinctions. Hence they lack social and moral knowledge.
This result has been independently confirmed by a more recent study that considers two
adult subjects who exhibited damage to their ventromedial cortices (similar to Elliot) but

336
Damasio, A. (1994: 45).

338
whose damage occurred at an early age (three months and fifteen months).
337
These two
subjects did exhibit characteristically psychopathic behavior—e.g., lying, stealing,
violence, and lack of remorse after committing such violations—presumably because
they, unlike subjects such as Elliot who had late-onset damage, did not have the
advantage of a lifetime of conditioning via normal emotional responses.
Significantly, the early-onset patients, unlike the late-onset patients, showed
serious deficits in their explicit knowledge of social and moral norms. Anderson and
colleagues suggest that this deficit is due to the fact that the ability to experience feelings
necessary for the normal acquisition of explicit knowledge of moral and social norms was
lost at such an early age, an age well before adequate maturation of many of the regions
of the PFC associated with moral cognition and hence an age well before the ability to
acquire moral and social knowledge.
In sum, various studies show that individuals who have lost the capacity for
empathic emotional responses in adulthood can nonetheless continue to accurately make
moral distinctions in most cases and offer compelling moral reasons in support of those
distinctions, even though they are not moved by those distinctions and reasons. On the
other hand, those who have never acquired the capacity for empathic emotional responses
cannot accurately make moral distinctions. Hence they are not able to acquire moral
knowledge, nor do moral distinctions move them. Anderson et al.’s psychopaths, Blair’s
psychopaths, and Damasio et al.’s prefrontal patients such as Elliot, vividly illustrate that
moral behavior (and social behavior in general) depends crucially on particular emotional

337
Anderson, S. W., A. Bechara, H. Damasio, D. Tranel, and A. R. Damasio (1999).

339
capacities. They therefore lend equal support to Hume’s positive claim that the edifice of
human morality rests on a foundation of “sentiment”, in particular sentiments that are the
product of empathy, or “sympathy”. They also echo Hume’s infamous claim in the
Treatise (Book II, Part III, Section iii) that, “‘Tis not contrary to reason [alone] to prefer
the destruction of the whole world to the scratching of my finger.”
Taken together, all of the above findings appear to lend at least some tentative
scientific support to the basic framework of Hume’s moral ontology, to wit: the
contention that moral properties are intersubjective, dispositional, emotion-dependent
properties. Hume’s moral ontology offers a plausible explanation of those findings. The
moral domain appears to be picked out by psycho-physiological processes, and those
same processes arguably determine the moral domain as well. The following chapters
attempt to further bolster this moral ontology with philosophical support by sketching out
how this moral ontology can also resolve the accommodation problem.

340

Chapter IX: Semantic Accommodation

§IX.1 Introduction

Chapter VIII has, in effect, argued for a contemporary reconstruction of Hume’s
moral ontology based upon evidence gathered in primatology, social psychology, and the
blossoming field of cognitive neuroscience. According to the proffered Humean moral
ontology, the following general characterization of the moral domain is true: an object X
belongs to the moral domain if and only if observation or contemplation of the emotional
effects of X (in a given situation) elicits particular feelings—viz., feelings of empathic
concern and feelings of approbation or (as the case may be) disapprobation toward X—
within those who satisfy certain conditions—conditions such as being psycho-
physiologically “normal” (e.g., lacking brain lesions to particular regions of the brain),
observing or contemplating X from an equanimous frame of mind (e.g., not angry,
depressed, or over-aroused), implementing various perspective-taking techniques, etc.
Moral properties, so conceived, are emotion-dependent properties ascribed to
objects. As Hume puts it, moral properties are “a relation, which nature has placed
between the form [of an object] and the sentiment.”
338
In short, particular emotional
states rooted in empathy determine the domain of moral properties. Just as Hume claims,
“our own sensations determine the vice and virtue of any quality, as well as those
sensations, which it may excite in others” (T 597). Following Hume, this project holds
that our moral judgments are descriptive, property-ascribing, truth-evaluable statements,

338
Of the Standard of Taste [henceforth SOT], paragraph 10.

341
some of which are true. There are moral properties and facts, and moral error and
disagreement are not only possible but do occur. As discussed, these results are
congruent with the first 5 presuppositions of moral practice and discourse.
Hume’s moral ontology is not without its complications however. One obstacle is
that Hume does not explain how any specific moral evaluative term or expression—e.g.,
‘...is morally good’ or ‘...is a virtue’—makes reference to such emotion-dependent moral
properties. I suggest that this is due to the fact that Hume, influenced by many of his
predecessors and contemporaries, was primarily concerned with giving an account of
moral psychology and moral experience itself rather than a precise account of moral
language. Evidence that Hume was neither concerned with the literal semantic content of
such terms as ‘virtue’ and ‘vice’, nor with precisely how people use such terms, can be
found in Hume’s work. For example, Hume writes:
It is fortunate, amidst all this seeming perplexity, that the question
[concerning how one should define or use ‘virtue’ and ‘vice’], being
merely verbal, cannot possibly be of any importance. A moral,
philosophical discourse need not enter into all these caprices of language,
which are so variable in different dialects, and in different ages of the
same dialect.
339

Hume is clearly much more concerned with describing the role that the sentiments
play within our moral practice than describing the precise connection those sentiments
have with our moral evaluative language. For instance, he goes on to state that “if, in
short, the sentiments are similar which arise from these endowments and the social

339
An Enquiry Concerning the Principles of Morals (henceforth EPM), appendix IV, paragraph 2.

342
virtues; is there any reason for being so extremely scrupulous about a word, or disputing
whether they be entitled to the denomination of virtues?”
340

Instead of reading Hume as offering explicit analysis of the literal meanings of
moral evaluative terms and expressions, I suggest he is more plausibly read as providing
a set of general application conditions for moral ascriptions. In other words, Hume only
provides conditions, in particular psychological conditions, for identifying (1) when an
object (e.g., an act or trait) belongs to the moral domain, and consequently, (2) when it is
generally apposite to use certain types of moral evaluative expressions (e.g., ‘...is a
virtue’, ‘...is good’, ‘...is right’, or ‘...morally ought to...’). Hume does not attempt to
give the literal semantic content of these expressions however. For instance, he neither
tells us whether there is any difference in meaning between these expressions, nor does
he fully specify their truth conditions.
In the end, then, Hume leaves us with the task of identifying the best analysis of
moral evaluative terms and expressions that is consistent with his moral ontology (as
interpreted by this project). This, in a nutshell, is the aim of this chapter. This task is
clearly reconstructive because, as I read Hume, he does not offer any semantic analysis.
Moreover, we will be appealing to contemporary standards to assess the analysis.
As far as I am aware, all of those who have offered semantic analysis of moral
evaluative terms and expressions consistent with (what is arguably) Hume’s feeling-
dependent conception of moral properties have taken what I call the traditional
approach. The traditional approach assumes that evaluative terms such as ‘good’ refer to

340
EPM, appendix IV, paragraph 6, Hume’s emphasis.

343
or describe the relevant moral property.
341
Given Hume’s moral ontology, this approach
holds that value terms such as ‘good’ refer to or describe a feeling-dependent property.
For instance, C. D. Broad writes: “Now for Hume the statement ‘x is good’ means the
same as the statement ‘x is such that the contemplation of it would call forth an emotion
of approval towards it in all or most men’.”
342
David Wiggins claims that, according a
sensible Humean account, “x is good if and only if x is the sort of thing that calls forth or
makes appropriate a certain sentiment of approbation.”
343
Wiggins generalizes this claim
stating: “...for each value predicate ϕ, there is an attitude or response of subjects... such
that an object has the property ϕ stands for if and only if the object is fitted by its
characteristics to bring down that extant attitude or response upon it and bring it down
precisely because it has those characteristics” (ibid).
This chapter sketches out an alternative semantic analysis of moral evaluative
expressions that rejects the traditional approach and yet is generally compatible with, and
in fact remarkably complementary to, Hume’s moral ontology. This novel approach is
preferable to the traditional approach for various reasons, a few of which will be
addressed in this chapter. We shall proceed by first focusing on an analysis of ‘good’ and
then expand the analysis in order to provide a more general account of evaluative terms.

341
Within the analytic tradition, this approach has its roots in G. E. Moore’s seminal work Principia Ethica
(1903). As we shall discuss later in this chapter, Moore’s famous “open question” argument simply
assumes that ‘good’ refers to a moral property. Moore, along with most others in roughly the first half of
the 20
th
century, believed his argument showed that ‘good’ resisted any “definition” or analytic reduction.
Moore took his argument to demonstrate that ‘good’ is “indefinable” and must therefore refer to a “simple”,
unanalyzable property.

342
Broad (1944: 84-5, his emphasis).

343
Wiggins (1998: 206).

344
§IX.2 An attributive analysis of ‘good’
Approximately half a century ago, Peter Geach forcefully argued that ‘good’ is
best analyzed as an attributive adjective rather than a predicative adjective.
344

Attributives are not independently predicative because they do not denote a property.
Instead they modify (explicitly or implicitly) a predicative term N that is independently
predicative. Geach demonstrates this feature by observing that if ‘good’ were
independently predicative, then the expression ‘x is a good N’ would be logically
equivalent to ‘x is good and x is an N’; but it is not, just as ‘x is a big flea’ is not logically
equivalent to ‘x is big and x is a flea’, as illustrated by the fact that a big flea is also an
animal, but it is not therefore a big animal. The term ‘big’ modifies the term ‘flea’.
Similarly, to predicate ‘is good’ of some object x is, in effect, to say of x that it is good
for or as an N, or simply a good N, where N denotes a comparison class F of which x is a
member (e.g., the kind of thing x is, such as a flea, a pet, a book, a sunset, etc.).
345
Other
examples of attributive predicate modifiers are ‘slow’, ‘tall’, ‘heavy’, and as already
mentioned, ‘big’.
346

Similar to ‘good’ and mutatis mutandis all other attributives, the attributive
predicate expression ‘is big’ simpliciter is analyzed as the complex predicate ‘is a big N’.
This complex predicate applies to an object x iff x is a member of the comparison class F

344
Geach, P. (1956: 33-42).

345
Geach (ibid) emphasizes the latter (viz., a good F), whereas Samuel Wheeler III (1972: 310-34)
emphasizes the former (viz., good for or as an F).

346
Interestingly, Wheeler (1972) argues that color terms such as ‘red’ also function as attributivesfor
instance something can be red for a face, or red as a wine, or red as far as hair color goes.

345
denoted by N, and x is larger in size than most other members of F.
347
While it is fairly
clear that the attributive adjective ‘big’ is applicable along some scale of size and, for
instance, ‘tall’ along some relevant standard of height, it may not be immediately obvious
what the relevant standard of evaluation is to which ‘good’ is applicable. One might
suggest as an answer that the relevant standard is one of value. This, however, seems no
more helpful than saying that ‘good’ applies along some relevant dimension of goodness;
we might just as well ask what the relevant standard is for assessing value. What is
needed is a clear sense of how we could go about applying the standard in question, as in
the case of bigness or tallness. How exactly does Geach intend to extend this sort of
analysis to the term ‘good’?
Unfortunately, Geach does not provide an analysis of the term ‘good’ of the type
given above for ‘big’, but he does hint at the term’s connection with “wants” and
“desires”. John Mackie is more explicit, stating: “to be called good a thing must be such
as to have some satisfying relation to something like interests,” and “‘good’, I think,
always imports some reference to something like interests or wants.”
348
Along these
lines, I suggest that ‘good’ always imports reference to ends (or goals).
349
The proffered
attributive analysis of ‘good’ can be stated as follows.
‘x is a good N’ means roughly: x is a member of the comparison class F
denoted by N, and x is such as to fulfill/further the contextually relevant
ends pertaining to F to a higher degree than most other members of F.

347
For a similar analysis, see Scott Soames (2002, Vol. 2: 148-52).

348
J. L. Mackie (1977: 62, 58).

349
See also Stephen Finlay (2005) who switches to talk of “ends” from talk of “interests” in (2004).

346
So, for example, x is a good Merlot iff x is a Merlot, and x is such as to fulfill
contextually relevant ends pertaining to Merlots (e.g., the relevant taste, bouquet, color,
viscidity, acidity, sugar content, etc.) to a higher degree than most other Merlots.
§IX.3 Troubleshooting the proffered attributive analysis of ‘good’
In this section, we will clarify the proffered attributive analysis of ‘good’ by
addressing a few problems associated with the analysis.
1. One problem for the analysis is that ‘good’ does not always seem to be used
attributively. Even Mackie, whose account of ‘good’ clearly shares much in common
with Geach’s, questions the attributive analysis. Mackie writes: “Contrary to what Geach
claims, ‘good’ is not always attributive in the sense of needing some determinate noun to
lean upon” (1977: 57). Geach claims that ‘good’ must modify a “determinate” noun
because a problem arises otherwise. In some contexts of utterance of a judgment
containing ‘good’—particularly moral contexts—the relevant ends appear to pertain to
such kinds of class of objects as acts, events, or states of affairs. However, such classes
of objects are “indeterminate” insofar as a variety of different kinds of ends—e.g., moral
ends, prudential ends, legal ends, political ends, etc—can be associated with these classes
of objects. Consequently, a problem arises: it is unclear how to give an attributive
analysis of good’ when ‘good’ modifies a predicative noun N in which there are many
different and varied kinds of ends associated with the comparison class F denoted by N.

347
Prima facie, a determinate analysis seems to be all but impossible in such
instances precisely because it is unclear what is the contextually relevant set of ends.
350
It
appears that the attributive account of ‘good’ needs to modify a determinate noun in order
to give a determinate analysis. According to Geach, ‘good’ cannot meaningfully modify
nouns such as ‘acts’ or ‘states of affairs’ because they are, as he puts it, “too empty” or
“vague”.
351
Geach concludes that such uses of ‘good’ are therefore “illegitimate.”
I disagree with Geach, and I agree with Mackie: ‘good’ does not need to lean on a
determinate noun. ‘Good’ can meaningfully modify such nouns as ‘thing’, ‘act’, and
‘event’. To say that x is a good act or a good event is not necessarily to say something
illegitimate or ambiguous. As discussed, the problem for the attributive analysis is that
there are many different and varied kinds of ends that can pertain to such classes of
objects as acts, events, or states of affairs. To avoid this problem, the attributive analysis
must hold that the relevant kind (or set) of end(s) is somehow specified, but how?
I suggest that there are two primary ways in which a speaker can indicate what is
the relevant kind (or set) of end(s) without the need of a “determinate” noun N to modify:
it can be indicated either explicitly by the utterance or implicitly by the context of
utterance. So, for example, in moral contexts the relevant kind of end is the moral

350
Scott Soames (2002: 151) also observes this same basic difficulty. Soames also points to a similar
problem of analysis where ‘good’ modifies predicates in which, apart from special stage-setting, it is
unclear what the contextually relevant ends would be, such as, for example, particles of dust or balls of lint.
I will not address this problem since it doesn’t pertain to the present discussion of value terms used in
moral contexts.

351
Geach (1956).

348
kind.
352
This reference to moral ends is indicated either explicitly—e.g., by using the
expression ‘morally good’—or implicitly via conversational implicaturei.e., the
context of utterance indicates that the relevant ends are moral ends.
353
Similarly, in non-
moral conversational contexts, the context can indicate reference to a particular kind (or
set) of non-moral ends.
354
For instance, the conversational context might implicitly
indicate that the relevant kind of ends is (non-moral) prudential ends, or it might do so
explicitly—e.g., by using the expression ‘prudentially good (N)’. This applies mutatis
mutandis to other kinds (or sets) of non-moral ends—e.g., rational ends, legal ends, etc.
Human beings are exceptionally good at picking up on such contexts of utterance
and thereby the relevant kind of end. Consider, for instance, the following examples:
(a) Giving to the poor is good.
(b) Exercising for at least 45 minutes daily is good.
Apart from special stage setting, (a) is quite naturally read as a moral claim. That is, it is
readily interpreted as saying something like: Giving to the poor is a morally good
thing/act to do. So interpreted, ‘good’ in (a) implicitly modifies a predicative noun—viz.,
‘thing’ or ‘act’—to which many different and varied kinds of ends can pertain. Yet (a) is
not typically taken to be illegitimate or ambiguous because the relevant ends are naturally
presumed to be moral ends. Thus, according to proposed analysis, (a) means roughly:

352
See also Finlay (2004, 2005). Whereas I give a Humean account of what makes an end moral, Finlay,
as we shall discuss later, offers a different and I think problematic conception of what constitutes a moral
end.

353
H. P. Grice (1989) developed the notion of conversational implicature and revealed its importance.

354
But it might not. In such cases, clarification will be needed before a definitive analysis can be given.

349
Giving to the poor is an act, and giving to the poor is such as to fulfill/
further the moral ends pertaining to acts to a higher degree than most other
acts in the comparison class F (e.g., those acts an agent might perform).
355

In other words, compared to all the other acts available to an agent (e.g., buying a
yacht, soaking in a tub, watching TV, eating a slice of pizza, etc.) giving to the poor tends
to fulfill moral ends to a higher degree. In the example above, one could (I think
convincingly) argue that the relevant comparison class should be further refined such that
giving to the poor is not being compared to all acts but rather to those acts pertaining to
one’s money. In other words, (a) is best understood as saying something like: Giving
one’s money to the poor fulfills moral ends to a higher degree than most other things one
could do with one’s money (e.g., buying a yacht or a slice of pizza, going to the movies,
investing in IBM, etc). It is noteworthy that judgments are often restricted to a particular
set of circumstances (C), and this can affect the comparison class. For instance, I might
want to know whether giving to the poor is a morally good thing to do in the particular
situation C in which I find myself. I might judge that it is not a morally good thing to do
in C, but still maintain that, generally speaking, giving to the poor is a morally good
act.
356

Regarding (b), it too is not taken to be either illegitimate or ambiguous because
the relevant ends are naturally presumed to be prudential ends. Apart from special stage

355
We shall discuss what constitutes a moral end in the following section.

356
The conversational context will often indicate whether a moral judgment is relativized to a specific set
C.

350
setting, I suggest (b) is naturally read as saying something like: Exercising for at least 45
minutes daily is a prudentially good thing/act to do. As such, (b) means roughly:
Exercising for at least 45 minutes daily is an act, and exercising for at
least 45 minutes daily is such as to fulfill/further the prudential ends
pertaining to acts to a higher degree than most other acts in F.
In other words, as far as acts go, exercising for at least 45 minutes daily is a prudentially
good one. Just as in the moral case, our judgments about what is prudentially good for us
can be relativized to a particular set of circumstances C. For example, I might want to
know whether it is prudentially good for my grandmother to exercise for at least 45
minutes daily given her asthma, weak heart, and high blood pressure.
2. Another related problem for the analysis is that ‘good’ does not always seem to
be used comparatively. For instance, Mackie disputes the comparative aspect of Geach’s
attributive analysis when he writes: “Something may be called good simply in so far as it
satisfies or is such as to satisfy a certain desire” (1977: 27). Thus, according to Mackie,
whether or not an object is good is wholly a matter of whether or not an object satisfies or
is such as to satisfy a given set of desires or interests. If so, then a comparison with other
objects would therefore be irrelevant to its being good.
I don’t deny the possibility—and indeed the plausibility—that sometimes ‘good’
might not involve comparison of objects. Following Mackie, if this were the case, then
‘good’ would be able to be used as an independent predicate and not as a comparative,
attributive adjective. It would refer to a relational property that holds between a

351
particular object (or set of objects) and a particular end (or set of ends).
357
Thus, in such
instances, ‘good’ would be analyzed roughly as follows:
‘x is good’ means roughly: x is such as to fulfill/further a set of
contextually relevant ends to a (specifiable) sufficiently high degree.
It is noteworthy that this non-comparative analysis of ‘good’ is congruous with
the comparative analysis given above. The only difference is that, in the comparative use
of ‘good’, ‘good’ functions as an attributive adjective that modifies a predicative noun N.
Hence the relational property turns out to be a comparative relation between objects that
belong to a comparison class and a particular end (or set of ends). And this comparative
relation is denoted by the complex predicate ‘is (a) good (N)’. In the non-comparative
use, there is no comparison class, and hence the relational property simply holds between
the relevant object(s) and end(s), and it is denoted by the simple predicate ‘is good’.
Although I do not wish to deny the possibility of a non-comparative, non-
attributive use of ‘good’, I would still like to express my reservations about this account.
I am not yet convinced that there is any putative instance of a non-comparative,
predicative use of ‘good’ that cannot be explained by the comparative, predicate-modifier
account presented above. In other words, all putative non-comparative, predicative uses
of ‘good’ may very well be disguised instances of the comparative, attributive use.
Throughout the rest of the project, I shall therefore remain agnostic about whether
the non-comparative, predicative analysis of ‘good’ is in some contexts of
utterance/belief correct. However, as far as I can tell, ‘good’ is used comparatively in

357
Finaly, S. (2005) defends this “end-relational” account of ‘good’.

352
moral contexts of utterance/belief. In moral contexts, it appears that we are always
comparing the relevant object (e.g. a particular act in C) to other objects of its kind (e.g.,
all other acts available in C), or so I suggest. Since this project is primarily concerned
with moral contexts of utterance/belief, we will confine our discussion to the
comparative, attributive predicate-modifier analysis of ‘good’.
Also noteworthy at this point is the proposed shift from reference to “interests”,
“desires”, or “wants”, a la Mackie and Geach, to reference to “ends” or “goals”. The
shift in terminology helps to distinguish ends and goals from psychological states such as
wants and desires. Mackie partially addresses this by saying that the object, in order to
be a good one, must be “such as to satisfy” the interest or desire in question rather than
simply saying it must “satisfy” that interest or desire. As he observes, “a good carving
knife is still a good one if it is never used, and never even needed” (1977: 56). Hence
Mackie remarks that in order for an object to correctly be considered good, “it is enough
that the thing should be such that it would satisfy wants [or] interests … if any were
brought to bear upon it” (ibid, my emphasis).
The proposed reference to ends and goals enables this project to completely drop
reference to psychological states such as wants, desires, or interests. The object in
question can be such as to fulfill contextually relevant ends or goals even if no interests,
wants, or desires are ever brought to bear upon it. Furthermore, the shift in terminology
is also congruous with the language deployed by Hume in discussing the role of causal
reasoning in sympathy, as well as by cognitive scientists in discussing the cognitive
processes involved in empathy and moral cognition. (I shall make more of this in §IX.5.)

353
3. Another possible threat to the proposed analysis of ‘good’ is G.E. Moore’s
famous “open question” argument.
358
The open question argument is basically designed
to show that any definition or descriptive semantic analysis of ‘good’ will inevitably fail.
No matter what definition or descriptive analysis is offered, Moore claims that it is
always an “open”, or meaningful, question to ask of whatever satisfies the definition or
analysis whether it is good. But Moore thinks that this question would be “closed”, or
trivially true, when asked using the correct definition or analysis of ‘good’. According to
Moore, it should be as trivial and self-answering as asking: “Granted that x is good, but is
x good?” On the basis of this open question test, Moore concludes that ‘good’ must
therefore refer to a “simple”, indefinable/unanalyzable property.
359

While it is widely acknowledged that Moore’s open question argument is flawed
insofar as it doesn’t adequately distinguish concepts/terms from the properties for which
they stand, many scholars nonetheless take it to function as an argumentative device that
demonstrates the futility of providing any descriptive analysis of ‘good’ (as used in moral
contexts).
360
This project challenges this contention by claiming that the attributive

358
G.E. Moore (1903, section 13).

359
In this respect, Moore thought that ‘good’ is like ‘yellow’, which also refers to a simple, indefinable/
unanalyzable property. But Moore also claims that the property of goodness is epistemologically different
from the property of yellowness insofar as we can tell that something is yellow by sense perception but we
can only tell that something is good by “intellectual intuition.” Moore (arguably) thought that this
epistemological difference is rooted in a metaphysical difference between the two properties: whereas
yellowness is a metaphysically “natural” property, goodness is a metaphysically “non-natural” property.

360
Horgan and Timmons (1992) have tried to revive Moore’s open question argument in a way that makes
the distinction between concepts and properties. They argue that their thought experiment, which they
label Moral Twin Earth, shows exactly what Moore (mistakenly) thought his original argument showed,
namely, that ‘good’ does not refer to any property that can be given a reductive or explanatory account.
However, unlike Moore, Horgan and Timmons deny that there are irreducible, inexplicable, “non-natural”
moral properties, and so they conclude that this demonstrates ‘good’ must not be a descriptive term.

354
analysis of ‘good’ withstands Moore’s open question argument. As it applies to the
proffered analysis, Moore’s open question test asks: Granted that x fulfills the
contextually relevant ends (e.g., moral ends) for (morally) evaluating Fs (e.g., acts) to a
higher degree than most other members of F, but is x a (morally) good F? This project
contends that this question is not “open”. It is, in fact, a self-answering question.
361

§IX.4 The attributive analysis of ‘good’ & Hume’s moral ontology
There are a few more prima facie problems with the proffered analysis. First, it is
far from clear how the proposed attributive analysis is compatible with Hume’s moral
ontology since, as stated, feelings of approval and disapproval do not even appear in its
analysis of ‘good’. Since the project claims to defend Hume’s (or at least a Hume
inspired) moral ontology, this matter must be addressed. How exactly can the proposed
analysis of ‘good’ be congruent with Hume’s (or a Humean) moral ontology?
362

Second, assuming that the analysis does somehow bring into its analysis feelings
of approval and disapproval such that these feelings determine moral properties, a
semantic puzzle emerges. If, as Hume maintains, moral properties are determined by
appeal to one of only two kinds of moral sentiment, namely, approval or disapproval,
then moral expressions presumably refer to one of only two kinds of moral property, to
wit: the property of being moral and the property of being immoral. Hence the following

361
See also Soames (2002: 151). Interestingly, Mackie (1977: 60-2) makes a somewhat similar point, but
attempts to explain why Moore’s question always appears “open”. He suggests that, in virtue of asking the
question as Moore does, the question implies that the considered object is to be evaluated using a different
standard (i.e., a different set of ends) than the one originally used. I, on the other hand, have asked Moore’s
question in such a way that the question holds constant a particular set of ends (as well as a particular
comparison class). Consequently, the question is thereby made self-answering and hence no longer “open”.

362
Blackburn (1985) argues that if ‘good’ is attributive, then a Humean moral ontology is thereby ruled out.

355
moral expressions all appear to somehow refer to the same moral property: ‘...is morally
good’, ‘...is morally right’, ‘...is a moral virtue’, and ‘...morally ought to...’. Yet
presumably none of these expressions are semantically equivalent since the evaluative
terms ‘good’, ‘right’, ‘virtue’, and ‘ought’ do not appear to be synonymous. But given
Hume’s feeling-dependent account of moral properties, how can Hume (or a Humean)
explain this apparent lack of synonymy between these terms and expressions?
363

Finally, I have argued that the conversational context sometimes helps (either
explicitly or implicitly) to narrow our conception of what counts as the relevant set of
ends. This is particularly important in moral contexts because in these contexts the
relevant comparison class is usually of the sort to which many different and varied kinds
of ends can pertain (e.g., acts, events, or states of affairs). As a solution to this problem, I
have proposed that, in moral contexts, reference is always made to moral ends. This
solution is only helpful, however, if we have an account of what constitutes a moral end.
And as of yet we have no such account. What is the account?
Before presenting the answer, let’s quickly review the semantic proposal. The
proffered analysis holds that the predicate ‘...is a morally good N’ or simply ‘...is good’
(when used in a moral context of utterance) applies to an object x iff x is a member of the
set F (denoted by N) and x is such as to fulfill/further the moral ends pertaining to F to a
higher degree than most other members of F. For example, an act is a morally good act

363
This is also a problem for other interpretations of Hume. Supposing for instance that the evaluative
terms just listed refer to, or merely emote, or express, the feeling of moral approval, then these terms would
presumably be interchangeable across cases without any change in meaning; but it seems clear that they are
not. Justin D’Arms and Daniel Jacobson (2000: 725) also mention this problem for Humean accounts, but
they do not consider any solution to it. See also Michael Smith (1994).

356
iff that act is such as to fulfill/further the relevant moral ends—viz., those pertaining to
the set of acts that comprise the comparison class
364
—to a higher degree than most other
acts of that set. Unfortunately, the account has yet to say what constitutes a moral end.
It is at this point that Hume’s emotion-dependent account of moral qualities enters
the analysis. According to the criteria of Hume’s moral ontology, an end e belongs to the
moral domain if and only if observation or contemplation of e being fulfilled or furthered
or (as the case may be) foiled or frustrated by an object elicits via empathy particular
feelings toward that object—viz., feelings of approval or (as the case may be)
disapproval—within observers under certain conditions. Thus, just as Hume claims,
moral distinctions depend upon human sentiment. With certain conditions in place, our
attitudinal responses determine moral properties, and these emotions also play a crucial
role in apprehending moral properties—just as Hume claims.
Significantly, along with Hume, this project holds that various ends can and do
have this dispositional, feeling-dependent quality. This is significant because it follows
from this, in conjunction with the proposed semantic analysis of ‘good’, that by
attributing this property to ends we can also thereby ascribe them to good-states (i.e.,
states of goodness) and thus to the objects (e.g., particular acts) to which such states (viz.,
states of moral goodness) are ascribed. Thus, in moral contexts, to say that a particular
set of objects x is (morally) good is, in effect, to ascribe to x the properties: (i) of being
such as to fulfill or further moral ends to a higher degree than most other members of the

364
To recapitulate, the range of the comparison class can vary. For example, the range might be limited
only to those acts available to an agent within a specified set of circumstances C, such as the circumstances
in which an agent finds herself.

357
comparison class to which x belongs, and consequently also (ii) of being such as to elicit
toward x (via empathy) feelings of approval within appraisers when appraisers (who meet
certain conditions) observe or contemplate the relevant ends being fulfilled or furthered
by x. In short, both ends and objects can have this (Humean) moral property.
§IX.5 Differentiating moral properties & values
The project proposes to generalize the account just given by maintaining that all
value terms (e.g., ‘good’, ‘virtue’, ‘right’, ‘ought’, etc.) import reference to ends/goals.
According to this account, values are to be generally conceived as relational properties
that obtain between objects—viz., the objects to which a value property is ascribed—and
a given set of ends/goals.
365
In short, an object has positive or (as the case may be)
negative value insofar as it furthers or fulfills or (as the case may be) frustrates or foils a
given set of ends/goals.
366

With this theory of value in place, we now have a Humean account of moral
value: an object has positive or (as the case may be) negative moral value insofar as it
furthers or fulfills or (as the case may be) frustrates or foils moral ends/goals, that is,
ends/goals that elicit particular emotions under certain conditions. As described above,

365
This general theory of value traces all the way back to Socrates who, in Xenophon’s Memoirs, Book III,
chapter 8, states: “Everything else that we use is considered to be fine and good in accordance with the
same standard—namely, the end for which it is serviceable” (Conversations of Socrates, translated by H.
Tredennick and R. Waterfield, 1990). Besides Socrates, Geach, and Mackie, the notion that value is a
relation between objects and ends, or something similar such as interests, needs, or wants, can also be
found in the work of George Santayana (1913); D.W. Prall (1921); R. B. Perry (1926); Paul Ziff (1960);
and more recently Stephen Finlay (2004, 2005).

366
To recapitulate, here and throughout this work I use ‘frustrate’ and ‘foil’ in contradistinction to the
difference marked out by ‘further’ and ‘fulfill’: one can perform an act that furthers an end or goal without
fulfilling it. Thus, as I use the terms, one can perform an act that frustrates an end or goal without foiling it.

358
by attributing moral properties to ends/goals we can thereby ascribe them to value-states,
as well as to the object(s) to which such value—viz., moral value—is ascribed.
The proffered analysis, as such, draws a distinction between moral properties and
value properties, even though moral properties are also closely associated with values.
Moral properties are emotion-dependent qualities. They are, in effect, relational
properties that hold between emotions and ends, (and thereby) value-states, and
objects.
367
Values, on the other hand, are not dependent upon our emotions. They are
relational properties that hold between ends and objects. In the characteristic way
described above, moral properties and value properties conjoin to produce moral value.
Moral evaluative thought/judgment is therefore about/refers to both of these properties.
This proposed integration of these distinct properties—viz., moral properties and
values—is remarkably congruous with the integrative approach discussed in the previous
chapter. Recall that, according to that approach, our ability to represent and evaluate
short- and long-term ends, goals, and outcomes is automatically linked to motivational
salience due to the observed integration of the different neural networks/ brain regions
associated with structured event knowledge, social perceptual and functional features,
and central motive states (i.e., the cognitive components that constitute what, at the
phenomenological level, is described as empathy). Thanks to this cognitive integration,
our observation or contemplation of particular short- and long-term ends/goals being

367
It might be thought that this is a significant departure from Hume’s view, but part one has, in effect,
already argued that it is not. Hume holds that morals are “a relation, which nature has placed between the
form [of an object] and the sentiment” (SOT, paragraph 10). The relevant form of the object is that which
accounts for the fact that the object fulfills or foils various ends to which we then sympathetically respond.
(See part one, especially chapters II and V, for further discussion; see also §IX.7)

359
fulfilled or furthered or (as the case may be) foiled or frustrated automatically evokes
particular emotional/motivational states in us (when certain conditions are met).
The suggestion of this project is that the moral domain, including moral value, is
marked out and apprehended by us due to this integrative, cognitive-cum-neurological
process of empathy. As we shall discuss, it is this automatic linkage between ends and
emotions (via empathy) that accounts for the special motivational and action-
guiding/normative force typically associated with our conception of moral value.
According to the semantic proposal given above, this special linkage is actually reflected
in the language we speak. The very notion of moral value involves the integration of
moral properties on the one hand and value properties on the other. In other words, it
involves a unique integration of particular conative emotions with (an object’s or
objects’) fulfilling/furthering or (as the case may be) foiling/frustrating of various ends.
It is noteworthy that the proposed analysis of value terms can already, on its own,
account for a link between our value judgments (e.g., our judgment of an object’s
goodness) on the one hand and guiding choices and actions, as well as expressing
approval/disapproval, on the other. Many have thought that this is impossible. For
instance, R.M. Hare (1952) argues that no such account can be given. He writes:
Let us suppose for the sake of argument that there are some ‘defining
characteristics’ of a good picture. It does not matter of what sort they are;
they can be a single characteristic, or a conjunction of characteristics, or a
disjunction of alternative characteristics. Let us call the group of these
characteristics C. ‘P is a good picture’ will then mean the same as ‘P is a
picture and P is C’… [However] if ‘P is a good picture’ is held to mean
the same as ‘P is a picture and P is C’, then it will become impossible to
commend pictures for being C; it will be possible only to say that they are
C… (84-85)

360

Hare’s point is that, for any “naturalistic” descriptive analysis of good put forth,
that analysis will subsequently fail to capture the commending function of good, a
function that appears to be an essential aspect of the notion of good. Hare contends that
this performative function of commendation is the “primary meaning” of good, although
he does think that good has a “secondary” descriptive meaning. However, this
“secondary” descriptive meaning varies across contexts. As Hare puts it, “there is no one
common property which is recognizable in all cases” (103).
I suggest that Hare is mistaken on all counts. The proffered attributive analysis
gives the “defining characteristic” or property of all things good, and that descriptive
analysis is the only meaning of good. The commending function of good is a byproduct
of its descriptive meaning. According to the analysis, the defining characteristic/property
that all good objects have in common is the characteristic/property of being such as to
fulfill (or further) the contextually relevant ends pertaining to a comparison class F (of
which the object is a member) to higher degree than most other members of F. To say
that an object is good just is to say that the object has this property. So, to take Hare’s
example, the defining characteristic/property of any good picture P is that the object P is
a picture, and P is such as to fulfill (or further) the contextually relevant ends pertaining
to pictures to higher degree than most other pictures.
Hare rightly points out that good-making properties/characteristics vary
depending upon the object(s) under evaluation and the standard (i.e., the set of ends) of
evaluation. But his argument mistakenly assumes that all descriptive analyses of good

361
must hold that good stands for the characteristics/properties that make the objects good.
There is plausible alternative for descriptivists. Values are relational properties that hold
between objects and ends. Goodness is just one instance of such a relational property.
Moreover, this end-relational account of goodness can explain why the characteristics to
which Hare appeals in his argument are good-making properties. They are the particular
characteristics (or what Hume calls the “form”) of objects in virtue of which those objects
fulfill/further the contextually relevant ends/goals to a higher degree than most other
objects belonging to the comparison class. The end-relational account also explains why
these good-making characteristics vary. The good-making characteristics will vary
depending upon the contextually relevant ends, as well as the particular comparison class.
Following in Hare’s footsteps, Allan Gibbard writes:
[T]he special element that makes normative thought and language
normative… involves a kind of endorsement—an endorsement that any
descriptivist analysis treats inadequately… In a community of stable,
widely accepted norms, this element of endorsement might be carried by
properties—the properties that, in everyone’s mind, qualify a thing for this
kind of endorsement… No one, though, has found such a property, and so
we still need a language fit for fundamental normative inquiry. (1990: 32-
34)

Can the account accommodate the endorsing/commending function that Hare,
Gibbard, and others
368
claim is essential to the notion of good and other normative
concepts/terms? Indeed it can. Someone who judges that some object x is, for instance, a
good F says that x is such as to fulfill the contextually relevant ends pertaining F to a
higher degree than most other members of F. This judgment is consequently normative
for—i.e., it serves to guide the choice/action of—anyone who has an interest in those

368
See, e.g., Dreier (1990), Horgan and Timmons (1992), Blackburn (1993, 1998), and Timmons (1999).

362
ends because the judgment indicates that that interest is more likely to be satisfied by
choosing/doing x than most of the other members of F. And insofar as the speaker issues
the judgment from the perspective of someone who has such an interest in those ends—
whether that be the speaker herself, someone else, or both the speaker and others—the
judgment that x is good also serves to convey commendation of x.
369
In this way, the
end-relational theory of value offers an explanation of how evaluative terms (e.g., ‘good’)
can be given a descriptive analysis that accounts for a connection with
approval/disapproval.
However, as Gibbard, Blackburn, Timmons, and others suggest, the link between
our judgments of moral value and our feelings of moral approval/disapproval is different
in important respects from the link between our judgments of value per se and approval/
disapproval. It seems that there is something special about ‘good’ and other evaluative
terms as used in moral contexts—perhaps this is the real point of Moore’s argument.
This project concurs. The proposed Humean account holds that, in this moral use,
the relevant feelings of approval/disapproval are byproducts of empathy. The interest we
take in moral ends is primarily due to our empathic feelings, which are the result of an
automatic linkage between the various cognitive components discussed in chapter VIII.
370

This explains why psychopaths take no real interest in moral ends and why those like

369
Cf. Finlay (2004, 2005).

370
This is not to deny that there can be competing interests that sometimes override our interest in moral
ends. Nor is it to deny that one could take an interest in moral ends for purely self-interested reasons, but
such a case would be an exception to the rule.

363
Elliot
371
take very little interest in moral ends: the automatic cognitive linkage is severed
in these individuals.
Furthermore, to judge that an object has moral value is to judge that it has inter
alia an intersubjective, dispositional, emotion-dependent property, namely, a moral
property. Consequently, a moral value judgment is expressive of feelings of
approval/disapproval even when the speaker takes no interest in moral ends. It is
expressive of feelings of approval/disapproval (even if not the speaker’s own feelings)
because these feelings, along with empathic concern, determine the domain of moral
properties, including the moral quality ascribed by the judgment. Thus, unlike judgments
of value per se, there is a special connection between our judgments of moral value and
these feelings of approval/disapproval, namely, a semantic/conceptual connection.
The analogy with color is probative.
372
When we judge that an object is a certain
color, say red, we are also typically experiencing a corresponding visual impression,
namely, a red impression. Our color impressions naturally guide our judgments of color
(arguably) because our color impressions (with certain conditions in place) determine the
domain of colors—i.e., colors are visual-impression-dependent properties. Consequently,
we typically assume that individuals are experiencing a color impression that corresponds
to the color judgment they utter, but of course they need not. Regardless, in either case,
judgments of color are expressive of color impressions insofar as the domain of colors is

371
See the discussion in chapter VIII.

372
The following discussion is taken from §II.V.

364
determined by those impressions. There is no saying what the colors ascribed by color
judgments are without ultimately appealing to the color impressions that determine them.
Similarly, this project maintains that there is no saying what the moral qualities
ascribed by moral value judgments are without ultimately appealing to the feelings of
approval/disapproval and empathic concern that determine those qualities. And this
explains why moral value judgments, unlike other value judgments, are expressive of
feelings of approval/disapproval even when the speaker of the judgment takes no interest
in moral ends, that is, even when the speaker is not expressing her own feelings of moral
approval or (as the case may be) disapproval.
The proposed semantic/conceptual link between our moral value judgments and
feelings of approval/disapproval helps to explain why some scholars have (mistakenly)
thought that we cannot issue moral judgments without these corresponding feelings. And
hence it also helps to account for the heated debates about the plausibility of the
“amoralist”, that is, someone who sincerely believes the moral judgments she issues
while at the same time being unmoved by them.
Hare, for instance, claims that when the amoralist utters the expression “ϕ-ing is
[morally] good” she is, in effect, using ‘good’ in inverted commas sense because she
actually means something like “ϕ-ing is in accordance with what other people judge to be
[morally] good” (1952: 124-6, 163-5). Hare is mistaken, but his error is understandable.
When the amoralist says that, “ϕ-ing is (morally) good”, she means roughly: “ϕ-ing is
such as to fulfill moral ends to a higher degree than most other acts (that comprise the
comparison class)”. But since the amoralist lacks the relevant phenomenal experience

365
used to identify moral ends, the only evidence she could have in support of her belief that
ϕ-ing is (morally) good would be acquired by observing whether ϕ-ing fulfills to a higher
degree than most other acts what others judge to be moral ends. The amoralist can best
determine this by observing/judging whether ϕ-ing is in accordance with what others
judge to be (morally) good.
373
Hence Hare is right that the amoralist makes this sort of
judgment, but he is wrong in thinking ‘good’ therefore has a different meaning for her.
In sum, our value judgments refer to the relation between the object(s) under
evaluation and the fulfilling or foiling of a given set of ends. According to the proposed
account, our belief-forming faculties regarding the fulfilling or foiling of certain short-
and long-term ends, goals, and outcomes tend to have an automatic link (via empathy)
with particular motivational states. Consequently, our judgments of moral value (e.g.,
moral goodness) normally have motivational salience.
374
Those who lack any empathic/
moral response to such considerations are (thankfully) a rare exception to the rule.
While it is true that moral value is a (complex) property for which people have a
contingent but sufficiently universal disposition to experience particular motivational
states, it is also true that this property is inherently tied to such motivational states. The
feelings of approval/disapproval determine the domain of moral value (at least when
certain conditions are met). Consequently, these feelings are, as it were, built-into the

373
The one rare exception to this occurs when the amoralist is like Elliot. In this case, the amoralist can
rely on her own knowledge/memory of what ends are moral ends, knowledge acquired in her past when she
did have the requisite phenomenal moral experience/response to learn on her own what ends are moral
ends. (See the discussion in chapter VIII.) It is noteworthy that individuals such as Elliot are about as close
as one can get to an empirical confirmation of the plausibility—indeed the actuality—of an “amoralist”.

374
This doesn’t entail that there are no competing motivations, as well as various techniques (e.g.,
distraction) to mitigate the motivational salience of consideration of those ends, goals, and outcomes.

366
semantics of the terms we use to refer to those properties. Similarly, while it is true that
the color of an object is a property for which people have a contingent but sufficiently
universal disposition to experience particular visual states, it is also (arguably) true that
this property is inherently tied to such visual states. These visual states determine the
domain of color (at least when certain conditions are met). Consequently, these states
are, as it were, built-into the semantics of the terms we use to refer to those properties.
As a result of this semantic/conceptual linkage, we can judge that a particular
object x has moral value in spite of the contingent psychological connection between
objects and ends on the one hand and certain conative emotions on the other. For
instance, many of us would claim that a particular object x has moral value even if there
was no (all but) universal disposition for empathic/moral concern. While such a claim
might initially appear to be in tension with the proposed account, in fact it is not. This is
because the proposed semantic/conceptual linkage opens the door for the possibility that
the terms ‘moral’ and ‘immoral’, and expressions (explicitly or implicitly) containing
them (or their derivatives), rigidly designate our actual dispositions to have the requisite
type of phenomenal experience.
Consider once again the analogy with color and color terms. Most of us would
not judge that objects (e.g., raspberries and strawberries) would be a different color
simply because humans’ color sensibilities were different. This is (arguably) because we
take our color terms to rigidly designate our actual color sensibilities. This sort of
rigidification of our color terms thereby rules out a certain kind of color relativism.
Similarly, many of us appear to take our moral terms and expressions to rigidly designate

367
our actual moral sensibilities. This sort of rigidification of our moral terms/expressions
thereby rules out a certain kind of moral relativism. David Wiggins is one of the first
scholars to suggest this sort of rigidification when he writes:
x is [morally] good if and only if x is the sort of thing that calls forth or
makes appropriate a certain sentiment of approbation given the range of
propensities that we actually have to respond in this or that way. (1998:
205, his emphasis)

§IX.6 Advantages over the traditional semantic approach
Wiggins’ Humean theory and the Humean theory proposed by this project agree
that there is a necessary link between moral properties and certain affective sensibility-
states (viz., feelings of approval and disapproval), and (possibly) an a priori connection
between our moral terms that denote these moral properties and our actual sensibilities
that give rise to these states. Our theories agree that, just as color terms and the
dispositional, intersubjective properties to which they refer seem to have a kind of “to-be-
seenness” conceptually built into them, so too do moral terms and properties appear to
have a kind of “to-be-pursuedness”, or action-guidingness, conceptually built into them.
However, Wiggins takes (what I call) the traditional approach in giving semantic
analyses compatible with a Humean-type moral ontology. He offers a sensibility theory
as a theory of value per se.
375
So, for instance, he takes a value term such as ‘good’ to
refer to a (Humean) moral property. Thus, when Wiggins generalizes his claim quoted
above about rigidification, he proposes to rigidify predicates of value per se. He writes:
[F]or each value predicate ϕ, there is an attitude or response of subjects
belonging to a range of propensities that we actually have such that an

375
See also, for instance, John McDowell (1985, 1987) and G. Sayre-McCord (1988, 1991).

368
object has the property ϕ stands for if and only if the object is fitted by its
characteristics to bring down that extant attitude or response upon it and
bring it down precisely because it has those characteristics. (1998: 205)

The theory advanced by this project does not agree with the traditional approach
to giving an account of value predicates. Values per se are not dispositional, attitude-
dependent properties as Wiggins and others suggest. Unlike moral properties, values per
se neither exhibit a necessary nor an a priori tie to conative states. Values are entirely
attitude-independent and thus distinct from moral properties. The distinction between
moral properties and values plays a crucial role in defending this project’s Humean
sensibility moral theory against various objections to which Wiggins’ and others
Humean-type sensibility moral theories are arguably subject.
376

For instance, there is arguably important differences between value terms as used
in moral contexts and value terms as used in non-moral contexts. Consequently, the
traditional account, insofar as it accepts such a difference, cannot offer a univocal
analysis of evaluative terms. For example, ‘good’ would have to be given a different
analysis in moral contexts from the analysis of ‘good’ as used in non-moral contexts.
The sort of rigidification Wiggins proposes for ‘good’ and other value terms (e.g., ‘right’,
‘virtue’, and ‘ought’) illustrates this difference. Clearly evaluative terms such as ‘good’,
as used in non-moral contexts, are not guaranteed a priori to be tied to our sensibilities,

376
It is noteworthy that the proposed view has already explained why judgments of value per se—e.g. that
some object x is a good—do prima facie appear to have direct ties with our emotional responses. Thus it
can explain why a sensibility theory of value per se does appear to be prima facie plausible. Consequently,
it can explain why it is rather easy to read Hume as offering a sensibility theory of value and evaluative
terms. Hume’s theory, however, is not forced into this reading. As we shall see, the proposed distinction
between moral properties and values offers an alternative that is remarkably compatible with Hume’s
account.

369
much less tied to our sensibilities as they actually are.
377
Furthermore, the traditional
approach is also hard pressed to account for the apparent lack of synonymy between
evaluative terms such as ‘good’, ‘right’, ‘virtue’, and ‘ought’ since it presumably holds
that each of these terms is tied to the same sensibility state, namely, moral approbation.
Among the advantages of the proffered semantic analysis of moral evaluative
language and the corresponding distinction between moral properties and values is that it
offers straightforward solutions to both of these problems. First, unlike the traditional
approach, it provides a univocal analysis of evaluative terms such as ‘good’, ‘right’,
‘ought’, and ‘virtue’. Each of these terms does not vary in meaning across moral and
non-moral conversational contexts. Instead, it is the relevant objects and ends that vary
across contexts. Second, according to the proposed analysis, moral evaluative
expressions such as ‘…is (morally) good’, ‘…is (morally) right’, ‘…(morally) ought
to…’, and ‘…is a (moral) virtue’ are not semantically equivalent precisely because the
evaluative terms ‘good’, ‘right’, ‘ought’, and ‘virtue’ are not synonymous. Each of these
evaluative terms has a different descriptive meaning.
Roughly, x is good means: x is such as to fulfill contextually relevant ends to a
higher degree than most members of the comparison class F; x is right means roughly: x
is such as to fulfill contextually relevant ends to a higher degree than all other members
of F; agent A ought to x means roughly: it is more probable that contextually relevant

377
Timmons (1999) forcefully argues that it is highly implausible to hold that evaluative terms such as
‘good’ function as rigid designators, even in moral contexts. This project concurs, but it also suggests that
one can plausibly hold that the term ‘moral’ and its derivatives rigidly designate our actual sensibilities.

370
ends are fulfilled by A x-ing than any other member of F;
378
and x is a virtue means
roughly: x is a quality that is such as to typically fulfill contextually relevant ends.
All of the above analyses entail that none of these evaluative terms per se refer to
moral properties. Rather, the addition of the adverb ‘morally’ or adjective ‘moral’, which
modifies these evaluative terms, explains why such evaluative expressions refer to ends
that belong to the moral domain.
379
In those instances in which the adverb or adjective is
not explicitly included in the moral utterance, (typically) the conversational context
indicates that the evaluative expressions containing these evaluative terms refer to moral
ends and hence moral value, rather than non-moral ends and non-moral value.
380

§IX.7 Congruity with Hume’s account
Thus far I have given an analysis of moral evaluative terms and expressions that is
generally compatible with Hume’s moral ontology (as I understand it). In concluding, I
suggest that the proffered analysis is actually quite complementary to Hume’s account.
As discussed, Hume argues that when standard conditions are satisfied our attitudes of

378
This is an analysis of the normative ‘ought’. Cf. Finlay “Oughts and Ends” (2008, forthcoming). I shall
say more about Finlay’s analysis of ‘ought’ and the analysis above in the following chapter.

379
Contrary to Blackburn’s (1985) claim, an attributive analysis of ‘good’ in not necessarily incompatible
with a (Humean) moral ontology that views moral qualities as dispositional, emotion-dependent properties.

380
As we shall discuss in the next chapter, the proposed distinction between moral and value properties also
presents an account that is superior to the end-relational theory of value alone. Most, if not all, of those
who have proffered the end-relational theory of value have tried to accommodate the various features of
moral discourse and practice by appealing to value properties alone. These theories thereby construe moral
properties simply as particular instances of value properties. Unfortunately, moral properties, as such, lose
their inherent tie to motivational states because all values per se exhibit only a contingent link with
motivational states. One important consequence of this is the fact that these theories cannot justify the
objective normative status of morality that the commonplace view presumes, or so I shall argue.

371
approval and disapproval determine the moral domain. However, as discussed in part
one, Hume also argues that our reasoning “paves the way” for these sentiments.
The final sentence, it is probable... that which renders morality an active
principle and constitutes virtue and happiness, and vice our misery: it is
probable, I say, that this final sentence depends on some internal sense of
feeling, which nature has made universal in the whole species. For what
else can have an influence of this nature? But in order to pave the way for
such a sentiment, and give proper discernment of its object, it is often
necessary, we find, that much reasoning should precede, that nice
distinctions be made, just conclusions drawn, distant comparisons formed,
complicated relations examined, and general facts fixed and ascertained.
(EPM, section I, paragraph 9)

As discussed, in Hume’s view the sort of reasoning that we deploy when making
our moral judgments is causal reasoning, in particular reasoning about the causal
tendencies of objects to fulfill or foil ends. For instance, Hume writes: “One principal
foundation of moral praise being supposed to lie in the usefulness of any quality or
action, it is evident that reason must enter for a considerable share in all decisions of this
kind; since nothing but that faculty can instruct us in the tendency of qualities and
actions…” (EPM, Appendix I, paragraph 2). The tendency of which Hume speaks just is
the tendency of objects (e.g., traits and actions) to fulfill or foil a given end or set of ends.
After all, in Hume’s view: “Usefulness is only a tendency to a certain end” (EPM,
Appendix I, part II, paragraph 2). More precisely, it is a tendency to bring about an end
that is agreeable either to the possessor of the object that has the tendency, or to those
who “have commerce” with her.

372
As we know, Hume holds that our moral sentiments, via the psychological
process of sympathy, are responsive to our evaluations of these tendencies of objects.
381

For instance, he writes: “Wherever an object has a tendency to produce pleasure in the
possessor, ...it is sure to please the spectator, by a delicate sympathy with the possessor”
(T 576-7). Thus, concludes Hume, “we must allow, that the reflecting on the tendency of
characters and mental qualities, is sufficient to give us the sentiments of approbation and
blame” (T 577). Given Hume’s moral ontology, it follows that the “distinction between
what is useful, and what is pernicious…is the same in all its parts, with the moral
distinction” (EPM, Section VI, part I, paragraph 5).
382

In sum, if an object is such as to fulfill or foil certain ends, then it is both useful
and moral or (as the case may be) pernicious and immoral. These will be ends that, upon
consideration of their being fulfilled or foiled, elicit attitudes of approval or (as the case
may be) disapproval within those who satisfy standard conditions. Since these attitudes
determine the moral domain, it is a short step to the conclusion that such ends belong to
the moral domain, i.e., that they are moral ends. Thus, that which is truly useful or
pernicious on Hume’s view is simply an object—e.g., an act, trait, or agent—that tends to
fulfill or (respectively) foil moral ends.
The proffered semantic analysis of moral evaluative terms and expressions is
clearly quite complementary with this reading of Hume. For instance, the analysis holds
that moral value terms function as attributives that modify a comparison class and this

381
See, for instance, T 316-24 and T 574-91. I explain in detail how sympathy generates moral sentiments
from these evaluations in my doctoral dissertation [reference omitted].
382
As I read Hume, we are to judge whether an object is truly useful or pernicious from the “moral view.”

373
reflects the importance that Hume assigns to comparison. “By comparison alone”, says
Hume, “we fix the epithets of praise or blame, and learn how to assign the due degree of
each” (SOT, paragraph 22). At the same time, the proposed analysis also makes explicit
the connection between our reasoned evaluations of objects (in terms of their fulfilling or
foiling particular ends) and our sentiments, both of which play a crucial role in Hume’s
account of moral judgment/thought.
For instance, Hume writes: “In moral decisions, all the circumstances and
relations must be previously known; and the mind, from the contemplation of the whole,
feels some new impression of affection or disgust, esteem or contempt, approbation or
blame” (EPM, Appendix I, paragraph 10). Following Hume, the proposed analysis offers
an explanation of the apparent essential connection between human conduct and moral
judgment/thought by appealing to particular conative attitudes (and “standard”
conditions). But the account also improves upon Hume by providing a straightforward
descriptive analysis of the content of moral thought/judgment that is consistent with that
account.
It is noteworthy that this account not only agrees with Hume’s contention that the
primary objects of moral evaluation are not acts but intentions (T 348)—more precisely,
the primary objects of moral evaluation are the agent’s intentions to bring about
particular ends by performing some act. The account is also able to explain why this is
the case. End-directed acts are primarily the kind of acts that are objects of our moral
value judgments, and in order for an act to be directed toward fulfilling or furthering a
particular end, the agent must intend to bring about that end by performing the act.

374
Hence the primary objects of such moral evaluations are the agent’s intentions to
facilitate (through action) the actualization of a particular end or set of ends. As Hume
points out (T 348-9), we tend to look past any accidental consequences of the act and
hold the agent morally responsible for her actions primarily because of her intentions to
bring about a particular end.
Hume observes that a notable exception to this rule is our moral evaluations of an
agent’s character traits (T 348-9). But even in this case, the agent’s repeated intentional
aiming at a particular end is presumably what establishes the habit or disposition in the
first place. And this fact explains why the agent is morally responsible for many of her
dispositions to act in certain ways even in those cases in which the acts themselves,
which are performed out of habit, may not be performed intentionally.
§IX.8 Summation
In sum, we have not only identified an alternative approach to analyzing moral
evaluative terms and expressions that is generally compatible with Hume’s moral
ontology; we have arguably identified the analysis that best fits Hume’s account and
putative semantic facts, such as the lack of synonymy between various evaluative terms
and expressions. Just as Hume argues, (i) moral distinctions depend upon human
sentiment (subject to standard conditions), and therefore (ii) sentiment plays a crucial role
in both the identification and evaluation of moral qualities. However, by explicitly
drawing the ontological distinction between values and moral properties, this project
marks a significant improvement over Hume. First, it offers a clear and concise account
of each kind of property, which Hume does not do. Second, this ontology also enables

375
the project to provide an explicit semantic analysis of moral and evaluative language,
which Hume also does not do.
To reiterate, I have claimed that Hume does not offer any semantic account of
moral evaluative language. So I am certainly not suggesting that the proffered analysis is
Hume’s. The analysis is undoubtedly reconstructive. For instance, Hume does not
distinguish between moral terms and value terms; nor does he draw the corresponding
distinction between moral properties and values.
383
However, as we’ve seen, Hume does
hold that there is an intimate relationship between our evaluations of the tendency of
objects to fulfill or foil particular ends, on the one hand, and our moral sentiments, on the
other. The above analysis is clearly in keeping with this account.
The proposed semantic analysis also has significant explanatory power. For
instance, the attributive analysis withstands Moore’s “open question” argument, and it
can explain how value terms such as ‘good’ need not modify a determinate noun. The
corresponding ontological distinction between moral properties and values opens up an
unexplored avenue for avoiding weighty objections to Humean-type sensibility theories,
objections to which others that do not employ that distinction arguably succumb. It also
enables the project to account for a special semantic/conceptual link between moral value
and motivation. As we shall now discuss, this link aids the project in accommodating the
rather peculiar normative force/authority typically associated with moral judgment.

383
Interestingly, Hume does at one point equate a “desir’d good” with a “design’d end”; see T 416-7.
Hume seems to recognize a distinction between value and moral value. He, unfortunately, does not
concisely explain the grounds upon which that distinction is made.

376

Chapter X: Normative Accommodation

§X.1 Moral normativity

As discussed in chapter I, the resolution of the accommodation problem is largely
a matter of providing an explanation that accommodates the fact that moral thought/
judgment can be practical in its issue to just the extent that it is.
384
This is in part a matter
of accommodating the connection that moral thought/ judgment has with motivation. For
instance, among other things, utterance of a moral thought/ judgment typically expresses
the speaker’s intention and motivation to behave in accordance with that judgment. It
also often motivates others to act similarly. Chapter IX has argued that this project can
accommodate these features of the practicality of morality.
There are, however, other important features that must be accommodated as well.
These pertain to the rather peculiar normative character of moral judgment. The content
of moral thought/ judgment is presumed to be action-guiding and thus normative.
Sincere moral evaluations are not only descriptions; they are also prescriptions. They
make a claim about what one ought or ought not to do or pursue. Hence they also make a
claim about how one ought or ought not to be motivated to act. In short, all moral
judgments appear to entail moral ought-judgments, and all moral facts appear to entail
moral ought-facts.

384
Cf. Michael Smith’s claim that the challenge is “to explain how deliberation on the basis of our values
can be practical in its issue to just the extent that it is” (1992: 330, his emphasis). Of course, this project
contends that moral value has a special connection with practicality as compared to non-moral value.

377
The commonplace view is that these normative (or ‘ought’) requirements of
morality apply to persons independently of their particular motivations and intentions.
Having motivations, intentions, or beliefs in opposition to a moral judgment neither
relinquishes one from nor diminishes its normative authority or force. Moral evaluations,
as such, function as practical reasons with a special—some say “queer”—kind of
normativity: they make a claim of obligation upon persons by giving them reasons—viz.,
normative or ‘ought’ reasons—to conduct themselves in a certain manner irrespective of
their own attitudes, desires, ambitions, intentions, etc. We typically expect persons to
accept and govern themselves in accordance with these reasons, no matter what their
antecedent motivations.
In order to stress this putative attitude-independence, moral evaluations, along
with their peculiar normative status (or reason-giving force), are sometimes described as
categorical. Moral evaluations also consequently have a kind of universality built into
them insofar as they make a claim upon anyone who finds him or herself in the relevant
circumstances to behave (or to refrain from behaving) in a specific manner, regardless of
that person’s desires, ambitions, intentions, etc.
This chapter aims to explain and defend these presumptions about the normative
character of moral thought/ judgment. We shall begin by picking up on the theme of the
previous chapter, namely, semantic analysis. Since the principal class of moral
propositions appears (prima facie) to be moral ought-propositions, this chapter opens
with an analysis of the modal auxiliary verb ‘ought’. According to the proposed analysis,
the normative term ‘ought’ has much in common with the proffered analysis of ‘good’.

378
Both analyses reference a contextually relevant set of ends and a contextually relevant
comparison class. Moreover, I shall argue that, as with ‘good’, there is nothing
particularly peculiar or special about the normative term ‘ought’. Like ‘good’, the special
normativity of ‘ought’ as used in moral contexts has everything to do with the (explicit or
implicit) invocation of moral ends and the feeling-dependent nature of these ends. This,
as we shall see, is remarkably congruous with Hume’s claims about the moral ought.
We will then use this analysis of ‘moral ought’ to help explain the above objective
presuppositions concerning the peculiar “categorical” normative character of moral
judgments and facts. According to that presupposition, the referents of true moral ought-
propositions are, in an important sense, attitude-independent facts. This is the sense in
which moral ought-propositions have normative, or reason-giving, force/authority over
persons irrespective of their antecedent attitudes, motivations, intentions, etc. As we
shall see, the proffered account can accommodate the truth of this presupposition.
§X.2 An analysis of ‘ought’
The proposed analysis of the modal auxiliary verb ‘ought’ draws heavily from the
recent work of Stephen Finlay (2008, forthcoming). Finlay argues (I think correctly) that
the most basic form of the modal term ‘ought’ is found in sentences such as:
(a) It ought to rain tomorrow.
In sentences such as (a), the auxiliary verb ‘ought’ is clearly not used normatively. In (a),
it is used to issue a straightforward predictive claim. According to Finlay’s analysis and
the analysis of this project, this predictive function of ‘ought’ is due to the fact that all

379
ought-propositions (both non-normative and normative ought-propositions) state
comparative probabilities. Roughly, (a) is analyzed as saying something like:
(a*) It is more probable that it rains tomorrow than any relevant alternative.
Generalizing this analysis, all ought-statements of the form It ought to be the case that p
mean roughly: It is more probable that p (e.g., that a particular state of affairs obtains)
than that any other member (e.g., any alternative state of affairs) of the comparison class
F obtains.
It is noteworthy that all modality statements—e.g., all statements containing one
or more of the following modal auxiliaries: can, could, may, might, must, ought, shall,
should, will, and would—always either explicitly or implicitly suppose that a particular
set of circumstances (C) obtain. Hence an implicit supposition that a particular set C
obtains needs to be annexed to the analyses above. The set C qualifies the modality; i.e.,
it functions to define the parameters of the modality assessment.
385
For instance, logical
modalities assess actual and/or possible outcomes (or states of the world) supposing that
the laws of logic obtain; metaphysical modalities assess actual and/or possible outcomes
(or states of the world) supposing that metaphysical laws obtain; nomic modalities assess
actual and/or possible outcomes (or states of the world) supposing that laws of science
obtain; and epistemic modalities assess actual and/or possible outcomes (or states of the
world) supposing that a particular set of beliefs/ knowledge obtain. Different kinds of
modalities yield different results insofar as, and to the extent that, the set of parameters C
vary—e.g., what is logically possible differs from what is nomically possible.

385
Cf. Finlay (2008, forthcoming).

380
As noted, the parameters C can be indicated either implicitly by way of the
conversational context or explicitly. For instance, logical modality can be explicitly
indicated by using a modal auxiliary verb such as ‘must’ and the adverb ‘logically’
yielding the expression ‘logically must’;
386
metaphysical modality can made explicit
using a modal expression such as ‘metaphysically may’,
387
etc. Various other restrictions
can also be included in C. For instance, returning back to the original example, sentence
(a) is a nomic modality; it supposes that the laws of nature apply. But being a predictive
claim, (a) also presumably includes as part of its circumstances C the supposition that
various facts obtain—e.g., various geographical factors and the fact that there is a low-
pressure front slowly approaching the speaker at the time she issues the judgment (a).
We now have an analysis of non-normative ought-statements, but what about
normative ought-statements? Following Finlay (ibid), this project holds that there is no
difference in analysis; all ought-statements are given a univocal analysis. The difference
between the non-normative and normative ‘ought’ is due entirely to the parameters
stipulated by C. Unlike non-normative ought-statements, normative ought-statements
contain, as part of C, reference to a contextually relevant set of ends/goals. Non-
normative ought-statements are transmuted into normative ought-statements by having C

386
Following Finlay (ibid), the statement It (logically) must be the case that p is analyzed as: Every
possible state of the world in which circumstances C (the laws of logic) obtain is such that p.

387
Following Finlay (ibid), the statement It (metaphysically) may be the case that p is analyzed as: At least
one possible state of the world in which circumstances C (the metaphysical laws) obtain is such that p.

381
qualify the statements as end-relational ought-statements. In fact, all modal judgments
turn normative when ends are brought into the set C to qualify the modality.
388

As with other circumstances that comprise C, the contextually relevant ends can
be brought to bear either implicitly (via conversational context) or explicitly. One can do
so explicitly by simply stating the relevant end/ends. Consider the following examples:
(b) Given the aim of not upsetting the boss, you ought to arrive at work on time.
(c) If you should want some free donuts, you ought to arrive at work on time.
(d) So as to make a good first impression, you ought arrive at to work on time.
(e) As a means to keeping your job, you ought to arrive at work on time.
(f) In order to be prepared for the meeting, you ought to arrive at work on time.
389

In each of the examples above, the antecedent clause qualifies the ought-
statement. However, in order to qualify an ought-statement as normative, one need not
refer to a specific end, as do the above examples. There are other ways to explicitly refer
to a set of ends. For instance, one can qualify an ought-statement by indicating that a
particular kind of end is relevant. As discussed in the previous chapter, one can do this
by simply using particular adverbs to modify the normative term. For example, one can
qualify an ought-statement by referencing prudential ends by with the expression
‘…prudentially ought to…’, or by referencing moral ends with the expression ‘…morally
ought to…’, or by referencing rational ends with the expression ‘…rationally ought to…’,

388
Cf. Finlay (ibid).

389
Finlay (ibid) prefers this locution, suggesting that all normative modal auxiliaries have an implicit “in
order that e…” operator—here e stands for the set of contextually relevant ends.

382
etc. Of course, the context of utterance or the subject matter can also implicitly indicate
the relevant kind of ends. Consider the following sentences.
(g) One ought to give to the poor.
(h) One ought to exercise for at least 45 minutes daily.
Similar to the examples given in the previous chapter regarding the term ‘good’,
(g), apart from special stage setting, is quite naturally read as a moral claim—viz., One
morally ought to give to the poor. And (h), apart from special stage setting, is readily
interpreted as a prudential claim—viz., One prudentially ought to exercise for at least 45
minutes daily. The similarity with the analysis of ‘good’ is striking. For instance, both
analyses not only reference a contextually relevant set of ends; similar strategies are used
to reference those ends. Both analyses also reference a comparison class.
Speakers of ought-statements often do not refer to a specific end or ends when
they issue ought-statements because they presume that the relevant set of ends is apparent
to the agent. Typically, the speaker assumes that the relevant ends are the agent’s ends;
i.e., they assume that the agent cares about those ends. In most cases, this explains why
normative ought-statements function to give guidance to those agents that the normative
judgments are directed toward. For instance, lawyers advise their clients about what they
legally ought to do, typically presuming that their clients care about legal ends, at least
insofar as these ends are relevant to the clients’ prudential ends. But this is not always
the case. Sometimes ought-statements have a normative function even supposing that the
agent does not care about the contextually relevant ends. Moral ought-statements are
consummate examples of this. We shall take up this important issue later in §X.4.

383
§X.2.1 Differentiating non-normative & normative ought-statements
Let’s now look more closely at how the reference to ends transmutes non-
normative ought-statements into normative ought-statements. To illustrate this, consider:
(i) You ought to arrive at the office on time.
Sentence (i) can be read both non-normatively and normatively. And the
difference between these two readings depends solely upon whether the relevant
circumstances C include reference to ends. According to the non-normative reading, (i)
is analyzed as meaning roughly: Given C, it is more probable that you arrive at to the
office on time than not. In this case, C presumably includes various states of affairs such
as for example: the time in the morning at which I, the speaker of (i), called you at your
house, the short distance from your house to the office, the lack of traffic during that time
of day, the speed at which you usually drive, etc. Assuming this is the case, (i) can
therefore be interpreted as: Given C, including the particular time I called you at your
house, the short distance from your house to the office, the lack of traffic during that time
of day, the speed at which you drive, etc., it is more probable that you arrive at the office
on time than not. On this reading, (i) is a non-normative comparative probability claim
stating a particular state of affairs, to wit: the time by which you are more likely to arrive
at the office given that certain conditions obtain. By stating this comparative probability,
(i) can be used to make a prediction about when you will arrive at the office.
On the normative reading, (i) is analyzed as meaning: Given C, it is more
probable that you arrive at the office on time than not. This analysis is not surprising
given that the project claims to offer a univocal analysis of ‘ought’. However, in this

384
case, C includes particular ends such as for example: your being prepared for the office
meeting, your eating some free donuts, your not upsetting the new boss, your making a
good first impression on the client, etc. Assuming this is the case, (i) can be read as
follows: Given C, including your being prepared for the office meeting, your eating some
free donuts, your not upsetting your new boss, your making a good first impression on the
client, etc., it is more probable that you arrive at the office on time than not. On this
reading, (i) is a normative comparative probability claim stating a particular state of
affairs, to wit: which means—viz., your arriving at the office on time or not—is more
likely given that certain ends obtain—e.g., given that you eat some free donuts.
An interlocutor might question whether the above comparative probability is
actually normative. As discussed, normative ought-statements are characteristically used
to guide choices and actions, to give advice and/or to criticize. But how can facts about
what an agent is more likely to do (or choose) given certain ends be used to give advice
and/or criticism about what to do or choose? It does not appear that the means that you,
the agent, are more likely to take in order to fulfill/further certain ends has any bearing on
what you ought to do given those ends. This is correct, but I suggest that the analysis of
(i) doesn’t indicate such facts. The interlocutor is misinterpreting the proffered analysis
of (i) due to an underlying ambiguity latent within that analysis.
The relevant facts stated by (i) cannot be facts about what you, the agent, are
more likely to do (or choose) given certain ends. Such facts could be used to make a
prediction about a particular outcome, but they cannot, in themselves, be used to give
advice or to criticize the agent or the agent’s action/choice. Hence these facts are clearly

385
irrelevant to the normative ought-statement (i). This is not the right way to understand
the normative analysis of (i). The normative ought-judgment (i) state