University of Southern California Dissertations and Theses

Defending compatibilism: a paradigm case argument

(USC Thesis Other)

Defending compatibilism: a paradigm case argument

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
DEFENDING COMPATIBILISM: A PARADIGM CASE ARGUMENT

by

Xiaoyu Zhu
_______________________________________________________________________

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

December 2009

Copyright 2009 Xiaoyu Zhu

ii

Dedication

For Kadri – She took my hand, opened my mind and touched my heart.

iii

Acknowledgments

I owe my foremost debt to my advisor Kadri Vihvelin who guided me through the
research and writing of my dissertation with her philosophical wisdom, inspirational
passion and hard work. I am also grateful to my dissertation committee member Janet
Levin for her advice and friendship throughout my years of graduate study at USC.
To Mark Schroeder, the graduate student advisor at USC and Tom O’Neil, the
Dean of Social and Behavioral Sciences at Antelope Valley College, I am indebted for
their tremendous moral and practical support.
I thank my husband, Yongqiang Chen, for being my love and soul mate. I thank
my family, friends and colleagues. Without their endless love, support and
encouragement, this project would never have been completed.
It is to Kai Nielsen that I owe my greatest debt. He is my role model and guide
both in philosophy and life.

iv
Table of Contents

Dedication ii

Acknowledgments iii

Abstract vi

Chapter One: Introduction 1
The Free Will / Determinism Problem 1
The Analysis Strategy 4
The Incompatibilist Challenge 7
The Paradigm Case Strategy 11
Outline 18

Chapter Two: Frankfurt’s Compatibilism 20
Classical Compatibilism 20
Frankfurt’s Compatibilism 22
Frankfurt’s Theory of Free Will and Moral Responsibility 27
Frankfurt’s Answer to the Incompatibilist Challenge 38
Free Will as Mental Occurrence 39
Conclusion 49

Chapter Three: Strawson’s Compatibilism 51
The Examples are Descriptive 54
The Examples are Normative 62
The Harris Case 67
Compatibilism 72
Objective Attitudes 75
Strawson’s Naturalism 78
Conclusion 81

Chapter Four: The Manipulation Argument for Incompatibilism 82
The Martian Device Argument 84
The Four Case Argument 92
The Soft-Line Reply and the Hard-Line Reply 95
Mele’s Criticism 105
Confusions Involved in Kane’s Idea of CNC Control 107
Conclusion 112

v
Chapter Five: The Paradigm Case Argument for Compatibilism 113
PCA as Ostensive Teaching 114
PCA as Elucidation 122
Conclusion 144

Bibliography 146

vi
Abstract

The objective of this study is to solve the free will/determinism problem by
defending a new form of compatibilism. Unlike most classical and contemporary
compatibilists, the author takes the Paradigm Case Argument (PCA) strategy instead of
the analysis strategy to defend compatibilism. She endeavors to offer a correct
conceptual analysis of free will as well as an adequate response to the incompatibilist
challenge.
While the analysis compatibilist wishes to discover conditions that are both
necessary and sufficient for applying the concept of free will, the PCA compatibilist
believes that there is no such thing for free will because not all correct uses of the concept
of free will have something in common. The author distinguishes two versions of PCA,
namely, PCA as Ostensive Teaching and PCA as Elucidation. She rejects the former and
advocates the latter because the latter regards paradigm cases of free will as conceptual
elucidations of free will while the former understands them as instances.
According to the author, the correct account of free will is a conceptual analysis
of free will consisted of the explanations of paradigmatic examples to which free will and
its surrounding concepts are correctly applied.

1

Chapter One: Introduction

1. The Free Will / Determinism Problem
It is common sense that we are persons with free will, who often, and perhaps
typically, act freely. We may freely take a walk in the morning to enjoy the fresh air or
choose freely between tuna sandwich and beef sandwich on the lunch menu. We say to a
friend, “It’s up to you which movie we will watch tonight”; we tell a co-worker, “It’s in
your power to finish your work before the deadline”; we warn our kid, “You’ve got a
choice: either wash your hands now or no ice cream”; we advise a colleague, “You could
have chosen a different topic to speak about at the meeting”. We use expressions such as
“it’s up to you”, “it’s in your power”, “you’ve got a choice”, or “you could have done
otherwise” under above circumstances because we believe that these are cases where our
friend, co-worker, kid or colleague has the freedom of will to make a choice and act
accordingly.
We distinguish normal adults from abnormal adults such as the mentally ill. We
do not think that it is in the power of an obsessive-compulsory disorder patient to stop
washing his hands excessively. The patient cannot choose freely between healthy
behaviors and unhealthy behaviors exactly because he suffers mental illness. We also
differentiate an adult from a child. When a small child spills a cup of coffee on the
carpet, we would not tell her as we would tell an adult, “You could have been more

2
careful.” The child does not have the ability to control her bodily movements well and
her mind is not developed enough to understand that she should have been more careful.
Normal adults do not always enjoy free will because sometimes they lose their
freedom to external force, duress, compulsion, anger, fear, etc. A prisoner cannot freely
visit his family or travel; a mother does not have much choice when her daughter is in the
hand of a kidnapper; a drug addict loses the power to get his work done.
Do normal adults under normal circumstances act with free will? Our common
sense is that they do. But do they act with free will if determinism is true, that is, if the
total state of the world at any time, in conjunction with all the laws of nature, entails a
unique future? This question, as known as the free will / determinism problem, is beyond
the scope of common sense and becomes the issue in the debate between incompatibilists
and compatibilists. The incompatibilist answer to the question is that if determinism is
true, even normal adults under normal circumstances do not act with free will. The
compatibilist disagrees and holds that the truth of determinism is not a threat to the
possibility of free will.
1

The project that my dissertation undertakes is to investigate the free
will/determinism problem, that is, whether we can exercise control over our actions,
choices and decisions if the world is deterministic. This problem is known as one of the
most difficult and persistent problems in the history of western philosophy. During the
last two millenniums, almost all the major philosophers and theologians have grappled

1
Some compatibilists, such as R. E. Hobart, hold that free will requires the truth of determinism. It seems
to me that a requirement as such is the incompatibilist ingredient mixed in compatibilism because a
compatibilist should regard determinism as totally irrelevant to free will. See Hobart, “Free Will as
Involving Determination and Inconceivable Without It”.

3
with it and offered solutions to it. However, none of the solutions seems to be
satisfactory to us, contemporary people living in the late twentieth and early twenty-first
centuries. As we become more and more knowledgeable in science and capable in
technology, our worry that we may not be our own masters does not go away but
becomes more and more acute. Thirty years after the birth of the first test tube baby,
prospective parents who use in vitro fertilization services are now choosing embryos
according to their sexes or tendency to develop genetic diseases. It will not be surprising
if one day we order babies based on their size, look, intelligence, talent or temperament in
the same manner as we order pizzas. If that happens, how much control would such a
baby have over his own desires, choices, or abilities in his adult life? Similar worries are
growing as we know more about the functions of our brain and how our beliefs and
behaviors can be drastically influenced by different drugs or unconscious thoughts. The
development of science has deprived us of our privilege of being the center of the
universe multiple times and in different dimensions. The unpleasant or even depressing
truth that we may have to face in our time might be something like that the chess world
champion Garry Kasparov has no more free will than the chess-playing computer Deep
Blue because if determinism is true a person’s choices and actions will be no more free
than a robot’s choices and actions when it makes a move according to its pre-
programmed codes.
Our achievements in science seem to encourage the belief that the world is
deterministic. Philosophers are not in agreement concerning this belief. Some hold that
the world is indeterministic and others simply maintain an agnostic position. But no

4
matter what one’s belief is concerning the empirical world, as long as one believes that
the truth or falsity of determinism is relevant to the possibility of free will, one is
considered an incompatibilist. Otherwise, one is a compatibilist who holds that
determinism and free will are compatible. The compatibilist view is in harmony with
both our scientific intuition and everyday experience because on the one hand the
scientific education we receive cultivates the idea that every event can be scientifically
explained by appeal to its antecedent causes and on the other hand we believe that we
make free choices, act freely and hold each other morally responsible on a daily basis.
Both intuitions seem to be obviously true and hard to deny, but it does not follow that
justifying compatibilism is an easy task because the incompatibilist challenge, that is, the
truth of determinism would make free will impossible, has always been a serious threat.

2. The Analysis Strategy
The most important and popular strategy for defending compatibilism used by
both classical and contemporary compatibilists is the analysis strategy. By taking this
strategy, the compatibilist endeavors to offer an analytic definition that gives necessary
and sufficient conditions for the application of the concept of free will.
The definition of free will offered by the classical compatibilist such as Hobbes or
Hume is that an agent enjoys free will if and only if there are no impediments to doing
what he wants or chooses to do (Hobbes 1651, Hume 1888). Free will characterized as
such is thought to be compatible with determinism because even if determinism is true as
long as one is unimpeded to do what one wants to do, one has free will. However, there

5
are two major problems with this account. First, the analysis as such is not sufficient
because a person suffering obsessive-compulsive disorder may be free to do what he
wants to do, but if his desire is a product of his mental illness, then even if he is
unimpeded when acts on it, he does not act with free will. Second, it does not answer the
incompatibilist’s claim that if determinism is true we are always unfree because we are
never able to choose or act differently from the way we actually choose and act.
In order to solve the first problem of classical compatibilism, Harry Frankfurt
developed a more complex definition of free will – the Hierarchical View of free will
(Frankfurt 1971). According to him, having and acting with free will is not being able to
do what one wants to do, but being able to do what one wants oneself to do, that is, acting
according to one’s second-order desires rather than first-order desires. A variation of the
Hierarchical View is the Reason View defended by Gary Watson and Susan Wolf
(Watson 1975, Wolf 1990). The Reason View holds that an agent has and acts with free
will if and only if his actions either are (or he has the capacity to make them be) in
accordance with his values (or the Good). Both views are insightful because each one
correctly captures a correct application of the concept of free will. However, neither
theory is satisfactory. They do not do justice to free will because they reduce a wide
variety of the applications of freedom into a single one. The false assumption on which
all these theories are based is that compatibilism must be defended by giving an analytic
definition of what free will is. As Frankfurt’s Hierarchical View is the leading theory in
the field which is extremely important and highly influential, I shall closely examine it
and try to argue not only that his particular definition fails but also that the very project of

6
trying to defend compatibilism by way of an analysis in the form of analytic definition
fails.
Frankfurt distinguishes free action from free will. He argues that the definition of
free will given by the classical compatibilist explains free action rather than free will. He
uses an example of three drug addicts to defend his notion of free will. The unwilling
addict desires both to take the drug and refrain from taking it because he does not
approve his drug addiction. According to Frankfurt, the drug addict’s desire for taking
the drug is first-order and desire to stay away from it is second-order. If he acts
according to his second-order desire, he acts with free will.
Frankfurt’s idea that the unwilling addict’s choice of staying away from drugs
manifests his free agency is confirmed by our intuition. However, contrary to what
Frankfurt believes, this is not the correct application of the concept of free will but a
correct application of it. There are more than one types of examples to which we may
correctly apply the concept. For instance, the free will recognized by the classical
compatibilist and rejected by Frankfurt (through his free action vs. free will distinction) is
one and the free will identified by the Reason View is another. The analysis
compatibilist is forced by his own requirement and ideal to focus his attention on only
one of the legitimate applications of our concept of free will. However, there is no such
thing as sufficient and necessary conditions for the concept of free will, even though it
may be true that some other concepts can be defined this way. The compatibilist should
recognize all the correct uses and regard them as equally valid.

7
Frankfurt’s account also fails because it cannot meet the incompatibilist
challenge. He mistakenly believes that his definition of free will is immune to the danger
posed by the truth of determinism. In fact, with respect to the challenge, Frankfurt’s
theory fares no better than the classical compatibilism. It offers no resource to respond to
the incompatibilist objection that if determinism is true there would be no reason to think
that one could have desired otherwise at the second-order level. If an agent’s second-
order desire is determined and thus could not have been otherwise, he does not have free
will even by Frankfurt’s own standard.

3. The Incompatibilist Challenge
Constructing an analysis of the concept of free will cannot answer the
incompatibilist challenge by itself, because no matter what definition of free will the
compatibilist comes up with, the incompatibilist will argue that the truth of determinism
well endangers it. Unless the compatibilist directly faces the challenge, it is the
incompatibilist rather than the compatibilist who is winning the debate.
The incompatibilist challenge has always been on the frontline throughout the
history of the free will debate, but it did not take its contemporary form until it was
forcefully articulated by the contemporary incompatibilists through the Consequence
Argument, first developed by Carl Ginet and later strengthened by Peter van Inwagen
(Ginet 1966, van Inwagen 1983), and the Manipulation Argument, a family of arguments
constructed by Robert Kane, Peter van Inwagen and Derk Pereboom (Kane 1996, van
Inwagen 1983, Pereboom 2001).

8
In his important and influential book An Essay on Free Will, van Inwagen
characterizes the thesis of determinism as follows:
For every instant of time, there is a proposition that expresses the state
of the world at that instant;

If p and q are any propositions that express the state of the world at
some instants, then the conjunction of p with the laws of nature entails
q.
2

Built on this definition of determinism, van Inwagen has constructed his famous
Consequence Argument:
If determinism is true, then our acts are the consequences of the laws
of nature and events in the remote past. But it is not up to us what
went on before we were born, and neither is it up to us what the laws
of nature are. Therefore, the consequences of these things (including
our present acts) are not up to us.
3

The argument can be reconstructed as follows:
(1) If determinism is true, then our acts are the consequences of the laws of
nature and events in the remote past.
(2) It is not up to us what went on before we were born.
(3) It is not up to us what the laws of nature are.
(4) Therefore, the consequences of these things (including our present acts)
are not up to us.
The Consequence Argument is a formal representation of a philosophical worry
that if determinism is true, then no one ever chooses or acts freely and thus our common

2
Peter van Inwagen, An Essay on Free Will, p.65.

3
Ibid, p.56.

9
sense that we have and act with free will under normal circumstances is false. For
example, if determinism is true, my present act of signing a contract will be the
consequence of the laws of nature and events in the remote past. Since it is not up to me
what happened before I was born and what the laws of the nature are, the consequence of
these things, namely, my act of signing the contract, is not up to me.
Van Inwagen’s book has had an enormous impact on the debate between
compatibilists and incompatibilists. Incompatibilism is now a respectable philosophical
thesis. But the book has been almost entirely ignored by compatibilists.
The incompatibilist challenge also manifests itself in the Manipulation
Argument. Van Inwagen invites us to consider a possibility that we are manipulated by
Martians through a tiny device implanted in our brain but mistake ourselves to be free
agents. If this is the case, our firmly held common sense belief that we have free will
under normal circumstances will turn out to be an illusion. If determinism is true, we will
be in the same situation, that is, all the choices and actions which appear to be free are
nothing but disguised necessitated-events in reality.
More recently, Pereboom has offered a more ambitious manipulation argument –
the Four Case Argument used as a positive argument for incompatibilism. Pereboom
presents four cases in each of which Professor Plum murdered Ms. White in the manner
that we intuitively would not hold Plum responsible. If our intuition is justified,
determinism is not compatible with free will or moral responsibility.
Pereboom’s Four Case Manipulation Argument has received more attention from
compatibilists but their responses are inadequate. If we look only at the literature where

10
compatibilists and incompatibilists actually engage each other, incompatibilists are
wining.
Instead of directly facing the incompatibilist challenge, most compatibilists
choose to make peace with it by conceding that the truth of determinism may render
someone unable to do otherwise but he is still morally responsible even if he could not
have done otherwise. Frankfurt’s seminal paper “Alternate Possibilities and Moral
Responsibility” paved the way to this direction (Frankfurt 1969). The leading figures
traveling on this path are John Martin Fischer and Mark Ravizza who developed
Reasons-Responsive Compatibilism (Fischer and Ravizza 1998). They agree with the
incompatibilist that if determinism is true one could not have done otherwise, but as long
as one’s desires are sensitive to rational considerations one is a morally responsible agent.
This view is also closely affiliated with Frankfurt’s Hierarchical View. Although the
positive component of the theory has made contribution to the free will literature, its
decision to sidestep the incompatibilist challenge is rather disappointing.
As a compatibilist, I will take the incompatibilist challenge seriously and try to
show that if we examine the arguments in favor of incompatibilism closely, we will see
that they contain serious logical defects. For example, both the Martian Device
Argument and the Four Case Manipulation Argument appeal to the idea of global
manipulation, but they fail to distinguish a regular from special kind of it. I will argue
that a regular global manipulation is something like brainwashing or mind control while
special global manipulation is not, on closer inspection, a coherent concept.

11
4. The Paradigm Case Strategy
Although the analysis strategy for defending compatibilism is the most popular
choice, it is not the only choice. Another strand of contemporary compatibilism did not
descend from classical compatibilism but owes its inspirations to G. E. Moore and
Ludwig Wittgenstein (Moore 1959, Wittgenstein 1953). In their attempt to defeat
philosophical skepticism, some anti-skeptic philosophers developed the Paradigm Case
Argument and applied it to a wide range of skeptical positions, one of which is
incompatibilism in the free will debate. Antony Flew first used the Paradigm Case
Argument to explicitly defend compatibilism. He offered an example where a young
man married the woman he loved under no duress to show that free will is nothing but
paradigm cases of one’s making free choices and acting on them (Flew 1955). Peter
Strawson’s paper “Freedom and Resentment” is another example of applying the
Paradigm Case Argument to the free will problem (Strawson 1962). He argues that if we
take a closer look at our moral practice of holding each other responsible, we shall see
that determinism is irrelevant to what we consider as praiseworthy or blameworthy.
The paradigm case strategy is different from the analysis strategy. It does not aim
at finding the common properties of everything to which the concept of free will applies.
It views the Analysis Compatibilist’s goal of specifying conditions necessary and
sufficient for the application of free will as misguided.
A Paradigm Case Compatibilist holds that the concept of free will should be
explained by means of paradigmatic examples because the similarity that we find in a
group of examples where we correctly employ the concept of free will can be very

12
different from that in another group. In other words, with respect to the concept of free
will, there are no conditions for applying it that are both necessary and sufficient.
The Paradigm Case Argument can be easily misunderstood by both its opponents
and proponents. A most-frequently-made mistake is to regard it as an argument that
attempts to prove the existence of the instances of the concept in question (i.e. free will).
According to this understanding, the concept in question is instantiated because the only
way to learn the meaning of it is through looking at the instances of it. For example, the
compatibilist tries to meet the incompatibilist challenge by using the fact that we
understand the meaning of free will to show that it must be instantiated regardless
whether determinism is true or false.
Let us consider an example. If someone denies that red things exist, we may
refute him by pointing to a red apple and maintaining that the only way to learn the
meaning of “red” is to be shown red objects. Similarly, when the incompatibilist denies
that we have free will given the truth of determinism, we may refute him by pointing to
the example that the young man married his girlfriend of his own free will and arguing
that if no instances of free will exist, we would never learn the meaning of the
expressions of “acting freely” or “of one’s own free will”.
Most people understand the argument more or less along this line. This is
probably why the argument dropped out of the picture quickly after its heyday in the
1950’s. The problems with the argument understood as such are fatal. In the first place,
appealing to facts can never help refute skepticism because whether the facts are the way
they appear to be is exactly what is in question. The compatibilist will not be able to

13
state a fact in favor of his view without begging the question. In the second place, one
can learn the meaning of a word without having to look at an instance of it in the world.
For example, we understand what “unicorn” means without having to see an actual
unicorn in the world. These criticisms have convinced many that the Paradigm Case
Argument has been damaged and defeated beyond repair. However, it seems to me that it
is too quick to jump to the conclusion. I believe that a properly formulated Paradigm
Case Argument will not only render these objections inapplicable but also enable the
compatibilist to offer a plausible account of free will, meet the incompatibilist challenge
and expose the problems with incompatibilism.
I shall distinguish two versions of the Paradigm Case Argument. The defeated
version is PCA as Ostensive Teaching and the version I will defend is PCA as
Elucidation.
I propose that the Paradigm Case Argument should be understood and
constructed as an argument that explains the meaning of the concept of free will rather
than that proves the instances of it in the world because the nature of the incompatibilist
challenge is not concerned with the instance but the meaning of free will. So like the
Analysis Compatibilist, the Paradigm Case Compatibilist offers a conceptual analysis of
the concept of free will. But unlike the Analysis Compatibilist, he does not aim at
constructing necessary and sufficient conditions for the application of the concept
because he realizes that not all the correct applications of free will have something in
common. The goal of the Paradigm Case Compatibilism is to refute incompatibilism by
showing clearly in view that free will is a concept that we use to draw distinctions

14
between coerced and uncoerced, impeded and unimpeded, compelled and uncompelled
choices or actions. Anyone who denies such a distinction is to question the meaning of
free will rather than the existence of its instances.
While the Analysis Compatibilist tends to accept one of the correct uses of free
will and reject all the others, the Paradigm Case Compatibilist endorses them all in their
original forms and shapes without imposing on them any artificial requirement. The
Paradigm Case Compatibilist does not say that “Free will must be such-and-such based
on my theory” but rather “Free will is such-and-such based on my observation of our
concepts” because he believes that a concept is defined by what it does in the human
practice rather than what it is said to be in philosophical theories. For example, one may
construct a theory that defines a twenty-dollar bill to be equivalent to a fifty-dollar bill,
but it does not follow that the value of the twenty-dollar bill is raised to fifty dollars,
because as soon as the value of the twenty-dollar bill is explained in terms of its buying
power, its difference from the fifty-dollar bill will become crystal clear. Similarly, once
what the concept of free will can do in our practice is laid bare before us, we will see that
determinism is irrelevant to it and that the incompatibilist who vows not to use it given
the truth of determinism will not be able to live without smuggling it in from the back
door.
But do we really have a concept of free will? Doesn’t the fact that sometimes we
disagree on its applications prove that we don’t have such a concept? To have a concept
of free will, we do not have to reach an absolute agreement on every single case of
application. If our disagreement on what is beautiful does not mean that we do not have a

15
concept of beauty, our disagreement on when an agent is free should not mean that we do
not have a concept of free will. Moreover, the meaning of a concept can be explained in
various ways, one of which is to give examples where the concept is applied. If giving
necessary and sufficient conditions of a concept is the only way to explain it, then we can
never learn how to use the expression “acting freely” or teach our kids or foreigners to
use it because we have not figured out what those conditions are yet. Finally, if we do
not have a concept of free will, we would not understand what the subject matter of the
free will problem is or what compatibilists and incompatibilists debate about. If it is not
the free will that we find crucial in our moral, legal and religious practices is said to be in
danger, why should we care what impact determinism has on it?
Granted that we do have a concept of free will, if determinism is true, does it still
exist? The incompatibilist worry is that if the world is proved to be deterministic then the
paradigm cases of free will will turn out to be unfree. There is such a worry because it
has been assumed that determinism is relevant to the concept of free will. I shall try to
show that the alleged relevance is illusory.
Let us assume that determinism is discovered to be true and that the
incompatibilist request is granted. As a result, we never make free choices or act freely.
But are we still interested in drawing the distinctions between jumping into the pool and
being pushed into the pool or between a real epileptic seizure and a pretended one?
Given our human nature and way of living, we certainly are. But we cannot use the old
expression “acting freely” any more because none of these instances is a genuine free act.
In order to address the old distinction, a new word, say, “kree” must be created in order to

16
carry on the old task. With the help of the new word, we are able to say that the person
who pretends to have an epileptic seizure acts kreely and the person who has a real
seizure does not act kreely. It is not hard to see that “acting kreely” is equivalent to
“acting freely”. The old concept of free will is stuck with us as long as we live in the
same way as we used to before determinism was discovered to be true or as we would if
determinism were discovered to be false.
We have seen that, if determinism is true, the incompatibilist’s interest in drawing
the distinction between one’s jumping into the pool and being pushed into the pool has
defeated his refusal to apply the expression “acting freely” to the act of jumping and
proved that his claim that “acting freely” does not have an application in the world is
false.
Strawson’s view represents a Paradigm Case approach. He delineates the
conceptual connections among the concepts such as moral responsibility, gratitude,
resentment, excuse and exemption. He introduces the distinction between a participant
reactive attitude and an objective attitude. He believes that he has shown the
compatibility between determinism and moral responsibility because the acceptance of
the truth of determinism would not and should not lead to the decay or the repudiation of
participant reactive attitude.
Like any Paradigm Case Argument, Strawson’s account of freedom and moral
responsibility can be interpreted in two ways. One reading is that he describes instances
or applications of the expressions of “acting freely”, “holding responsible” and “applying
excuses or exemptions”. Interpreted as such, Strawson’s view is susceptible to an easy

17
attack. For example, one may point out that our everyday reactive attitudes can be
appropriate or inappropriate, fair or unfair, depending on what the person in question
actually does. If the instances (as opposed to the concept) of reactive attitudes are
constitutive of free will and moral responsibility, an observer’s resentment may well
convict an innocent person.
As Strawson clearly distinguishes an internal justification from an external one, it
seems to me that the first reading overlooks this crucial distinction. The distinction is
originally from later Wittgenstein who believed that a justification is possible only within
a practice and the practice itself cannot be justified or refuted from an external point of
view. As far as our moral practice of holding responsible is concerned, praise or blame
can be justified only internally. For example, Viggo is resented by his peers because they
believe he is a murderer when in fact he is innocent. Obviously, having the reactive
attitude of resentment towards Viggo is inappropriate and unfair. A particular instance of
holding someone responsible by having reactive attitudes can be correct or incorrect but
there is nothing correct or incorrect about holding responsible or showing reactive
attitudes as a whole because we can neither justify nor refute a practice from a point of
view outside the practice.
The incompatibilist, however, demands such a justification. His challenge is that
if we cannot justify the distinction between the cases where we hold responsible and the
cases where we apply excuses or exemptions, the distinction is arbitrary and therefore
should be allowed to move to somewhere else, namely, between determinism and

18
indeterminism.
4
I agree with Strawson that a practice itself as well as the concepts that
the practice makes possible can neither be justified nor refuted.
By taking this idea seriously, I believe that the first reading does not do justice to
Strawson’s view. A more charitable way to read him is to regard the examples he gives
as a statement or elucidation of internal connections among the concepts of gratitude,
resentment, responsibility, excuse and exemption, which in turn are all closely related to
the concept of free will. In other words, he is not offering an empirical report of our daily
moral practice but painting a logical geography of all the concepts surrounding freedom
of will. If this is correct, the examples are normative rather than descriptive because they
serve as explanations of meaning of the concept in question and thus setting the standard
of correctness.
Strawson’s way of defending compatibilism seems to me correct, but his response
to incompatibilism is inadequate. Like Frankfurt, he is dismissive about the
incompatibilist challenge and does not adequately explore what goes wrong with
incompatibilism.

5. Outline
In this thesis I will defend my view by first examining two representative
examples of the two main kinds of compatibilist accounts – Frankfurt’s higher order
desire account of acting with free will and Strawson’s reactive attitudes account of moral
responsibility. I will argue that these accounts fail as defenses of compatibilism

4
Of course, incompatibilists are not aware of this fact. They think they have discovered the true
distinction, that is, the distinction between determinism and indeterminism.

19
regardless of their other merits. I will then examine versions of the manipulation
argument for incompatibilism and show that different versions of this argument all fail. I
will argue that the compatibilist who defends compatibilism only by way of the analytic
definition approach cannot effectively criticize these arguments. Finally, in my last
chapter, I will sketch an alternative way of defending compatibilism – the Paradigm Case
Argument. I will explain why the Paradigm Case defense is not vulnerable to the
criticisms that are viewed as fatal to the traditional version of the Paradigm Case
Argument.
In short, I shall offer a conceptual analysis (as opposed to instances) of the
concept of free will through paradigm cases (as opposed to analytic definitions) and
argue that the incompatibilist denial of this analysis necessarily involves logical defects.

20

Chapter Two: Frankfurt’s Compatibilism

1. Classical Compatibilism
Frankfurt’s Hierarchical View is one of the most influential contemporary views
and a representative theory that takes the analysis strategy to defend compatibilism
(Frankfurt 1971). It purports to offer an analytic definition that gives necessary and
sufficient conditions for the application of the concept of free will.
Frankfurt’s theory is a descendent of Classical Compatibilism developed by
empiricists Thomas Hobbs and David Hume in the modern era. Both Hobbs and Hume
held that free will and moral responsibility are compatible with causation and
determinism because liberty (or, freedom) is nothing but one’s being able to do what one
desires or inclines to do.
5

According to Hobbs, “freedom” is “the absence of opposition”
6
, namely, the
impediments or constraints of one’s movements. For instance, a free man is free while a
prisoner is not; a healthy person is free while a patient confined to bed by illness is not.
Hobbs believes that freedom and determinism are compatible. He argues,
Liberty and necessity are consistent: as in the water that hath not only
liberty, but a necessity of descending by the channel; so likewise in the
actions which men voluntarily do, which, because they proceed their
will, proceed from liberty, and yet because every act of man’s will and
every desire and inclination proceedeth from some cause, and that

5
Hobbes and Hume used “necessity” to mean what we mean by “determinism”.

6
Thomas Hobbes, Leviathan, p.145.

21
from another cause, in a continual chain (whose first link is in the hand
of God, the first of all causes), proceed from necessity.
7

Just like flowing water has its causes, a person’s free action has antecedent causes
too. It is not inconsistent to say that a voluntary action is both free and caused.
Hume defended a similar view. He held that if a person is morally responsible for
a voluntary action, he must be the cause of this action. So free will or moral
responsibility is not only compatible with causation but also requires it. He distinguishes
two kinds of liberty,
Few are capable of distinguishing betwixt the liberty of spontaneity, as
it is call’d in the schools, and the liberty of indifference; betwixt that
which is oppos’d to violence, and that which means a negation of
necessity and causes. The first is even the most common sense of the
word; and as ‘tis only that species of liberty, which it concerns us to
preserve, our thoughts have been principally turn’d towards it, and
have almost universally confounded it with the other.
8

‘Tis only upon the principles of necessity, that a person acquires any
merit or demerit from his actions, however the common opinion may
incline to the contrary.
9

The liberty of spontaneity is the freedom opposed to violence or constraint while
the liberty of indifference is the freedom opposed to causation or necessity. What is
valuable and important to us is the liberty of spontaneity rather than the liberty of
indifference. Since freedom is the absence of a certain kind of causes rather than
causation itself, determinism does not undermine free will or moral responsibility. Those

7
Ibid., pp.146-147.

8
David Hume, A Treatise of Human Nature, pp.407-408.

9
Ibid., p.411.

22
who think that free will is incompatible with determinism fail to discern the differences
between the two kinds of liberty.
According to the Classical Compatibilist’s analysis of freedom of will, freedom is
not contrasted with causation or necessity but with impediment, constraint, or coercion.
If a person is able to do what he wants to do, the action is a free or voluntary action and
the person is morally responsible for performing it. However, one problem that has been
identified with this theory is that this definition is not sufficient. If someone’s desire
results from a compulsive-obsessive disorder, addiction, or other pathological condition
even if he is able to do what he desires to do, he still does not have free will. In order to
solve this problem, Harry Frankfurt proposed a more complex compatibilist account that
has come to be known as “Hierarchical Compatibilism”.

2. Frankfurt’s Compatibilism
In his article “Freedom of the Will and the Concept of a Person” (Frankfurt 1971),
Frankfurt constructs a more complex analytic definition of free will. He argues that the
word “person” does not merely connote membership in the species of Homo sapiens, but
is designed to “capture those attributes which are the subject of our most humane concern
with ourselves and the source of what we regard as most important and most
problematical in our lives.”
10
According to Frankfurt, these attributes may be absent in
some human species but found in some nonhuman species.

10
Harry Frankfurt, “Freedom of the Will and the Concept of a Person”, in Free Will: Critical Concepts in
Philosophy, edited by John Martin Fischer, Vol. IV, p.8.

23
The characteristics that differentiate persons from non-persons are the persons’
“capacity for reflective self-evaluation that is manifested in the formation of second-order
desires.”
11
Some desires are of the second order because they are desires based on the
reflective evaluation of the first-order desires. For example, a person on diet may have
conflicting first-order desires – a desire to eat the cheesecake and a desire to stay away
from it. If he wants one of these desires to be effective, then he has formed a second-
order desire.
Frankfurt distinguishes a special kind of second-order desire from the rest of
second-order desires. If someone not only wants to have a certain desire to be among his
first-order desires but also wants it to be effective, his second-order desire is a second-
order volition. One has a second-order volition only if one wants a certain desire to be
his will, namely, an effective desire. A desire may be effective even if the person does
not succeed in acting on the desire. For example, a person’s desire to lift his arm is
effective even when he is struggling to lift it while a strong man holds it down.
One may be curious about what it is like for someone to feel so depressed that he
committed suicide. One wants to experience the desire to commit suicide but does not
want this desire to be effective. One’s desire to have the desire to commit suicide is a
second-order desire, but not a second-order volition. But if one not only wants to acquire
the desire to commit suicide but also wants it to be effective, then one has a second-order
volition.

11
Ibid.

24
Frankfurt believes that “it is having second-order volitions, and not having
second-order desires generally” that characterizes a “person” and differentiates persons
from wantons.
12

He offers an example of two drug addicts to illustrate the distinction between a
person and a wanton. One addict is the unwilling addict who fights desperately against
his addiction and tries everything that may help him quit drugs even though his desire for
drugs is too powerful to resist. The other is the wanton addict who does not have a
second-order volition because he does not care about the fact that he is a drug addict. His
lack of concern is “due either to his lack of the capacity for reflection or to his mindless
indifference to the enterprise of evaluating his own desires and motives.”
13

On Frankfurt’s account, only persons are capable of either enjoying or lacking
freedom of will because wantons do not have second-order volitions and thus “freedom
of his will cannot be a problem for him.”
14
A person enjoys free will only if his will is in
accordance with his second-order volition, so the unwilling addict’s will is not free if he
takes the drug, but if he succeeds in one of his struggles and refrains from taking the
drug, he enjoys free will by doing so.
Frankfurt invites us to consider yet another addict, the willing addict, who
approves his own addiction and “would not have things any other way”
15
. He tries to do
everything he can to strengthen his desire for drugs. When the willing addict takes the

12
Ibid., p.11.

13
Ibid., p.13.

14
Ibid., p.15.

15
Ibid., p.18.

25
drug, his action is in conformity with his second-order volitions. However, this fact
alone does not qualify his will as free because freedom of will requires an additional
condition.
A person’s will is free only if he is free to have the will he wants. This
means that, with regard to any of his first-order desires, he is free
either to make that desire his will or to make some other first-order
desire his will instead. Whatever his will, then, the will of the person
whose will is free could have been otherwise; he could have done
otherwise than to constitute his will as he did.
16

Acting in accordance with one’s second-order volition – acting of one’s own free
will – is not sufficient for having free will. Another necessary condition is that one’s will
is under one’s control, that is, one could have made a different first-order desire one’s
will had one had a different second-order volition. For example, someone has four first-
order desires A, B, C and D. He is moved to act by his second-order desire to make D his
will. If he has free will, he could have willed otherwise, that is, he could have made A,
B, or C his will instead. A person’s will is free if and only if he wills (effectively desires)
in accordance with his second-order volition and he could have willed otherwise. By
this, Frankfurt means that the person’s first-order motivation – his will – is under the
control of his second-order motivation. If the person had had a different second-order
volition, he would have willed differently.
This definition of free will is different from the classical compatibilist definition
of free will. According to the classical compatibilism, a person’s will is free if and only
if there are no impediments that prevent him from successfully doing what he wants or
chooses to do. Frankfurt believes that his definition is incorrect because the classical

16
Ibid.

26
compatibilist conflates freedom of action and freedom of will. Freedom of action is the
freedom to do what one wants to do and freedom of will is the freedom to will
(effectively want) what one wants oneself to will and could have willed otherwise.
Frankfurt also distinguishes freedom of will from acting of one’s own free will.
This distinction is relevant to moral responsibility.
For the assumption that a person is morally responsible for what he
has done does not entail that the person was in a position to have
whatever will he wanted.
17

This assumption does entail that the person did what he did freely, or
that he did it of his own free will.
18

If one is held morally responsible for what he has done, he does not have to have
free will, but he must have done it of his own free will. Frankfurt believes that the
“relation between moral responsibility and the freedom of the will has been very widely
misunderstood”.
19
He argues,
It is generally supposed that … a satisfactory theory of the freedom of
the will necessarily provides an analysis of one of the conditions of
moral responsibility. The most common recent approach to the
problem of understanding the freedom of the will has been, indeed, to
inquire what is entailed by the assumption that someone is morally
responsible for what he has done. In my view, however, the relation
between moral responsibility and the freedom of the will has been
widely misunderstood. It is not true that a person is morally
responsible for what he has done only if his will was free when he did
it. He may be morally responsible for having done it even though his
will was not free at all.
20

17
Ibid.

18
Ibid.

19
Ibid.

20
Ibid.

27
Frankfurt’s theory breaks up the close relationship between free will and moral
responsibility. According to him, one can be held morally responsible for performing an
action even when one’s will is not free. Frankfurt says that the unwilling addict is not
morally responsible for taking the drug because he can meaningfully and truthfully say
that the desire that moves him to act is not his will, that his desire for the drug forces or
compels his action, that he acts against or despite his will. But since the willing addict
acts on the desire that he wants himself to act on, he cannot truthfully say these things
and, according to Frankfurt, he is morally responsible for taking the drug. The wanton
addict is not responsible for his actions because he does not have second-order volitions
in the first place. In sum, one is morally responsible for performing an action if and only
if one’s action is in accordance with one’s second-order volition.
Frankfurt believes that this definition of free will is compatible with determinism
because determinism is not relevant either to the question of whether persons have
second-order volitions or to the question of whether they succeed in conforming their
actions to their second-order volitions.

3. Frankfurt’s Theory of Free Will and Moral Responsibility
According to the classical compatibilists, if one is able to move as one wishes or
do what one desires to do without impediments, one has free will and is morally
responsible for performing the actions. As Frankfurt has rightly pointed out, it is widely
held that moral responsibility is closely related to free will. By defending a Hierarchical

28
View of Compatibilism, Frankfurt has broken up this connection which is something, I
shall argue, not worth doing.
I will apply Frankfurt’s theory of free will and moral responsibility to an example.
Cara, Aisha, and Bea are roommates and taking the same math class. They all find
working on math problems difficult and boring and wish they could play video games
instead. Although they are often miserably torn by their conflicting desires of studying
math and playing games, they remain in the class because they need the class to graduate.
Cara is aware of her inner struggle but does not care which desire gets the upper
hand. Sometimes, she fails to complete her homework because her desire to play video
games is so strong that she cannot resist it. Other times, when her fear for failing the
class outweighs her desire to play, she turns her homework in.
Aisha and Bea are not as carefree as Cara. They both care about which desire will
be effective. Aisha would like to see herself studying math as much as possible and not
wasting time playing video games. In contrast, Bea has a different preference. She
understands that she needs to pass the math class in order to graduate but she does not
want to sacrifice too much for achieving that goal. Sometimes, she is overtaken by the
fear for failing the class and will study, but whenever this happens, she does not respect
herself as much. She wants to see herself staying steady and playing video games as
much as possible.
By Frankfurt’s standard, Cara is a “wanton” because she does not have a second-
order volition. She lacks free will by default and is not responsible for her actions
regardless whether it is studying math or playing video games.

29
Both Aisha and Bea are “persons” because they have second-order volitions.
Only persons can be free or unfree and only persons can be held morally responsible.
Aisha’s second-order volition is to make her desire for studying math her will. If she is
successful, she enjoys free will because she is able to do what she wants herself to do.
She is certainly morally responsible for such an action and should be praised for her
achievement. However, as it often happens, she plays rather than studies because she
often finds her desire to play irresistible. It is like losing a war. She feels that she is
helplessly violated by a powerful external force. To apply Frankfurt’s theory, Aisha does
not have free will and is not responsible for playing video games.
Bea’s second-order volition is to make her desire for playing video games an
effective desire. She is proud of herself whenever she sees Cara and Aisha studying and
herself playing. She regards her own action as rebellious and heroic. According to
Frankfurt, Bea is morally responsible for her action but does not enjoy free will. She is
responsible because her action is in accordance with her second-order volition and she
lacks free will because her will is not under her control. Even if she had wanted herself
to study math instead (that is, even if her second-order volition had been different), she
would still have played video games.
Occasionally, Bea’s fears become overwhelming and she studies too. But when
this happens, she feels being forced by an external force and acting against her will.
Frankfurt would say Bea neither has free will nor is responsible for studying.
Now let us examine the applications of Frankfurt’s theory to Cara, Aisha and Bea
cases. Suppose that they all failed the math class as a result of unable to study hard

30
enough to earn a passing grade. Aren’t they responsible for their failure? Don’t they
deserve the grades they have earned? It is true that Cara was a wanton and did not care
which desire of hers would get the upper hand, but few would say she should be excused
because of that.
Unlike Cara, Aisha did care, but her will power was not strong enough to deliver
a success. Isn’t she responsible for her poor performance in the math class? Our
common sense is that she certainly is even though she felt that her desire to play was
external to her or forced upon her.
Frankfurt’s verdicts about the cases where Aisha refrained from playing video
games and where Bea played the games are in agreement with our intuition because
Frankfurt also thinks that they should be held responsible.
One may try to defend Frankfurt’s theory by pointing out that second-order
volitions do not have to be formed explicitly and consciously on each and every occasion
for Frankfurt also states that “the conformity of a person’s will to his higher-order
volitions may be far more thoughtless and spontaneous than” what happens to the
unwilling addict.
21
It is possible that Cara may not deliberately form a second-order
volition but she nevertheless has one implicitly. If someone asks her which desire
between the two she prefers to be effective and she answers “play”, then she has an
implicit second-order volition. So Cara should be held responsible for earning an “F”.
This move aligns Frankfurt’s theory with our common sense, but it cannot always be the
case. Otherwise, no normal human adults will ever be wantons and Frankfurt’s
distinction between “person” and “wanton” will be pointless. Thus, it must also be

21
Ibid., p.16.

31
possible that Cara answers, “I don’t have a preference because I don’t care if my desires
are desirable.” But this claim certainly does not and should not excuse her for her
responsibility.
With regard to Aisha’s case, one may reply that Aisha should be held responsible
for failing the class because she suffered weakness of the will rather than a physiological
addiction. It is true that her desire to play video games is not a pathological one as it is
the case with the unwilling addict’s desire to take the drug, but if Frankfurt’s theory is
only applicable to pathological cases then it cannot account for the concepts of free will
and moral responsibility that has concerned us in the free will debate. In order to
preserve the generality of Frankfurt’s analysis of free will and moral responsibility, Aisha
should not be regarded as responsible for her actions of playing video games. But this
idea obviously flies in the face of our common sense.
I hope that Cara and Aisha cases have shown that Frankfurt’s theory of moral
responsibility is counterintuitive. If we go along with our intuition and believe that Cara
and Aisha are responsible, then Frankfurt’s theory of free will will not sound as useful as
it initially does. If all normal human adults under normal circumstances are responsible
for what they choose or want to do, what is the point of re-defining free will in terms of
second-order volitions?
Frankfurt’s notion of free will seems to me equally problematic. When he insists
that a wanton, such as Cara, does not have free will when she acts because she does not
have second-order volition, he is not describing or reporting an empirical fact about
Cara’s will, but giving a new definition to the phrase “free will”. He defines it in terms

32
of second-order volition and by this definition Cara does not have free will. But this
move does not help shed light on our common sense that Cara deserves receiving an “F”
because she had the free will to choose between studying and playing, could have studied
more if she chose to, but unfortunately spent too much time playing. Frankfurt’s theory
does not help us understand what we mean by saying that Cara had the choices or could
have done better. Instead, it denies that she had the freedom of will. So it seems clear
that he is not discussing the same freedom of the will that has concerned us here. Rather,
he is interested in a different notion of free will, which is a much narrower one than the
one we usually find important.
The same problem can be found with Aisha and Bea cases. When Aisha plays,
she hates herself and wishes that she is studying instead. But her self-hatred and guilty
feelings do not make her action or will unfree. We are all familiar with the syndromes of
weakness of the will. When the alarm clock goes off, we push the snooze button; when
we are on a diet, we pick up another piece of cake; when we make a plan to go to the
gym, we find excuses to stay at home. We all hate our bad choices, which go against our
reasons, values, and even self-esteem, but few will say that our free will is taken away
from us when we suffer the weakness of the will or think that we are not responsible for
doing what we do.
According to Frankfurt, Bea does not have free will when she plays video games
because her effective desire could not have been otherwise as it would be much harder to
make herself study rather than play. This observation is certainly true. Some people tend
to do whatever they find easy to do, whether it is to continue to be an alcoholic or an

33
irresponsible spouse. But when they act in accordance with their preferences or second-
order volitions, their free will is not taken away from them. They are not only choosing
and acting freely but also responsible for what they do.
Frankfurt conceives free will as an achievement or honor which is not to be
granted to anyone but only to those who not only care about what kind of life they live
but also succeed in living the kind of life they want to live. Inspiring as it is, it is not the
same concept of free will that we find in the free will problem.
In fact, Frankfurt fails to provide a condition that is both necessary and sufficient
for the application of “free will” when understood in an ordinary way, because his
analysis is neither necessary nor sufficient. It is not necessary because some clear cases
of free will do not involve higher-order desires. Let us consider the following
conversation between Dale and Ed:
D: Did you work overtime last weekend?
E: Yes, I did.
D: Did your boss force you to? I know he is mean.
E: No, not this time. I stayed at work of my own free will because I wanted to
finish the project as soon as possible.
Ed’s working overtime is not only a free action but also a manifestation of his free
will. However, Frankfurt would have to deny this because Ed doesn’t form any second-
order volition.
It may be contended that the relevant second-order desires, like the relevant first-
order desires, are just preferences. If Ed preferred working overtime, then he had a

34
second-order volition which does not have to be formed explicitly since Frankfurt holds
that “When a person acts, the desire by which he is moved is either the will he wants or a
will he wants to be without.”
22

This may be true with a person, but by Frankfurt’s standard, Ed is a wanton. Ed is
a rational creature and capable of reasoning. He had a desire to finish his project last
weekend, so he took the necessary means to achieve his goal, which is working overtime.
He is not qualified as a person because “he is not concerned with the desirability of his
desires”. Frankfurt writes,
Nothing in the concept of a wanton implies that he cannot reason or
that he cannot deliberate concerning how to do what he wants to do.
What distinguishes the rational wanton from other rational agents is
that he is not concerned with the desirability of his desires
themselves.
23

Frankfurt’s notion of free will is not sufficient because one’s second-order
volition may be a result of manipulation or coercion. For example, Farah asks Gina:
F: Do you work overtime every weekend?
G: Yes, I do.
F: Do you want to do it?
G: Yes, I do. If I don’t, I will suffer more. I will not only feel sorry for myself
but also get myself fired because my boss hates to see us showing our unhappiness at
work.

22
Ibid., p.14.

23
Ibid., p.12.

35
Gina’s second-order volition is to make one of her first-order desires, working
happily, to be her will. But in this case not only Gina’s labor is forced; her second-order
volition is forced as well. If first-order desire can be forced, coerced, or manipulated,
second-order desires are susceptible to the same thing. Appealing to the second-order
does not help avoid these problems. Thus it is not always true that acting in conformity
to one’s second-order volition ensures that one enjoys freedom of will.
Frankfurt seems to have trouble justifying his definition of free will against
objections. One of the famous criticisms of Frankfurt’s notion of free will is from Gary
Watson (Watson 1975). He argues that it is unclear why the second-order volition is
given the authority of agency as it can be necessitated or compelled the same way that the
first-order desire is necessitated or compelled. Moreover, if the higher order of a desire
is, the more authority it has, there will be an infinite ascent of desires. In place of
Frankfurt’s definition of freedom, Watson offers his own. According to him, what makes
the unwilling addict free when he refrains from taking drugs is not his acting in
accordance with his second-order volition but with his evaluation system. But why is it
so? Watson’s account fares no better than Frankfurt’s in terms of finding justifications.
If someone thinks that the concept of free will should not be applied to the unwilling
addict at all even when he refrains from taking drugs, then whatever we find in the
unwilling addict’s psyche or evaluation system would be useless for the efforts of
refuting this person.

36
Frankfurt himself tries to justify his definition of free will. The first attempt he
makes is to differentiate internal desires from external desires.
24
He hopes to show that
second-order desires have the authority of one’s agency because they are internal to one.
There are two questions he needs to answer: (1) How do we know they are internal to one
and (2) why does acting on one’s internal desires make one’s will free?
To answer the first question, Frankfurt appeals to the distinction between internal
and external bodily movements. Internal bodily movements are, for example, the
movements one makes when one is pretending to have an epileptic seizure and external
bodily movements are the movements one makes when one is having a real epileptic
seizure. Similarly, internal desires are those one feels comfortable with if not proud of,
such as one’s desire to remain drug-free; external desires are those desires one is fighting
against, such as one’s desires for drugs.
Frankfurt’s example of internal and external bodily movements is supposed to
help us recognize internal and external desires. But it can hardly do so. The
characteristics of “internal” and “external” do not refer to physical locations but purely
metaphorical as we don’t actually see the principle of a desire is inside or outside a
person. The analogy between mental movements and bodily movements doesn’t help
either because it is unclear why we cannot say that a real epileptic seizure is internal to
one while the pretended seizure is external to one. At least, the real seizure, or the
principle of the bodily movements of a real seizure, is in the body rather than attached to
it by pretense.

24
See Frankfurt, “Identification and Externality,” in The Importance of What We Care About, 1988.

37
We understand that the movements of a real seizure are not up to the patient while
the movements of a pretended seizure are up to the person who pretends it. However,
achieving such understanding is not because we find out if the movements are internal or
external to the agent first and then decide if we attribute the movements to him. Rather,
our understanding of the differences between a real seizure and a pretended seizure
enables us to recognize the internality or the externality of the movements.
So the internality or externality of the principle of a bodily movement cannot be
determined independently from the fact that second-order desires are considered as
internal while the first-order desires external. In other words, the distinction between
internal and external desire cannot justify Frankfurt’s claim that second-order desires
have the authority of agency. Those agreeing with Frankfurt will not need the external
vs. internal explanation and those disagreeing with him will find it useless.
Another attempt Frankfurt makes to justify his definition of free will is to argue
that an agent’s second-order desires have the authority of representing him because he is
completely satisfied with it, will not regret about acting on it, or approve it
wholeheartedly. All these moves fail because one’s experience of self-satisfaction can be
temporary, wrong, or self-deceptive. It is not hard to think of an example where someone
is wholeheartedly satisfied with pursing self-destructive goals, such as the willing addict.
Even if certain self-satisfaction is neither wrong nor self-deceptive, we may still wonder
why it indicates free agency. Obviously, it is in need of justification the same way the
second-order desires are.

38
Frankfurt has successfully discovered the conceptual connections between free
will and other expressions such as “second-order desire”, “internal”, “external”, “self-
satisfaction” and “wholeheartedness”, but he is not aware that none of these expressions
has a firmer ground than others. In other words, they are all at the same level and no one
can justify any other. We may draw a conceptual map about them by elucidating their
connections, but it is wrong to think that some of them are more fundamental than others.

4. Frankfurt’s Answer to the Incompatibilist Challenge
Frankfurt believes that his definition of free will is immune to the threat of
determinism because so long as an agent effectively wants what he wants himself to
want, his will is free regardless of whether it is determined by an antecedent cause or
occurs randomly. He claims that
My conception of the freedom of the will appears to be neutral with
regard to the problem of determinism. It seems conceivable that it
should be causally determined that a person is free to want what he
wants to want. If this is conceivable, then it might be causally
determined that a person enjoys free will.
25

On the other hand, it seems conceivable that it should come about by
chance that a person is free to have the will he wants. If this is
conceivable, then it might be a matter of chance that certain people
enjoy freedom of the will and that certain others do not.
26

Frankfurt believes that neither the truth nor the falsity of determinism is relevant
to his definition of freedom, but this is not true. On Frankfurt’s account, the willing
addict does not have free will because given the overdetermination of his first-order

25
Frankfurt, “Freedom of the Will and the Concept of a Person”, p.19.

26
Ibid.

39
desire to take the drug, his will (effective desire) could not have been otherwise. In other
words, even if the willing addict had wanted a different first-order desire to be his will,
his desire to take the drug would still have triumphed. If so, how does he argue that the
truth of determinism is irrelevant? On the one hand, if the truth of determinism means
the “overdetermination” of the unwilling addict’s effective desire, then the unwilling
addict does not have free will even when he refrains from taking the drug. On the other
hand, if the effectiveness of a certain desire is not under the agent’s control but a result of
chance, as the falsity of determinism would imply, a person is not free “to have the will
he wants”
27
.
It is misleading to say, as Frankfurt does, that it may be a causally determined fact
or a matter of chance that certain people enjoy free will and certain people are not as
lucky. It is certainly true that some people are lucky and others are not no matter what
makes it so. But this is not what the threat of determinism is about. Rather, determinism
is about if there is still such thing as “luck” given the truth of determinism. If
determinism is true, no effective desires could have been otherwise and thus free will is
never possible. So it seems rather clear to me that Frankfurt fails to answer the
incompatibilist challenge.

5. Free Will as Mental Occurrence
Frankfurt believes that having freedom of action, namely, being able to do what
one wants to do, is not a necessary condition for having free will because the question of

27
Ibid.

40
one’s will concerns nothing but one’s desires and “to deprive someone of his freedom of
action is not necessarily to undermine the freedom of his will”.
28

For example, if Hank refused to accept the fact that his legs were paralyzed and
often tried to walk with them as if they were healthy, Frankfurt should say Hank enjoyed
freedom of will because
Even though he is no longer free to do what he wants to do, his will
may remain as free as it was before. Despite the fact that he is not free
to translate his desires into actions or to act according to the
determinations of his will, he may still form those desires and make
those determinations as freely as if his freedom of action had not been
impaired.
29

The underlying picture of this idea is that when someone acts with free will his
bodily movements are caused by his mental occurrences. Free will is to be located in the
mind rather than the body. The loss of the ability to move one’s body does not affect
what happens in the mind, so when one acts freely it is the mind that determines whether
the agent is acting with free will. I shall examine this idea closely and try to argue that
the root of the mistakes found in Frankfurt’s account, shared by both compatibilists and
incompatibilists alike, is nothing but to see volition as a mental occurrence.

(1) Mental State
Now let us look at the first possibility of regarding free will as a mental
occurrence closely, that is, to conceive free will as a mental state which an agent is in
when he acts freely. We may call the mental occurrence of free will “S state”. For

28
Ibid., p.14.

29
Ibid., p.15.

41
example, when I jump into the swimming pool to swim, I am in the state of S at the
moment of jump. But if I fall into the pool due to another person’s push, my mind is not
in the S state. On this view, free will is a mental state which occurs only when the agent
acts freely and it is its occurrence that qualifies the act as a free act. In other words,
being in the mental state S is a condition both necessary and sufficient for acting with
free will. Now let us investigate what S can possibly be.
Does S consist of the thoughts and ideas that the agent has in mind when he acts
with free will? At the moment when I jump into the pool, I may think that “I hope the
water is not too cold” or “I can finally swim after a long day”. But do these thoughts
render the action free? They certainly don’t because I can easily think of them when I sit
in my office or even when I am pushed into the pool.
It is possible that the mental state that makes the action free is not an idea or
thought which the agent is aware of, but a physiological process hidden from our view
which can be detected by scientific measures only. But this thesis overlooks the logical
priority of the concept of free will over the physiological process.
Suppose that I am perplexed by the concept of kindness and hope to understand it
better. A neuroscientist tries to help and tells me that kindness denotes a mental state K.
Whenever K occurs, the agent acts kindly and if the agent acts kindly, K must occur.
Does this information help? It certainly does not. What scientists discover is nothing but
a correlation between kindness and the mental state K. Scientists form hypotheses and
try to discover as many empirical connections as possible. If their investigation finds that
all kind behaviors correlate with lower blood pressure or the light up of certain areas of

42
the brain, then we are informed with one more piece of empirical knowledge. But when
they discover this, they must have taken the concept of kindness for granted. Even if the
correlation can be established by experiments, K cannot help us understand what
kindness is because it has already presupposed the concept.
Let us consider another example. A lie detector can be used to detect lies because
of the established correlations between lies and the liar’s physiological states such as the
change in heart rate. However, the detector cannot help us understand what dishonesty is
because the possibility of establishing the correlation presupposes the concept of
dishonesty. The reason why lie detector is useful to us is exactly because the
physiological evidence accompanying dishonesty detected by the polygraph matches our
existing concepts of honesty and dishonesty in most cases. Similarly, if the mental state
S is correlated with acting with free will, the correlation cannot help explain the meaning
of free will because it has presupposed it.
Some philosophers believe that what makes an action free is the first person
experience of acting with free will. If I scrutinize my experience at the first moment of
jumping into the pool, I may discover that I am actively making the effort to initiate the
jump. But when I am pushed, I do not have such experience. So it must be the presence
or lack of the exertion of power that makes difference in these two cases. Some
psychologists believe that acting with free will involves kinaesthetic and innervative
sensations.
30
Is the exertion of power the freedom of will? Does the kinaesthetic and
innervative sensation make an action free?

30
William James, The Principles of Psychology, in the series of The Britannica Great Books Of The
Western World, Vol. 53, pp.767-835.

43
It is not true that we always have such experiences or sensations when we act
freely. For example, when we say that Isaac married Jada of his own free will, we do not
mean that he had the experience of exerting power or that he had kinaesthetic and
innervative sensations when he got married. Instead, we mean that Isaac was in love with
Jada and married her under no coercion. If Isaac and Jada’s marriage was forced by
gunpoint, even if Isaac did have the experience of exerting power, we would not say that
he married Jada of his free will. If this is right, we may conclude that the first person
experiences of power exertion or having certain sensations are neither necessary nor
sufficient for an action to be free.
One may also find it tempting to conceive free will as a mental state where the
volition is free, namely, not being forced or compelled. In believing so, one has
mistakenly drawn an analogy between volition and a person. A person is free if he is
neither constrained nor compelled but volition cannot be free on its own. Although we
talk about free will, free action, or free decision, what we mean is not that volition, action
or decision is free or unfree by itself, but rather the agent is free or unfree when he
performs an action or makes a decision. Moreover, if volition is a mental state, how does
it initiate anything without an antecedent cause? If there is an antecedent cause, then it is
necessitated by the cause and turns out to be unfree.

(2) Mental Act
Now it seems that we are only left with one possibility when assuming that free
will is a mental occurrence, that is, volition is not a mental state that the agent finds
himself in but a mental act that the agent performs.

44
Some people conceive the volition as a link in a causal chain leading to an action.
It is free if it situates at the beginning of the chain and unfree if it has antecedent causes.
Let us consider two events – I raise my right arm and my right arm is raised by another
person. Both events can be seen as bodily movements brought about by antecedent
causes. If the first cause is my will, it is free; otherwise, it is unfree. But if we can say
that the act of willing is free because it is not caused by anything other than the agent,
then why don’t we simply say that my action of raising my right arm is free? Why do we
have to push the explanation one step back? One may also be tempted to think that
volition is not an act but a small agent sitting in one’s mind who can initiate actions, but
if so the original problem has been duplicated and an infinite regress awaits ahead.
Even though “to will” is a verb, it does not signify an action. However, it doesn’t
follow from this that it must signify a mental state or an experience because it is neither a
mental act nor a mental state.

(3) Trying
One may point out that the person who has lost freedom of action can still enjoy
free will as long as he is trying to execute his determination. This is another theory of
free will which attempts to pin down the concept of free will by focusing on trying and
making efforts. The act of will seems to manifest itself in such ways as trying to lift
weight, resisting temptations, or overcoming obstacles, so it is concluded that the
difference between cases with free will and the cases without is the occurrence or absence
of trying.

45
Even though trying and making efforts are conceptually related to free will, we do
not always try when we act freely. For example, “trying” makes sense only when one is
impeded, constrained or in the process of learning a new skill. The difference between I
am listening to the lecture and I am trying to listen to the lecture is that in the latter case
there are certain hindrances such as the hallway is noisy, or I am too tired to concentrate.
Moreover, we cannot always try to do something. One cannot try to communicate with
someone in Spanish if one does not speak Spanish at all.
In fact, trying to move my hand is logically connected with the movement of my
hand. The connection is logical or conceptual rather than causal. In other words, it is not
the trying that causes the intended movement. It is simply a movement with efforts as
opposed to a movement without any effort.

(4) The Causal Theory
The causal theorist analyzes acting with free will as a bodily movement caused by
a desire (or other pro attitudes) and a related belief. I will use “desire” to represent all the
pro attitudes that one may have when acting with free will in the following discussion.
For example, my act of opening the window is a free act because it is caused by my
desire for fresh air and my belief that I will let the air in if the window is open.
Donald Davidson defends such a view in his article “Actions, Reasons and
Causes”. He holds that “rationalization is a species of causal explanation”
31
. When we
try to explain why someone performs an action, we must name its primary reason which
is consisted of the pro attitude and the related belief. The primary reason justifies an

31
Donald Davidson, “Actions, Reasons and Causes” in Essays on Actions and Events, p. 3.

46
action by explaining it and this explanation is also a causal explanation because the
primary reason is necessarily the cause of the action.
It seems to me that there are two problems with this theory. First, it mistakes the
conceptual connection between desire and action as a causal connection. The concept of
desire is internally related to the concept of action. Desiring a pie means that one will try
to get a pie or eat it when one has it. But it is not the desire that causes one to act. It is
true that sometimes a desire may cause an action, such as a compulsive desire for drugs
may cause the drug addict to take the drugs, but it is not always the case. In regular
cases, our actions are not caused by a combination of desire and belief, because the
combination does not necessitate an action the way a cause necessitates its effect. The
action only takes place when we decide to act. Actually we have many desires we do not
act on. For example, I have a desire to take a nap when sitting in a meeting and believe
that I will have my desire satisfied if I step out of the room and go home. However this
desire and belief pair does not cause me to act as such. I remain sitting. Apparently, it
does not necessitate actions as the causal theorist suggests.
However, a Humean causal theorist may rebut it by arguing that the desire in the
desire and belief pair is not a single desire but the net balance of all the desires. One may
have a desire to do something but at the same time have another desire not to do it. For
example, in the above case, I not only have a desire to go home but also a desire to be
polite. The first desire pushes me to the direction of leaving the room but the second one
holds me back. So after the two desires fight against each other, the stronger one wins
and causes me to remain sitting in the room.

47
On this view, we always act on our strongest desire and it is logically impossible
to act without having any desire for the action or act on a desire that is not the strongest.
This is certainly not an empirical discovery. After all, how do we measure the strengths
of desires? If I make a right turn instead of left when I could have turned left, what is the
method of measuring the strengths of my desires of turning right or left? There is no
independent way of finding out which desire is stronger. The only way to identify the
strongest desire is to find out how the agent acts and the strongest desire will be the
desire for so acting. In this way, “the strongest desire” has lost its ordinary meaning and
become synonymous of free action or acting with free will.
Secondly, the causal theorist confuses “reason” with “cause”. The causal theory
cannot capture the idea of one’s having a reason to do something or trying to reach a
certain goal as opposed to one’s being forced to make certain bodily movements. One’s
performing an action in order to achieve a goal cannot be analyzed as a desire and belief
pair which causes a bodily movement. If it is the case, then we should be able to say
“Look, my arm is rising and now starts to wave. It must be caused by my desire to get
the attention from the guy across the street and my belief that one can usually attract
another’s attention by waving to him”. Sometimes it is true that I have to wait and see
what my final decision is when I am waving between two choices, but this is different
from guessing what another person is going to do.
If the causal theory is right, then what I do when I act freely will be forming a
causal hypothesis between my desires and beliefs and my action. For instance, whenever
the winter arrives, the tree in my backyard blossoms. So I reach the conclusion that the

48
arrival of winter causes the blossom of the tree. Similarly, I have discovered that
whenever I have a desire for saying something in class and the belief that raising my hand
would attract my teacher’s attention to call me, I would find my hand is up. Since it is
just a hypothesis, it is possible that my actions surprise me. I observe what I do the same
way that I observe what someone else’s does. I guess what I will do the same way that I
guess what another person with same desires and beliefs will do.
Here is another example. Assuming that I have the desire to kill someone and
believe that if I pull the trigger I will shoot him. My desire of killing him makes me
nervous and the anxiety causes my index finger trembles and that in turn causes the
trigger pulled. According to the causal theorist, my pulling the trigger is done of my own
free will, but apparently it is not.
The causal theorist may defend Davidson and argue that a reason can still be a
reason which explains why an action is performed, but meanwhile it is also a cause
because actions certainly take place in the “realm of causality”
32
. If a reason is left alone
and allowed to be a reason, then its meaning can be differentiated from that of a cause.
For example, the reason for my jumping into the pool is that I want to swim and the cause
of my falling into the pool is that someone pushed me. The causal theorist certainly does
not want to deny the contrast between the two cases. All he wants to emphasize is that
the reason is also a cause, even though this cause is different from the cause in the second
case. But what is the point of calling a reason “cause” if the “cause” here does not even
have an ordinary meaning? Instead of shedding light on any existing mystery, it causes
fresh confusion.

32
Ibid, p.19.

49
The causal theorist may also propose an analytic definition that a person acts
freely if and only if his actions are caused by his beliefs and desires in the right kind of
way. But this definition does not help explain or identify a free action because the
concept of free action is logically prior to the definition. This can be clearly shown when
we try to pin down “the right kind of way”. We will see that whenever the “way” is of
the “right kind”, it will coincide with what we ordinarily mean by acting freely.
So it seems to me wrong to conceive acting with free will as a bodily movement
that is caused by a pair of desire and belief, because acting freely is not composed of a
passive bodily movement and a mental occurrence. Moreover, volition is not a hidden
mental occurrence in the brain. If it were such a thing, we would not have the concept in
the first place, because we would not know how to employ it, or how to distinguish
correct from incorrect employment of the word. But since we do have the concept and
know how to use it, the technique of applying it must not be hidden or mysterious.
In sum, the concept of free will is not a name of a mental occurrence but a
concept we use under various circumstances to address certain needs of ours and serve
different purposes.

6. Conclusion
In this chapter, I examined in detail the most influential contemporary Analysis
Compatibilist theory – Frankfurt’s Hierarchical View of Compatibilism. I tried to show
that Frankfurt’s analytic definition of freedom of will is incorrect because the concept of
free will does not fit the explanation model of listing necessary and sufficient conditions

50
for the application of it. Frankfurt’s account also fails as a compatibilist view because it
does not adequately respond to the incompatibilist challenge. The root of Frankfurt’s
mistake, probably shared by many others, is to regard free will as a mental occurrence. I
tried to uncover it and detailed its problems.

51

Chapter Three: Strawson’s Compatibilism

The compatibilist account offered by Peter F. Strawson does not aim at
discovering conditions that are necessary and sufficient for the application of the concept
of free will, but reminding us of what our moral practice looks like and how the concepts
such as moral responsibility, excuse and exemption are internally related (Strawson
1962).
He discusses two opposing views on determinism’s possible impact on free will
and moral responsibility. One view is held by the pessimist who is an incompatibilist and
believes that the truth of determinism would render our moral notions, such as praising,
punishing or blaming inapplicable. The other view is held by the optimist who is a
compatibilist and thinks that our concepts of moral obligation and responsibility are in no
way threatened by determinism because the truth of determinism does not undermine the
purpose of moral practice, namely, regulating human behaviors in a way that will
produce the best outcome and maximize social utility.
Although Strawson defends a compatibilist view, he finds the optimist account
unsatisfactory because it overlooks an important aspect of our moral practice, that is,
reactive attitudes, a wide range of attitudes and feelings that we have toward each other.
He also rejects pessimism on the ground that the pessimist mistakes the truth of
determinism for a universal exemption or excuse for holding someone morally
responsible.

52
In order to remind us of what legitimate excuses and exemptions are, he describes
in detail our participant reactive attitudes when we interact with others in everyday life.
We may have attitudes toward others such as “gratitude, resentment, forgiveness, love,
and hurt feelings” when the relationships we stand to others are family, colleagues,
friends, lovers, etc. Our natural reactive attitude to behaviors manifesting good will is
gratitude and to behaviors manifesting ill will is resentment, but we are also capable of
suspending this attitude if some valid excuses can be presented, such as “He didn’t mean
to”, “He hadn’t realized”, “He didn’t know”, “He couldn’t help it”, “He had to do it”, “It
was the only way” or “They left him no alternative”. When we are convinced that one or
more of these excuses are applicable to a certain offensive behavior, we suspend or
abandon the resentful feelings and no longer hold the offender morally responsible.
33

Besides excuses, Strawson discusses another category of pleas – we apply
exemptions to qualified individuals either at the time action or all the time. There are two
types of exemptions. Type-1 exemptions are pleas such as “He wasn’t himself”, “He has
been under very great strain recently”, or “He was acting under post-hypnotic
suggestion” and Type-2 exemptions are the pleas such as “He’s only a child”, “he’s a
hopeless schizophrenic”, “His mind has been systematically perverted” or “That’s purely
compulsive behaviour on his part”. When we apply exemptions to an agent, we abandon
our reactive attitude and adopt an objective attitude. An objective attitude differs from a
reactive attitude because it may include “repulsion, fear, manipulation or treatment” but
cannot include “resentment, gratitude, forgiveness, anger or the love that adult feel

33
P. F. Strawson, “Freedom and Resentment”, in Free Will, edited by Gary Watson, 1982, p.75.

53
reciprocally for each other.”
34
The reactive attitude is a participant’s attitude toward his
equals who are either held responsible or excused and the objective attitude is a detached
researcher’s attitude toward his objects that are exempted from moral responsibility.
By describing standard reactive attitudes and objective attitudes, Strawson
believes that he has shown the compatibility between determinism and moral
responsibility because the acceptance of the truth of determinism would not and should
not lead to the decay or the repudiation of participant reactive attitude, as he put it,
“Would, or should, it mean the end of gratitude, resentment, and forgiveness; of all
reciprocated adult loves; of all the essentially personal antagonism?”
35
He believes that
the answer is no, because the truth of determinism implies that all reactive attitudes
would, or should, be replaced by objective attitudes, but given our human nature we
would not and should not live a life devoid of deep interpersonal emotional reactions and
commitments.
36

Strawson’s account can be read in two ways. One reading is that he describes
actual instances of holding responsible, applying excuses or exemptions. The examples
he offers are descriptive. An alternative reading is that he describes our actual practice of
holding responsible. The examples are paradigm cases where expressions surrounding
the concept of free will, such as “acting freely”, “holding responsible” and “applying
excuses or exemptions” are correctly employed. This reading grants the examples a
normative role.

34
Ibid., pp.78-79.

35
Ibid., p.80.

36
Ibid., p.81.

54
It is crucial to distinguish a practice from an instance. A practice is the standard
way of acting or reacting. For example, we typically blame an adult for breaking an
expensive vase but do not blame a two-year-old for doing the same thing. A practice sets
up the standard of correctness and an instance reports whatever people do, such as a
mother blaming her two-year-old for breaking her favorite vase. An instance can be
appropriate or inappropriate, correct or incorrect according to the existing standard. If we
do not make a distinction between a practice and an instance, we would not be able to
make the judgment that the mother who blames her two-year-old is reacting
inappropriately.
I shall try to show that the first reading renders Strawson’s account utterly
implausible and the second reading is a more charitable way of interpreting him.

1. The Examples are Descriptive
The first interpretation of Strawson’s account takes it as speaking about actual
instances of holding responsible and applying excuses or exemptions. This is considered
as the standard reading of Strawson defended by philosophers such as Jonathan Bennett,
Gary Watson, and Tim Scanlon.
Bennett in “Accountability (II)” holds that
This is one mark of the non-propositional nature of blaming, praising
etc. in Strawson’s account: feelings are made central, and are not tied
systematically to any propositions about their objects. My feeling of
indignation at what you have done is not a perception of your objective
blameworthiness, nor is it demanded of me by such a perception. It

55
expresses my emotional make-up, rather than reflecting my ability to
recognize a blame-meriting person when I see one.
37

Under this reading, Strawson offers an account of moral responsibility in terms of
feelings and emotional reactions. Praising or blaming is no longer tied to a perception or
belief of praiseworthiness or blameworthiness, but rather an expression of reactive
attitudes. The facts about moral responsibility are nothing ‘over and above’ the facts
about our dispositions to have reactive attitudes under certain circumstances.
Similarly, in his paper “Responsibility and the Limits of Evil: Variations on a
Strawsonian Theme”, Gary Watson characterizes Strawson’s account of moral
responsibility as an “expressive theory” because holding someone responsible is to be
explained or justified by the viewer’s emotional reactions, namely, reactive attitudes. He
writes,
Whereas traditional views have taken these attitudes to be secondary to
seeing others as responsible, to be practical corollaries or emotional
side effects of some independently comprehensible belief in
responsibility, Strawson’s radical claim is that these “reactive
attitudes” are constitutive of moral responsibility; to regard oneself or
another as responsible just is the proneness to react to them in these
kinds of ways under certain conditions. There is no more basic belief
which provides the justification or rationale for these reactions. The
practice does not rest on a theory at all, but rather on certain needs and
aversions that are basic to our conception of being human. The idea
that there is or needs to be such an independent basis is where
traditional views, in Strawson’s opinion, have gone badly astray.
38

In Strawson’s view, there is no such independent notion of
responsibility that explains the propriety of the reactive attitudes. The

37
Jonathan Bennett, “Accountability (II)”, in Free Will and Reactive Attitudes: Perspectives on P. F.
Strawson’s “Freedom and Resentment”, edited by Michael McKenna and Paul Russell, 2008, p.55.

38
Gary Watson, “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme”, in Free Will:
Critical Concepts in Philosophy, edited by John Martin Fischer, 2005, p.107.

56
explanatory priority is the other way around: It is not that we hold
people responsible because they are responsible; rather, the idea (our
idea) that we are responsible is to be understood by the practice, which
itself is not a matter of holding some propositions to be true, but of
expressing our concerns and demands about our treatment of one
another.
39

According to Watson, Strawson’s view is radically different from the traditional
views. Traditionally, reactive attitudes can be fair or unfair, appropriate or inappropriate,
justified or unjustified based on the basic belief about the agent’s culpability. In other
words, our holding someone responsible by showing our reactive attitudes towards him
can be justified by the fact that he is indeed responsible. Strawson rejects these beliefs
from the traditional views and holds that “there is no more basic belief which provides
the justification or rationale for these reactions” because holding responsible is nothing
but “expressing our concerns and demands about our treatment of one another”
40
.
Strawson’s view is utterly implausible when interpreted as such because he would
deny something apparently right, that is, we correctly hold someone responsible for doing
something because he is responsible for doing it. Under this reading of Strawson, one’s
reactive attitude, say, resentment can well convict an innocent person. For example,
according to the traditional views, Othello’s reactive attitudes of feeling jealous and
betrayed were unjustified because his wife Desdemona was not unfaithful to him. But
according to Watson’s reading of Strawson, she was responsible for committing the
crime of adultery simply because Othello’s emotional reactions were what they were.

39
Ibid., p.108.

40
Ibid.

57
Since Watson doesn’t distinguish a practice from an instance, naturally he does
not treat them differently. So his comments on Strawson cannot be consistently
interpreted as the second reading. In fact, many of them suggest that he reads Strawson
in the first way, especially when he addresses the differences between Strawson’s view
and traditional views.
Let us use libertarianism as an example of the traditional view. The libertarian
holds that Kaya’s resentment toward Leo, a drunk driver who killed a child, is secondary
to or justified by an independent belief that (1) Leo is responsible for his victim’s death
because he killed the child while driving drunk and he is a normal healthy adult and (2)
determinism is false, so Leo had free will and could have done otherwise.
When read in the second way, Strawson would also think that there’s an
independent belief that explains and justifies Kaya’s negative reactive attitudes toward
Leo. He would agree with the libertarian up to the last point because he thinks that
determinism is irrelevant to Leo’s responsibility. Strawson discusses in detail on what a
justifying belief could be. For instance, it could be “Leo is not a child but a 45 years old
man!” as the exemption that applies to a child would not apply to Leo, “He could have
stopped drinking at the party before it was too late” for he had no excuse to drink more
than he should, or “He should have known the danger of driving drunk because he is a
normal healthy adult” as what exempts a mentally ill would not exempt Leo.
Watson believes that Strawson’s view is radically different from the libertarian
view, however, it is not because Strawson denies (2), namely, the relevance of

58
determinism, but because Strawson denies (1), namely, the existence of an independent
belief that justifies Kaya’s reactions to Leo. Watson remarks,
There is no more basic belief which provides the justification or
rationale for these reactions.
41

In Strawson’s view, there is no such independent notion of
responsibility that explains the propriety of the reactive attitudes.
42

The explanatory priority is the other way around: It is not that we hold
people responsible because they are responsible; rather, the idea (our
idea) that we are responsible is to be understood by the practice, which
itself is not a matter of holding some propositions to be true, but of
expressing our concerns and demands about our treatment of one
another.
43

To think that Strawson denies (1) can only be explained by reading Strawson in
the first way, that is, to interpret Strawson’s claim that our practice of holding responsible
does not require a justification at the level of instances. According this reading, holding
Leo responsible is not justified because he is responsible but because Kaya finds his
action of driving drunk resentful.
This is probably why Watson finds Strawson’s discussion on excuses and
exemptions puzzling. He expresses his puzzlement in the following passage,
In one way, Strawson is anxious to insist that these attitudes have no
“rationale,” that they neither require nor permit a “rational
justification” of some general sort. Nevertheless, Strawson has a good
deal to say about the particular perceptions that elicit and inhibit
them.
44

41
Ibid., p.107.

42
Ibid., p.108.

43
Ibid.

44
Ibid., p.109.

59
If Strawson is read in the second way, there is nothing puzzling about these two
things that Strawson does, because they are at different levels. What Strawson insists is
that our having reactive attitudes or holding each other responsible as a practice does not
have rationale, but instances of having reactive attitudes or holding responsible do have
rationale and justification.
Watson’s interpretation of Strawson seems to me incorrect because it overlooks
the fact that Strawson distinguishes a justification inside a practice from a justification
outside a practice. An instance of holding someone responsible is operating within our
practice of holding responsible, so it can be correct or incorrect, justified or unjustified.
What Strawson rejects is not this kind of internal justification but the attempt to justify
the practice of holding responsible as a whole by trying to ground our moral practice and
moral concepts in certain beliefs such as the falsity of determinism or increasing social
utility.
Strawson’s view is understood by Watson as a view that explains moral
responsibility by our natural emotional reactions. But this is misunderstanding of
Strawson. First, reactive attitudes do not explain or justify holding responsible but rather
are constitutive of it. Secondly, what does not require or permit rational justification is
the fact that we have reactive attitudes at all rather than why we feel a particular action
resentful. The former does not rest on any theoretical ground (neither the falsity of
determinism nor social utility) and the latter is explained or justified by the facts about
the action in question and the circumstances under which it is performed.

60
One may defend Watson’s reading of Strawson by pointing out that in the
“Blaming and finding fault” section of the same article, Watson also states that
Consider the following common view of blame and praise: To blame
someone morally for something is to attribute it to a moral fault, or
“short-coming,” or defect of character, or vice, and similarly for
praise. Responsibility could be construed in terms of the propriety
conditions of such judgments: that is, judgments to the effect that an
action or attitude manifests a virtue or vice.
45

As I understand the Strawsonian theory, such judgments are only part
of the story.
46

The above passages clearly show that under Watson’s reading Strawson’s view is
not radically different from the “common view of blame and praise” because Strawson
agrees with the common view that reactive attitudes are reactions “to the quality of the
other’s moral self as exemplified in action and attitude” even though he thinks that it
should not be the whole story. What would complete the story is the range of emotional
reactions to others’ good or ill will. These reactions are constitutive of moral
responsibility rather than secondary to it.
According to this understanding of Watson’s reading of Strawson, if you stepped
on my foot by accident, but I mistakenly think you did it on purpose, and resent you for
it, then Strawson would say (and Watson reads him as saying) that my reactive attitude is
unjustified. It is unjustified because, according to Strawson, our reactive attitudes are
responses, not to the actual harm that has been caused by the behavior of another person,

45
Ibid., p.110.

46
Ibid.

61
but to the intention and attitude that the behavior manifested. Most specifically, we
resent the fact that the other person displayed either ill will or indifference towards us.
Similarly, in the case of Desdemona, since she was innocent of the crime Othello
accused her of, her intentions were blameless. Othello’s beliefs about her attitudes are
mistaken and because of this his reactive attitudes are unjustified.
I agree that it is clearly shown in this section that Watson reads Strawson as
wanting to keep the traditional view as part of his own view so that the reactive attitudes
can be justified or criticized by reference to the other person’s intention. However, a
claim as such flies in the face of the remarks he has been making in the earlier quotes,
especially in the remarks where he explains how Strawson’s view is radically different
from the traditional views, such as according to Strawson “There is no more basic belief
which provides the justification or rationale for these reactions.”
We may certainly grant this interpretation of Watson’s reading of Strawson, that
is, to read Strawson as saying that reactive attitudes are to be justified by the other
person’s intention or the quality of his will. If so, how does Strawson’s account differ
from the traditional views, say, the Kantian view? For Kant, the fundamental
justification of a good will is the will’s ability to pass the Categorical Imperative, but
what is it for Strawson? Does he have one? If he does not, would he adopt Kant’s
standard? It is also unclear what it means by saying that reactive attitudes are
constitutive of moral responsibility rather than secondary to it when they are to be
justified by some basic beliefs about responsibility. Watson does not address these
questions at all. As most of his writing, especially his discussion of the Harris case,

62
focuses on our reactive attitudes towards the offender (instead of the offender’s intention)
as justifications of the offender’s culpability, we have reason to think that Watson reads
Strawson as saying that there is no “independent notion of responsibility that explains
the propriety of the reactive attitudes”
47
, which is a view radically different from the
traditional views rather than accommodating those views.
Scanlon’s interpretation of Strawson is similar to Watson’s. He takes Strawson to
hold that moral judgments always involve the expressions of reactive attitudes, so he
raises objections as follows:
…it is not clear that moral judgments need always involve the
expression of any particular reactive attitude. For example, I may
believe that an action of a friend, to whom many horrible things have
recently happened, is morally blameworthy. But need this belief, or its
expression, involve a feeling or expression of moral indignation or
disapproval on my part? Might I not agree that what he did was wrong
but be incapable of feeling disapproval toward him?
48

When Strawson is read in the first way, he does seem to suggest that the
expression of a reactive attitude is necessary for holding someone responsible. If so,
Scanlon’s objection is certainly valid. But there is a more charitable way of reading
Strawson.

2. The Examples are Normative
An alternative interpretation of Strawson’s view is that the instance of the reactive
attitude in a particular case is not constitutive of holding someone responsible in that

47
Ibid., p.108.

48
T. M. Scanlon, “The Significance of Choice”, in Free Will, edited by Gary Watson, 2003, p.362.

63
case, but rather the concept of reactive attitude is constitutive of the concept of moral
responsibility. In other words, the concept of moral responsibility is internally (or,
conceptually) related to the concepts of emotional reactions such as gratitude and
resentment.
In order to understand Strawson’s view correctly, it is important to differentiate
justification at two levels, that is, the application level and the concept level. What
Strawson denies is the possibility of justifying the concept of moral responsibility, such
as what the consequentialist and the libertarian are trying to do. But he does not deny the
possibility of justifying a particular instance of holding someone responsible or one’s
having a particular reactive attitude.
When Strawson is interpreted this way, Scanlon’s alleged objection turns out to
be a confirmation of Strawson’s account. In the example Scanlon offers, he certainly
doesn’t have to feel resentment towards his friend because his friend is someone “to
whom many horrible things have recently happened,” which is, on Strawson’s account, a
typical Type-1 exemption that requires withdraw of his immediate reactive attitude of
resentment. Moreover, Scanlon carefully stipulates that this person is his friend and the
harm is not done to Scanlon personally. When all the excuses and facts are considered,
the example does not counter but support Strawson’s view.
In fact, if the conceptual relation between concepts is depicted carefully and
correctly, as Strawson has done, it would be difficult to construct a reasonable case
outside this conceptual map because when we describe an example, we can do nothing

64
but employing the existing concepts and follow the logic that governs these concepts,
otherwise what we say would not make any sense.
The following objection raised by Watson is also based on the first interpretation
of Strawson. He writes,
Thus, reactive attitudes depend upon an interpretation of conduct. If
you are resentful when jostled in a crowd, you will see the other’s
behavior as rude, contemptuous, disrespectful, self-preoccupied, or
heedless: in short, as manifesting attitudes contrary to the basic
demand for reasonable regard. Your resentment might be inhibited if
you are too tired, or busy, or fearful, or simply inured to life in the big
city. There are causal inhibitors. In contrast, you might think the
other was pushed, didn’t realize, didn’t mean to… These thoughts
would provide reasons for the inhibition of resentment. What makes
them reason is, roughly, that they cancel or qualify the appearance of
noncompliance with the basic demand.
49

A reactive attitude one actually has does depend on one’s interpretation of the
conduct in question, but it does not follow that the correctness or appropriateness of the
attitude depends on one’s interpretation alone and there is no objective way to find out if
the interpretation is justified. If I resent the harsh push on my back, my immediate
emotional reaction of resentment towards the person behind me may be a quick
interpretation of the push to be malicious. But when I’m told that he could not help
pushing me because he was pushed by another person, is this still my interpretation of the
conduct? If one still thinks it is, we may watch the videotape that has recorded what
happened. The point is that one’s interpretation of the conduct in question is not where
reasons locate.

49
Watson, p.109.

65
Watson seems to think that the existence of “causal inhibitors” posts threat to the
normativity of reactive attitudes. But it seems to me wrong to think so. When we say a
chair is red, we do not mean that it is red in a room with red lightings, when we wear red-
tint glasses or if we are color-blind, rather we mean that the chair is red when it is in
daylight viewed by people with healthy eyes. But it does not follow that the redness of
the chair depends on one’s interpretation of the color. Rather, this is what we mean by
color red. Similarly, when Strawson talks about resentment, normal conditions have been
assumed, that is, the conditions under which the causal inhibitors that Watson has in mind
are not applicable. So it is wrong to think that “reactive attitudes depend upon an
interpretation of conduct”
50
.
Watson thinks that Strawson takes our ordinary moral practices for granted, so he
complains,
If, for whatever reason, reactive attitudes are sensitive to historical
considerations, as Strawson acknowledges, and we are largely ignorant
of these matters, then it would seem that most of our reactive attitudes
are hasty, perhaps, even benighted, as skeptics have long maintained.
In this respect, our ordinary practices are not as unproblematic as
Strawson supposes.
51

Watson’s worry is that in our everyday life we often don’t know enough about
another person’s history so our immediate reactive attitudes are often hasty and unfair.
This is certainly true and is also why there is such thing as getting to know or understand
someone better. But this is a problem for Strawson’s view only when it is understood as
an “expressive theory”. If interpreted in the second way, Strawson’s account would not

50
Ibid.

51
Ibid., p.129.

66
be susceptible to the problem because under this interpretation reactive attitudes can be
appropriate or inappropriate.
Under the second reading, Strawson offers a normative account of the concepts of
reactive attitude, excuse and exemption, that is, when we react in accordance with it, we
are reacting correctly or appropriately. In other words, he is describing a norm or
standard of correct reactions rather than reporting whatever people do in everyday life, so
his account leaves no room for incorrect reactive attitudes either caused by
misinterpretations of the conduct or dysfunctions of the person who reacts. This doesn’t
mean that there cannot be inappropriate or incorrect reactive attitudes at the level of
instances, but these instances of reactive attitudes (i.e. the mother holds her two-year-old
responsible for breaking the vase.) are not what Strawson’s account of reactive attitudes
is about.
This is why I find Watson’s remark that “reactive attitudes depend upon an
interpretation of conduct” and examples that manifest attitudes “contrary to the basic
demand for reasonable regard” an indication of the fact that he reads Strawson in the first
way.
Watson mistakenly thinks that his worry is the same as what the skeptic has in
mind. He worries whether our everyday reactive attitudes towards others are fair. The
skeptic’s worry is more fundamental; he thinks that if determinism is true, we cannot
fairly hold anyone responsible for anything. These two worries sound pretty similar to
each other but in fact they are radically different. In the former the distinction between
fairness and unfairness remains where it is, even though it may be true that most of our

67
reactive attitudes are unfair. However, in the latter the distinction is removed from where
it is usually drawn and placed in a new location, that is, between determinism and
indeterminism. As a result, the fair and appropriate reactive attitudes according to the
standard that Watson has in mind will turn out to be unfair and inappropriate by the
skeptic standard given the truth of determinism.

3. The Harris Case
The Harris’ case is introduced and used by Watson to raise more objections to
Strawson’s view. Watson believes that Strawson would have difficulty in accounting for
the fact that the history of Harris gives pause to our reactive attitudes of resentment
because the recognition of the historical factors that helped Harris become who he is
would lead one down to the incompatibilist path. He wonders if “the expressive theory
[can] explain why the reactive attitudes should be sensitive to such an explanation”.
52

If we adopt the second reading, I don’t see why it should be a problem for
Strawson’s view to show sensitivity to Harris’ childhood experiences. Harris was abused
as a boy and didn’t have a normal family that most people have. This is why his history
gives pause to our immediate reactive attitude of resentment. On Strawson’s account, the
information about individuals’ background or “how they came to be” can well serve as a
reason for the viewer to modify his reactive attitudes.
Watson worries that the sensitivity to Harris’ childhood experiences would
become “grist for the incompatibilists’ mill”.

52
Ibid., p.122.

68
A plea of this kind is, on the other hand, grist for the incompatibilists’
mill. For they will insist on an essential historical dimension to the
concept of responsibility. Harris’s history reveals him to be an
inevitable product of his formative circumstances. And seeing him as
a product is inconsistent with seeing him as a responsible agent. If his
cruel attitudes and conduct are the inevitable result of his
circumstances, then he is not responsible for them, unless he was
responsible for those circumstances. It is this principle that gives the
historical dimension of responsibility and of course entails the
incompatibility of determinism and responsibility.
53

It is unclear why the belief that Harris’s childhood experiences have important
influences on how he behaves as an adult has an incompatibilist implication. All the
incompatibilists believe the relevance of determinism to free will and moral
responsibility. But in the Harris’ case, the relevance is between an abused childhood and
responsibility rather than between determinism and responsibility. In order to recognize
a traumatic childhood, one does not have to believe that determinism is true. For
instance, I believe that the death of the plant in my backyard is a result of the sudden
change of weather, I can hold this belief without believing that determinism is relevant to
this fact. In other words, there is nothing wrong with saying that the death of the plant is
caused by the weather regardless whether determinism is true or false. Similarly, the
belief that smoking for thirty years caused Malik’s lung cancer has nothing to do with the
truth of determinism. So, if I think that Harris’s cruelty is a result of his traumatic
childhood, I don’t automatically accept the truth of determinism. In fact, the truth of
determinism is irrelevant to the belief that our behaviors, personalities, and habits are
partially shaped by our environment.

53
Ibid., p.123.

69
Another objection raised by Watson is that complex emotional response to
Harris’s case on Strawson’s account is impossible because the facts about Harris’s moral
responsibility are nothing over and above the fact that other people have reactive attitudes
towards him. He writes,
How and why, then, does this larger view of Harris’s life in fact affect
us? It is too simple to say that it leads us to suspend our reactive
attitudes. Our response is too complicated and conflicted for that.
What appears to happen is that we are unable to command an overall
view of his life that permits the reactive attitudes to be sustained
without ambivalence. … In fact, each of these response is appropriate,
but taken together they do not enable us to respond overall in a
coherent way.
54

It would be a problem for Strawson to account for this kind of complex emotional
response if Strawson is read in the first way, that is, to think that he denies there is an
independent belief about Harris’s responsibility that justifies our reactive attitudes
because Harris’s responsibility is to be explained by our reactive attitudes.
When Strawson is read in the second way, the reactive attitudes that people have
towards Harris must be justified in the sense that they correspond to the fact about his
moral responsibility. In other words, we hold Harris responsible to certain degree
because he is responsible to certain degree. So our justified reaction to Harris is to resent
him (for his free choices) and pity him (for his unfortunate circumstances) at the same
time. If someone reacts differently, his reaction is inappropriate or incorrect according to
our existing standard described by Strawson. What gives us pause is Harris’s abused
childhood, a force outside his control, which partially caused him to be cruel and lack of
empathy. Harris himself was also partially responsible for being the way he was because

54
Ibid., p.123.

70
he could have chosen better throughout the years of growing up and being an adult, at
least, as well as his siblings did.
Strawson never says that one cannot have both reactive attitude and objective
attitude at the same time toward the same person. After all, life is not so simple. It
should be perfectly fine if “each of these responses is appropriate” and they cannot be
taken together to form a simple view.
The ambiguity and complexity we find in the Harris case should not be a
philosophical problem because they are concerned with the application of concepts rather
than concepts themselves. Let us say, we try to apply the concepts of black and white to
two objects X and Y in order to find out which is darker. Let us say X is a white board
with black dots and Y is a black board with white dots. The question “Which is darker?”
is a hard question because neither X nor Y has one color. The Harris case is difficult in a
similar way. Both his unfortunate childhood and his own choices (so far nothing in the
story shows that he was so traumatized that he was incapable of making free choices) led
him to where he was when he committed the brutal murder, so it is not surprising that we
have mixed emotional reactions towards him and his conduct.
One may raise another objection that “if I had had Harris’s upbringing and genetic
endowment, I would have turned out the same way.” This objection assumes that our
free choices and decisions are determined (as opposed to influenced) by our upbringing
and genetic endowment. But is it right to identify someone with his genes? When two
individuals with exactly the same genes, will they always react to the same situation in
the same way? If the objection is just a way of saying “If I were Harris, I would have

71
turned out the same way”, then it is certainly right, because if I were Harris, I would no
longer be me and Harris certainly turned out the way he actually had turned out.
However, it is hard to see how this can be an objection to the fact that we are justified to
resent Harris’s evil deeds.
I think that Strawson’s compatibilist account is susceptible to criticisms, but it is
not susceptible to the criticisms that Watson has advanced, because it seems to me that
Watson’s criticisms are based on a misreading, at least, uncharitable reading of Strawson.
A charitable way of reading Strawson is the second one, that is, to see him as
making the claims that (1) holding someone responsible correctly is nothing but (or,
constitutive of) having justified reactive attitudes toward this person (and vice versa), (2)
holding someone responsible correctly is justified by the fact that this person is
responsible, and (3) one’s responsibility is determined within the practice where excuses
and exemptions are among the considerations while determinism is not.
One may ask a question regarding (2) and (3) that which fact renders someone
who caused another person a harm or injury responsible? There can be at least two kinds
of fact – one is that the harm-doer’s intentions or will was bad and the other is that the
harm-doer does not fall under one of the standard excuses or exemptions.
It seems to me that on the second reading of Strawson these two facts are actually
one and the same. To say someone’s intention was bad amounts to say that this person
harmed another and he cannot be excused or exempted. There is no independent way of
determining the moral quality of a harm-doer’s intention or will.

72
4. Compatibilism
According to Strawson, our interpersonal relationship and reactive attitudes such
as resentment and gratitude will remain the same regardless whether determinism is true
or false, so determinism is irrelevant to moral responsibility. Strawson argues,
It does not seem to be self-contradictory to suppose that this [replacing
reactive attitudes with objective attitudes] might happen. So I suppose
we must say that it is not absolutely inconceivable that it should
happen. But I am strongly inclined to think that it is, for us as we are,
practically inconceivable. The human commitment to participation in
ordinary inter-personal relationships is, I think, too thoroughgoing and
deeply rooted for us to take seriously the thought that a general
theoretical conviction might so change our world that, in it, there were
no longer any such things as inter-personal relationships as we
normally understand them precisely is being exposed to the range of
reactive attitudes and feelings that is in question.
55

A sustained objectivity of inter-personal attitude, and the human
isolation which that would entail, does not seem to be something of
which human beings would be capable, even if some general truth
were a theoretical ground for it.
56

On Strawson’s view, the incompatibilist suggestion that we suspend or abandon
reactive attitudes and adopt objective attitudes if determinism is true is not theoretically
inconceivable but practically impossible. Given our human nature and the way we live,
we cannot “take seriously the thought that a general theoretical conviction” will make us
abandon our deep commitment to reactive attitudes. We may have “a theoretical ground”
to do so, but it cannot be done because our reactive attitudes are “too thoroughgoing” and

55
Strawson, p.81.

56
Ibid.

73
our “commitment to participation in ordinary interpersonal relationships” is too deeply
rooted in our human nature to give up.
57

Strawson’s response does not seem to me adequate because he only explains why
we would not abandon reactive attitudes when reason requires us to do so but does not
answer the question why we should not. As what Strawson is fully aware of,
…the real question is not a question about what we actually do, or why
we do it. It is not even a question about what we would in fact do if a
certain theoretical conviction gained general acceptance. It is a
question about what it would be rational to do if determinism were
true, a question about the rational justification of ordinary inter-
personal attitudes in general.
58

What would be rational to do if determinism is true? If our ongoing moral
practice is unjustifiable and thus irrational to be continued, then it should be ended even
though it might be a difficult task.
In response Strawson points out that a practice as a whole is not “something that
can come up for review” as rational or irrational because “our natural human
commitment to ordinary inter-personal attitudes” is “part of the general framework of
human life”.
59
But Strawson does not further explain why we cannot choose something
that belongs to the general framework of human life and why the truth of determinism
does not provide us with a good reason to make such a choice.

57
Strawson’s argument is similar to Hume’s solution to his own skepticism. Kant has shown that Hume’s
skeptical argument is confused because its reason, the empiricist rationality, is misguided. In the case of
causation, it is not that our nature is too stubborn to follow reason, but that the alleged reason is in fact a
disguised confusion.

58
Strawson, p.82.

59
Ibid., p.83.

74
Watson raises a similar objection to Strawson’s compatibilist conclusion. He
wants to show that the truth of determinism is relevant to our reactive attitudes, but the
reason he offers to support his point seems to me incorrect. He writes,
For determinism seems to entail that if one had been subjected to the
internal and external conditions of some evil person, then one would
have been evil as well. If that is so, then the reflections about moral
luck seem to entail that the acceptance of determinism should affect
our reactive attitudes in the same way as they are affected in Harris’s
case. In the account we have suggested, then, determinism seems to
be relevant to reactive attitudes after all.

Actually, this conclusion does not follow without special metaphysical
assumptions. … To congratulate me on these grounds would be to
congratulate me on being myself. …
60

It is true that “if one had been subjected to the internal and external conditions of
some evil person, then one would have been evil as well”, but it does not follow from this
that “the acceptance of determinism should affect our reactive attitudes in the same way
as they are affected in Harris’s case”
61
because the two ways in which our reactive
attitudes are affected are completely different.
To say that “if one had been subjected to the internal and external conditions of
some evil person, then one would have been evil as well” is amount to say that “if one
happened to be an evil person, then one would have been evil”. A statement as such is
necessarily true because it has the form “if p is true then p is true”. It is unclear how this
truth should affect our reactive attitudes.

60
Watson, p.126.

61
Ibid.

75
Every person is certainly who he or she is, but it does not contradict the fact that a
person also participates in making himself or herself because there are not only “external
conditions” but also “internal conditions” for being a certain individual. A traumatic
childhood is an external condition that exempts only part of Harris’s moral responsibility.
Otherwise, even if Harris were brought up by loving parents and had a normal childhood,
the truth of determinism would still require us to abandon our reactive attitudes and adopt
objective attitudes towards him. So the way in which the Harris case affects our reactive
attitudes is different from the way in which the truth of determinism affects our reactive
attitudes if it has any influence at all. Obviously, the incompatibilist owes us an
argument for the claim that determinism is an exemption or an excuse. In Chapter 4, I
will consider some arguments and argue that they all fail.
Although Watson’s argument is problematic, his complaint about Strawson’s
answer to the incompatibilist challenge is right on the point because Strawson’s response
is far from adequate.

5. Objective Attitudes
According to Strawson, the incompatibilist insists that the truth of determinism
requires us to abandon our reactive attitudes and adopt objective attitudes. Strawson
concedes that such a requirement might be theoretically valid even though it is
impractical. I will argue that Strawson fails to understand how radical the implications of
determinism are, according to the incompatibilist. The truth of determinism cannot

76
impose such a requirement on our moral practice because both objective and reactive
attitudes imply the possibility of free will.
First, according to Strawson, we naturally adopt reactive attitudes towards adults
and objective attitudes towards children. This observation is certainly correct, but he is
not aware that when we interact with a child, we are not interacting with a robot. The
way we manipulate a child’s behaviors, choices, or desires is radically different from the
way we change a robot’s movements. We may use a child’s desire for chocolate to
convince him to wash hands or use his fear for toothache to persuade him to brush teeth.
No matter what desire we want a child to have, if we are successful, the desire has to
become his own. But there is no such thing with a robot because the robot, unlike the
child, is not even a potentially free agent. The objective attitude requires us to treat the
child as a potentially free agent, or as a partly developed free agent, so the incompatibilist
claim cannot be that we should abandon reactive attitudes in favor of objective attitudes.
Secondly, even though we usually adopt objective attitudes towards a child as he
is not a fully-developed free agent, it does not follow that a child should not be held
responsible for whatever he does. I am sure we all agree that a schoolboy should be held
responsible for not doing his homework; a potty-trained three-year-old should be held
responsible for wetting her pants; and a healthy toddler should be held responsible for not
sleeping through the night. If it is fair to hold a child responsible for certain behaviors,
the objective attitude implies the possibility of free will.
Thirdly, if determinism is true, a thoroughly consistent incompatibilist will find
that no concept of action will be possible. For example, no one will be able to do

77
anything, such as to raise a hand, go home, stop running, or change an attitude. If so, we
will not be able to suspend or abandon reactive attitudes and adopt objective attitudes. It
is not that we do not have the ability to do so, but rather there is no such thing as
“suspend”, “abandon”, or “adopt” because no one will be able to initiate an action.
Susan Wolf describes in detail a scenario where we treat each other solely with
objective attitudes:
Imagine for a moment what a world would be like in which we all
regarded each other solely with the objective attitude. We would still
imprison murderers and thieves, presumably, and we would still sing
praises for acts of courage and charity. We would applaud and
criticize, say ‘thank you’ and ‘for shame’ according to whether our
neighbors’ behavior was or was not to our liking. But these actions
and words would have a different, shallower meaning than they have
for us now. Our praises would not be expressions of admiration or
esteem; our criticism would not be expressions of indignation or
resentment. Rather, they would be bits of positive and negative
reinforcement meted out in the hopes of altering the character of others
in ways best suited to our needs.
62

Wolf has correctly pictured a world which substitutes all reactive attitudes with
objective attitudes, but she has not realized that this new world cannot be what the
thoroughly consistent incompatibilist has in mind if determinism is true, because if
determinism is true, nobody would be able to “imprison” someone, “sing” something,
“applaud”, or “criticize” because all these concepts imply the possibility for one to
originate an action. These actions are compatible with objective attitudes but
incompatible with determinism.

62
Susan Wolf, “The Importance of Free Will” in Free Will and Reactive Attitudes: Perspectives on P. F.
Strawson’s “Freedom and Resentment”, edited by Michael McKenna and Paul Russell, pp.72-73.

78
A consistent incompatibilist should not think that the truth of determinism
requires us to make any change about our moral practice, let alone adopt a universal
objective attitude towards everybody. If a criminal could not help but committing a
crime, having reactive attitudes toward the criminal is also something an observer could
not help but doing. It would be logically inconsistent if the incompatibilist excuses the
criminal but holds the observer responsible.
It is certainly hard to be a thoroughly consistent incompatibilist because he would
not be able to use many words, terms, expressions that we use on a daily basis. For
instance, he is not entitled to use expressions such as “we should” or “we should not”
because the concept of the norm of our behavior implies the possibility of free will and
agency.

6. Strawson’s Naturalism
In his paper “Strawson’s Way of Naturalizing Responsibility”, Paul Russell
argues that Strawson’s compatibilism fails because he has to adopt token-naturalism in
order to refute Pessimism, but token-naturalism is an implausible view.
According to Russel, two kinds of naturalism can be distinguished in Strawson’s
refutation to Pessimism even though Strawson himself does not make such a distinction.
One kind of naturalism is type-naturalism which is a refutation to type-pessimism. The
type-pessimist believes that “if determinism is true, then we are not justified in being
disposed or prone to reactive attitudes and that we must, therefore (somehow) rid

79
ourselves of this type or species of emotion.”
63
In response, the type-naturalist claims
that our liability to certain emotional reactions “is natural to humans and requires no
general justification of any sort”
64
, so it is impossible for us to disengage from reactive
attitudes such as resentment and gratitude at this level.
The other kind of naturalism is token-naturalism which is a refutation to token-
pessimism. The token-pessimist holds that “we can and must cease to entertain reactive
attitudes toward any and all individuals who are morally incapacitated and that we are
capable of ceasing altogether to engage or entertain reactive attitudes insofar as we have
reason to believe that everyone is incapacitated in the relevant ways” and that if
determinism is true “then we are, indeed, all morally incapacitated.”
65
The token-
naturalist holds that “to take up the objective attitude… involves ceasing to entertain
(tokens of) reactive attitudes toward some or all individuals” and does not “involve
giving up our disposition or proneness to such attitudes.”
66

Russel argues that token-naturalism is far less plausible than type-naturalism
because in particular cases we are able to suspend or abandon our reactive attitudes in
accordance with reason but we are unable to suspend or abandon reactive attitudes
altogether even when reason requires us to do so. However, since Strawson’s target is

63
Paul Russell, “Strawson’s Way of Naturalizing Responsibility”, in Free Will and Reactive Attitudes:
Perspectives on P. F. Strawson’s “Freedom and Resentment”, edited by Michael McKenna and Paul
Russell, p.97.

64
Ibid.

65
Ibid., p.97.

66
Ibid., p.98.

80
token-pessimism, he has to embrace the implausible view of token-naturalism. This is
why his view fails.
It seems to me that the mistake that Strawson makes by thinking that the truth of
determinism requires us to abandon reactive attitudes and adopt objective attitudes
renders his account susceptible to Russel’s criticism. The weakness that Russel has
identified is that if Strawson agrees that we are able to change an attitude in one case then
we should be able to do so in more than one or even all the cases if needed because the
requirement would not be any different in nature but only in quantity.
I do not think that Strawson’s view as the way it is is able to respond to this
objection, because the objection is the incompatibilist challenge that Strawson has not
sufficiently addressed. The challenge cannot be met unless it is viewed from a correct
compatibilist point of view. For example, it can be pointed out that the token-pessimist is
self-contradictory. If everyone is morally incapacitated then who is able to abandon one
attitude and adopt another? Apparently, if someone is a pessimist (or, incompatibilist),
he cannot be a token-pessimist. Can he be a type-pessimist? The answer is still no,
because if determinism is true then who is able to rid himself of “certain type or species
of emotion”? So, it seems to me that the pessimist (or, incompatibilist) challenge is
problematic in its own right and Strawson’s response has shown that he fails to recognize
this fact.

81
7. Conclusion
In this chapter, I explained another important contemporary compatibilism – Peter
F. Strawson’s reactive attitude view and discussed two ways of reading it. The first
interpretation of it is to view it as trying to explain moral responsibility through instances
of emotional reactions. Under this reading, all the examples of reactive attitudes and
objective attitudes that Strawson offers are descriptive. This is the standard way of
reading Strawson. I tried to argue that this reading renders Strawson’s account utterly
implausible. There is a more charitable way of reading Strawson, that is, to view all the
examples of emotional reactions related to moral responsibility as normative so that
Strawson is delineating the conceptual connections among concepts surrounding free will
such as moral responsibility, reactive attitude, excuse and exemption. I also discussed
Paul Russell’s criticism of Strawson’s compatibilism and argued that Strawson has no
resource to refute Russell because he mistakenly believes that the truth of determinism
requires us to abandon reactive attitudes and adopt objective attitudes. It seems to me that
Strawson’s compatibilism fails because it cannot answer the incompatibilist challenge.

82

Chapter Four: The Manipulation Argument for Incompatibilism

There are different kinds of arguments against compatibilism and/or in favor of
incompatibilism. The Manipulation Argument I will be concerned with in this chapter is
usually considered as a serious threat to compatibilism as well as a compelling defense
for incompatibilism. As almost all contemporary compatibilists have, in the years after
Frankfurt constructed the Hierarchical View of compatibilism, come to agree that
freedom of will should be distinguished from freedom of action, acting intentionally no
longer suffices for acting with free will because acting with free will is acting in
accordance with second-order desires or values. But this view is unable to resist the
Manipulation Argument, a powerful counterexample to the Hierarchical View.
In the following discussions, I shall use “acting freely”, “acting of one’s own free
will”, “acting with free will”, and “morally responsible actions” interchangeably because
they are all conceptually related to one another and to the central concept in the free will
debate, namely, the concept of free will. Instead of believing these expressions are
always logically equivalent, I am convinced that there are enough overlapping cases to
which any of these expressions can be applied correctly. So focusing on these cases
should be sufficient for our present purpose of investigating into the nature of the
Manipulation Argument.
I shall examine some highly important and influential versions of the
Manipulation Argument – Peter van Inwagen’s Martian Device Argument, Derk

83
Pereboom’s Four Case Manipulation Argument and some of Robert Kane’s arguments
and argue that they all fail.
Van Inwagen invites us to consider the possibility that we are not free agents but
mistake ourselves to be free. We think we can choose, decide and act freely when in fact
we are controlled by a tiny device implanted in our brains by Martians. As a result, our
enjoyment of free will is illusory. If so, standard examples of free agency are not genuine
cases of freedom and thus the Paradigm Case Argument, one of the arguments in favor of
compatibilism, fails.
Pereboom has offered a more ambitious manipulation argument – the Four Case
Argument used as a positive argument for incompatibilism. Pereboom presents four
cases in each of which Professor Plum murdered Ms. White in the manner that we
intuitively would not hold Plum responsible. If our intuition is justified, determinism is
not compatible with free will or moral responsibility.
The incompatibilist Robert Kane introduces two kinds of control in his book
Significance of Free Will: one is constraining and the other is not. When someone is
under constraining control, he is forced to do something against his will. However, if he
is under nonconstraining control, he does not “feel frustrated or thwarted” because he
only wants or desires what his controller wants him to want or desire. For example, when
S is subject to a constraining control, he may be chained or threatened; when S is subject
to a nonconstraining control, he may be brainwashed. If S is not aware that he is subject
to a nonconstraining control, then the control is a “covert” nonconstraining control,

84
namely, CNC control. The CNC control is also called “global control” or “global
manipulation”.
67

Both the Martian Device Argument and the Four Case Manipulation Argument
appeal to the idea of global manipulation, but they fail to distinguish the regular kind of it
from the special kind of it. The regular kind refers to something like brainwashing or
mind-control. A successfully brainwashed agent can be distinguished from an un-
brainwashed agent even though they may both have undergone the same training session.
Examples range from political brainwashing in the Chinese Cultural Revolution when the
masses believed that Chairman Mao could never make mistakes to religious
brainwashing in various cults. However, the special kind of global manipulation is
different. An agent subject to this kind of manipulation cannot be distinguished from a
free agent even by a trained psychologist who closely questions and examines the agent.
I shall try to argue that both the Martian Device Argument and the Four Case
Manipulation Argument fail because they rely on the idea of the special kind of global
manipulation.

1. The Martian Device Argument
The Martian Device Argument is constructed by van Inwagen as a
counterexample to the Paradigm Case Argument, one of the arguments in support of
compatibilism. He believes that it is “possible to think of propositions that are consistent
with all our observations and which have the consequence that no one can do otherwise

67
Robert Kane, The Significance of Free Will, 1996, p.65.

85
than he does.”
68
I shall quote a passage related to the Martian Device Argument (M) first
and then the argument itself:
If we should discover that some particular person – Himmler, say –
acted as he did because a Martian device implanted in his brain at the
moment of his birth, had caused all his decisions, then we should hardly
want to say that Himmler had free will, that he could have helped what
he did, that he had any choice about the way he acted, or that he ever
could have done otherwise. And I don’t see why matters should be
different if we discovered that everyone was ‘directed’ by a Martian
device: then we should have to make these judgments about everyone.
69

(M) When any human being is born, the Martians implant in his brain a
tiny device – one that is undetectable by any observational technique we
have at our disposal, though it is not in principle undetectable – which
contains a ‘program” for that person’s entire life: whenever that person
must make a decision, the device causes him to decide one way or the
other according to the requirements of a table of instructions that were
incorporated into the structure of the device before that person was
conceived.
70

Under the Martian manipulation, Himmler’s free decision to quit smoking or take
a vacation will not be genuinely free because it is brought about by the Martian device. If
we recognize the lack of free will in Himmler, we should doubt about ourselves as well
because we could be in a similar situation, that is, covertly manipulated by the Martians.
Of course, the chance of the existence of some exterritorial creatures who outwit human
beings is slim, but if determinism is true, determinism will play exactly the same role as
the hostile Martians because the way we are subject to the global manipulation
administered by the Martians is the same as the way we are subject to the power of a
deterministic world.

68
Peter van Inwagen, An Essay on Free Will, p.109.

69
Ibid., p.110.

70
Ibid., p.109.

86
The Martian Device Argument heavily relies on the idea of global manipulation.
Now let us take a closer look at the idea through a thought experiment.
Himmler, Jimmler and Timmler were normal and healthy human triplets born and
brought up on Earth. One day all three of them were kidnapped by the Martians. When
they reappeared on Earth after a few weeks, they reported that the Martian scientists had
implanted some tiny devices in their brains for a Martian research project. They
implanted an H-shaped device in Himmler’s brain, a J-shaped device in Jimmler’s brain
and a T-shaped device in Timmler’s brain. But the Martians didn’t tell them what the
devices were and what functions they had.
In order to determine the functions of the Martian devices, the human scientists
watched the brothers closely and studied them carefully. After a while, they believed that
they had figured out what the H and J devices did to Himmler and Jimmler respectively.
Himmler used to tell everyone that he wanted to go to a Graduate School but
when he received five admission letters from five distinguished graduate schools, he
turned them all down. His friends were surprised at his decision. They asked him why
he spent so much time and energy on applying but ended up not going. He could not
figure out why either. Similar things happened every now and then ever since his trip to
Mars. For example, he made an important appointment to meet with his business partner,
but on his way to the meeting he stopped at a park and sat there for the rest of the day.
When his boss demanded an explanation, he could not offer any. The best thing he could
say was, “I suddenly felt like driving to the park and sitting there.”

87
Based on plenty of evidence, the human scientists concluded that the H device
could interfere with Himmler’s practical reasoning process. Whenever it was active,
Himmler would be unable to choose, decide, judge or act freely. So it is reasonable to
say that the function of the H device was to damage Himmler’s free will occasionally.
Unlike Himmler, Jimmler never said or did anything contradicting himself or
defeating his own purposes. He always liked music although did not have any talent in it.
Ever since his trip to the Mars, he began to manifest extremely strong confidence in his
musical talent. He firmly believed that he was as gifted as Mozart and anyone who
doubted this was simply jealous of him. He quit his firefighter job, sold his house, left
his wife and children, and broke up with his friends who tried to stop him so that he could
devote himself to composing music.
It seemed obvious to the human scientists that, like the H device, the J device had
a negative impact on Jimmler’s practical reasoning process as well. It accomplished this
job by boosting his ego and confidence to a dangerous degree. But unlike the H device,
the J device’s influence is not on and off, but a permanent, consistent and ongoing
process. So the human scientists concluded that Jimmler’s free will has been
permanently damaged by the J device.
Things were different with Timmler. The evidence showed that nothing about
Timmler was abnormal. He made reasonable choices and judgments in pursuing his
goals. The T device did not seem to have any influence on him. So the human scientists
believed that Timmler could enjoy free will the way any healthy normal human being

88
did. Compared to his brothers, Timmler was very lucky because his capacity for practical
reasoning and acting freely were left untouched by the Martians.
At this point, the Martian scientists dropped a message on earth to point out that,
contrary to what the human scientists believed, Timmler was no longer a free agent after
his trip to Mars because the T device contained a program for his entire life from the
moment of implanting the device to his death. The human scientists were highly alert.
However, learning from the past experience, they knew that Martian scientists’ words
should never be taken at face value.
It is claimed by the Martians that the T device “contains” every single detail of
Timmler’s life in the sense that whenever he must make a decision, “the device causes
him to decide one way or the other according to the requirements of a table of
instructions that were incorporated into the structure of the device”.
71
The idea is that
Timmler can never make a genuinely free choice or decision because it is the T device
that chooses or decides for him. But is this scenario logically possible? In other words,
is it logically possible for the T device to contain Timmler’s life without containing other
people’s lives as well as the complete states of the world? Let us consider an example.
At a certain point of Timmler’s life, he will have to decide either to run into a
burning house to save a child or stay away from it and remain safe. Timmler’s decision
is to save the child. In order to contain this decision, the T device must contain the
terrorist’s decision to follow his leader by setting the house on fire. In order to contain
the terrorist decision, the T device must also contain the terrorist leader’s decision to
launch this particular attack. In addition to all the decisions involved in the complex

71
Ibid.

89
causal network, the T device needs to know all the facts of the world other than human
decisions. Timmler will not drive by the house on fire if he arrives in the city two days
earlier as it would be the case if his flight were not delayed by the bad weather.
Apparently in order to contain Timmler’s decision to save the child, the T device needs to
be good at weather forecasting as well. But how plausible is it for an empirical device to
be as omniscient as God?
Not being misled by the Martians, the human scientists maintained that there is no
such thing as knowing every single detail of one’s life without knowing every single
detail of everyone else’s life and every single detail of the world. But nothing empirical,
certainly including the T device, is capable of accomplishing this kind of supernatural
work.
The Martian scientists respond that the T device is a computer program with a
long but finite list of instructions of the form “If you are ever faced with a choice
between running into a burning house to save a child, or staying safe, save the child” and
“If you are ever faced with a choice between running into a burning house to save an
adult, or staying safe, stay safe” and so on.
The human scientists were not moved by this argument because they knew that in
order for the program to accommodate all the possibilities and leave no gap, all the
possibilities must be known to the Martian scientists when the program was designed.
For example, at the moment when Timmler was about to run into the burning house, a
terrorist in a car started a shooting rampage and some pedestrians including a child were
injured. What should Timmler do? Saving the child in the house or the one on the

90
street? If the program leaves the decision to him, then Timmler is left with free will; if
the program decides for him, then it must be omniscient.
Since Timmler did not change or show any sign of abnormality, the human
scientists believed that the T device did not have any influence on him or his free agency.
The Martian scientists wondered how humans could reach such a conclusion. It was true
that Timmler was as normal as everyone else on earth, but being normal was by no means
a guarantee of having free will.
In order to point out this mistake to the human scientists, the Martians made
public a dark secret that all human beings (except Himmler and Jimmler) were like
Timmler because everyone’s brain was implanted a T device at birth. The human
scientists quickly confirmed the accuracy of the message because it was not difficult to
detect the device by using the newly invented Martian device detector, but they still could
not figure out what the T device’s function was. It could be the case that the T device
replaced the work that had previously been done by parts of Timmler’s brain, but as long
as it is a perfect replacement as if it were a successfully transplanted organ, we may still
consider Timmler’s brain the same as its original one.
A few days later, the Martians disclosed the intended function of the T device.
According to them, the T device was designed to deprive human beings of their free will.
It is supposed to be more vicious and harmful than the H and J devices. The H device
only takes one’s free will away occasionally. Although the J device takes one’s free will
away permanently, its influence can be easily discovered through one’s abnormal
behaviors. But the T device is different. It takes one’s free will away completely from

91
birth to death and accomplishes this task without showing even the slightest trace of it
because all human beings considered normal are in the same situation.
The human scientists were baffled by the second message from the Martians
because the T device obviously did not do anything. The Martians explained that had the
T device not been implanted in the human brains, humans would be able to live a life
with free will. But what would a life with free will be like and how would it differ from
the current human life? The Martians’ answer was, “Without being equipped with the T
device, human beings will be able to choose and do otherwise. For example, a child
without the T device who is eating vanilla ice cream could have chosen to eat chocolate
ice cream instead.” The human scientists asked the Martian scientists to identify such a
child between the two one of whom was accidentally left T-device-free at birth, but the
Martians could not tell without using their T device detector. This further confirmed the
human scientists’ belief that the T device does not make any difference at all. By telling
humans that the T device had deprived them of free will, Martians were either trying to
confuse them or themselves confused.
If we apply the technical terms used by the incompatibilist to the story, we may
say that Himmler was subject to a local manipulation and Jimmler a global manipulation.
Was Timmler manipulated? Agreeing with the Martian scientists, the incompatibilist
would say Timmler was also subject to a global manipulation the same way Jimmler was.
But the story has shown that this is not true because the Martian scientists could not
prove that the T device had any influence on Timmler.

92
Van Inwagen’s Martian Device Argument will be able to refute compatibilism
only if Timmler is also manipulated because in the argument the truth of determinism is
supposed to play the same role as the T device. However, the idea that T device impairs
one’s free will is either a lie or a sign of confusion.
Now let us consider if the Manipulation Argument can be used to support
incompatibilism. Since Pereboom has constructed a more ambitious version of the
Manipulation Argument than van Inwagen’s Martian Device Argument, I shall now turn
to examine Pereboom’s Four Case Manipulation Argument and try to show that it fails to
establish an incompatibilist view.

2. The Four Case Argument
Pereboom invites us to consider four cases in each of which Professor Plum
murdered Ms. White in the manner that we intuitively would not hold Plum responsible.
If our intuition is justified, determinism is not compatible with free will or moral
responsibility.
In case one, Professor Plum is locally manipulated by “radio-like technology”.
Case 1. Professor Plum was created by neuroscientists, who can
manipulate him directly through the use of radio-like technology, but he
is as much like an ordinary human being as is possible, given this
history. Suppose these neuroscientists “locally” manipulate him to
undertake the process of reasoning by which his desires are brought
about and modified – directly producing his every state from moment to
moment. The neuroscientists manipulate him by, among other things,
pushing a series of buttons just before he begins to reason about his
situation, thereby causing his reasoning process to be rationally egoistic.
Plum is not constrained to act in the sense that he does not act because
of an irresistible desire – the neuroscientists do not provide him with an
irresistible desire – and he does not think and act contrary to character

93
since he is often manipulated to be rationally egoistic. His effective
first-order desire to kill Ms. White conforms to his second-order desires.
Plum’s reasoning process exemplifies the various components of
moderate reasons-responsiveness. He is receptive to the relevant pattern
of reasons, and his reasoning process would have resulted in different
choices in some situations in which the egoistic reasons were otherwise.
At the same time, he is not exclusively rationally egoistic since he will
typically regulate his behavior by moral reasons when the egoistic
reasons are relatively weak – weaker than they are in the current
situation.
72

Plum’s action would seem to satisfy all the compatibilist conditions we
examined. But, intuitively, he is not morally responsible because he is
determined by the neuroscientists’ activities, which are beyond his
control. Consequently, it would seem that none of these compatibilist
conditions is sufficient for moral responsibility.
73

In this case, according to Pereboom, Plum is not responsible for murdering White
because “he is determined by the neuroscientists’ activities, which are beyond his
control”.
Case 2. Plum is like an ordinary human being, except that he was
created by neuroscientists, who, although they cannot control him
directly, have programmed him to weigh reasons for action so that he is
often but not exclusively rationally egoistic, with the result that in the
circumstances in which he now finds himself, he is causally determined
to undertake the moderately reasons-responsive process and to possess
the set of first- and second- order desires that results in his killing Ms.
White. He has the general ability to regulate his behavior by moral
reasons, but in these circumstances, the egoistic reasons are very
powerful, and accordingly he is causally determined to kill for these
reasons. Nevertheless, he does not act because of an irresistible desire.
74

72
Derek Pereboom, Living Without Free Will, 2001, pp.112-113.

73
Ibid.

74
Ibid., pp.113-114.

94
Pereboom believes that in this case Plum is not responsible because “his action is
determined by the neuroscientists’ programming, which is beyond his control”.
75

Case 3. Plum is an ordinary human being, except that he was
determined by the rigorous training practices of his home and
community so that he is often but not exclusively rationally egoistic.
His training took place at too early an age for him to have had the ability
to prevent or alter the practices that determined his character. In his
current circumstances, Plum is thereby caused to undertake the
moderately reasons-responsive process and to possess the first- and
second-order desires that result in his killing White. He has the general
ability to grasp, apply, and regulate his behavior by moral reasons, but
in these circumstances, the egoistic reasons are very powerful, and
hence the rigorous training practices of his upbringing deterministically
result in his act of murder. Nevertheless, he does not act because of an
irresistible desire.
76

According to Pereboom, Plum is not responsible in the third case because his
action of killing is determined by the rigorous training, which is beyond his control.
Case 4. Physicalist determinism is true, and Plum is an ordinary human
being, generated and raised under normal circumstances, who is often
but not exclusively rationally egoistic. Plum’s killing of White comes
about as a result of his undertaking the moderately reasons-responsive
process of deliberation, he exhibits the specified organization of first-
and second-order desires, and he does not act because of an irresistible
desire. He has the general ability to grasp, apply and regulate his
behavior by moral reasons, but in these circumstances the egoistic
reasons are very powerful, and together with background circumstances
they deterministically result in his act of murder.
77

According to Pereboom, there are no relevant differences between the fourth case
and the first three cases, so there would be no reason to think that Plum is responsible in

75
Ibid.

76
Ibid, p.114.

77
Ibid., p.115.

95
the last case if he does not seem to be so in the first three cases. If so, determinism is
incompatible with free will and moral responsibility.
Pereboom’s Four Case Argument can be reconstructed as follows:
(1) Plum1-3 are not responsible for killing White.
(2) The only thing Plum 1-3 have in common is that their actions are
manipulated actions which are caused by a deterministic causal process that
traces back to factors beyond Plum’s control.
(3) If determinism is true, then every act, including paradigmatically free and
responsible acts (i.e. Plum 4) are no different in any relevant respect from a
manipulated act, that is, are caused by deterministic factors beyond S’s
control.
(4) Therefore, if determinism is true, no one is ever responsible for performing
a paradigmatically free and responsible act.
I shall consider two typical replies to the argument first and then argue that the
Four Case Argument fails because it appeals to the idea of the special kind of global
manipulation which we have reasons not to trust.

3. The Soft-Line Reply and the Hard-Line Reply
In his article “A Hard-Line Reply to Pereboom’s Four-Case Manipulation
Argument”, McKenna reconstructs Pereboom’s Four Case Argument as follows
78
:

78
Michael McKenna, “A Hard-Line Reply to Pereboom’s Four-Case Manipulation Argument” in
Philosophy and Phenomenological Research, Vol. 77, pp.142-159.

96
1. If S is manipulated in manner X to A, then S does not A of her own free will
and is therefore not morally responsible for A’ing.
2. An agent manipulated in manner X to A is no different in any relevant
respect from any normally functioning agent determined to do A from CAS.
3. Therefore, if S is a normally functioning agent determined to A from CAS,
she does not A of her own free will and therefore is not morally responsible
for A’ing.
By “CAS”, McKenna means the Compatibilist-friendly Agential Structure. If a
person acts from the structure of CAS, he satisfies all the conditions that a compatibilist
would require for free will and moral responsibility. McKenna introduces the terms of
“soft-line reply” and “hard-line reply” according to which premise is being rejected. A
soft-line reply rejects premise 2 and a hard-line reply rejects premise 1. In other words, a
soft-line reply tries to find some differences between a manipulated agent and a
paradigmatically free and responsible agent in a deterministic world and a hard-line reply
denies that a manipulated agent does not act of her own free will or not morally
responsible for her actions.
In some cases it is possible to find differences between a free agent and a
manipulated one, but in others it is impossible. For example, in van Inwagen’s Martian
device case, all human beings are controlled by Martians, which means that no
paradigmatically free agents can be distinguished from the manipulated ones. So I agree
with McKenna’s judgment that
When faced with carefully crafted instances of MA, many thoughtful
compatibilists have tended toward the soft-line reply, denying whenever

97
possible that the incompatibilist has truly captured CAS. In my
estimation, this is the wrong compatibilist tendency. It leaves open an
easy incompatibilist rebuttal via a slight revision to the example so that
manner X gets right all that is required for CAS.
79

Since a more “carefully crafted instances” of Manipulation Argument will
eliminate all the differences between paradigm freedom and manipulation which will
make the argument immune to the soft-line reply, McKenna has developed a hard-line
reply to the Four Case Argument. He argues that it is not clear that Plum in any of them
is not free or morally responsible.
He proposes to fix our attention on “salient agential and moral properties”
80
by
reversing the order of Pereboom’s four cases and adding two additional ones. He begins
with Pereboom’s Case 4 and calls it Case 6 instead. He argues that since Pereboom
“cannot begin by presuming that determinism rules out free will and moral responsibility,
he cannot begin by denying that Plum is free and responsible”
81
in this case. Then he
moves to an additional case that is not Pereboom’s:
In Case 5, causal determinism is replaced by God, who foreknows each
of Plum’s acts and his entire life history. God brings about the entire
state of the world at every moment. God does this by “setting the entire
world in motion” in certain ways. Everything else unfolds instead just
as it would in Case 6.
82

79
Ibid., p.143.

80
Ibid., p.152.

81
Ibid., p.153.

82
Ibid., p.152.

98
Applying Pereboom’s generalization strategy to Case 5 would result in the
intuition that Plum in this case is also free and responsible because “it seems arbitrary to
make theological determination itself have a relevant difference here”.
83

Then he adds one more case, Case 4, which is not Pereboom’s:
In this case, god does not do the work, but a deity, Diana, zeroes in just
on the introduction of the zygote that will become Plum. Unlike God,
Diana cannot foresee the unfolding of every event in Plum’s world as
applied to every person, but she has just a glimmer of god’s knowledge,
enough that she can see just how the introduction of this zygote will
yield Plum, who will one day kill Ms. White.
84

Intuitively, Plum should be free and responsible in this case as well because “Case
4 is just a localized version of the knowledge possessed by God in Case 5” and “there is
no relevant difference as applied to these two cases.”
85

Cases 3 to1 are the same as Pereboom’s Cases 3 to 1, only presented in a reversed
order. McKenna believes that “Cases 3 and 2 just involve the same causal processes as
Case 4 and 5, but via dumb luck. And Case 1 is just the dumb luck played out over time
instead of as in Case 2, all up front at the beginning.”
86
So he concludes that “Pereboom
is not entitled to presume that in Case 1 Plum is not free or morally responsible. This is
so by the light of his own generalization strategy”.
87

83
Ibid., p.153.

84
Ibid., p.152.

85
Ibid., p.153.

86
Ibid.

87
Ibid.

99
Since Pereboom’s argument relies on our intuition about manipulative features, he
begins with a clear manipulation scenario and then slowly crosses the line between
manipulation and freedom. On the contrary, McKenna appeals to our intuition about
agential and moral properties, so he begins his series of cases with a paradigmatically
free and responsible deterministic agent. This is probably why McKenna calls the
situation a “dialectical stalemate”
88
because there is no obvious reason why we should
adopt one order of presenting the cases rather than the other. Apparently, which
conclusion one reaches depends on what intuition one has. It is true that McKenna’s
presentation of the cases may help bring out a compatibilist intuition in someone whose
intuition is ambiguous. But if a person has a strong compatibilist intuition, he may find it
natural to judge Plum as free and responsible at least in Case 4 even when the four cases
are presented in its original order. However, if one has an incompatibilist intuition, one
will certainly find Pereboom’s conclusion the most natural, probably even when the cases
are presented in the reversed order. McKenna is fully aware of this situation as he writes,
In my estimation, the disagreement between Pereboom and me plays out
in a dialectical stalemate. The intuitions that he seeks to elicit do not
demonstrably trump the ones that I seek to elicit. Given my treatment of
Cases 1 through 6, it is not clear that Plum in any of them is not free or
morally responsible.
89

Obviously, McKenna’s strategy does not help expose the real problem with
Pereboom’s argument. When Pereboom and McKenna appeal to our linguistic intuition,
they do not prove anything other than announce their own positions regarding the free
will problem. Pereboom speaks for incompatibilists and claims, “This is our intuition”

88
Ibid., p.154.

89
Ibid.

100
and McKenna answers on behalf of compatibilists, “Our intuition is different.” It seems
to me that the only way to break the stalemate is to clearly demonstrate what goes wrong
with the Four Case Argument. I believe this can be done by matching the Plum cases
onto my Himmler, Jimmler and Timmler case.
According to Pereboom’s vague description of Plum 1 that the “neuroscientists
manipulate him by…pushing a series of buttons just before he begins to reason about his
situation, thereby causing his reasoning process to be rationally egoistic” but “he is as
much like an ordinary human being as is possible”
90
, Plum 1 is both manipulated and like
an ordinary person. Since detailed evidence of the existence of manipulation is absent,
Plum 1 can be seen as corresponding to either Himmler or Timmler. If we are able to
spot some weird moments when we compare him in these moments with his normal state
or with other people, then he is locally manipulated like Himmler. The role of the H
device is played by the neuroscientist’s push on the buttons of a radio-like equipment.
However, if Plum 1 is only “rationally egoistic” to a reasonable degree as we often see in
any other egoistic individual, then as in Timmler’s case there would be no reason to think
that Plum 1’s egoistic characteristic is a result of a manipulation.
Both Plum 2 and Plum 3 correspond to Jimmler because in the second and third
cases Plum is subject to a global manipulation even though the causes of the
manipulation are different. Plum 2’s manipulation is created by the neuroscientists and
Plum 3’s manipulation is produced by training and brainwashing. Although Pereboom
claims that Plum 2 and 3 are manipulated, he falls short in describing the details of the

90
Pereboom, pp.112-113.

101
result of the manipulation. Simply stating that a certain act is caused by the
neuroscientist or the brainwasher does not necessarily make it the case because the
Martian scientists are false in making the claim that Timmler is manipulated by their T
device.
Pereboom’s description of the four cases is ambiguous. For example, in Case 2, it
is unclear why Pereboom says that Plum is “causally determined to kill for these
reasons”
91
rather than “he kills for egoistic reasons”. There are no details available to
show that Plum’s act of killing is causally determined. Pereboom believes that in this
case Plum is not responsible because “his action is determined by the neuroscientists’
programming, which is beyond his control”
92
. But as told by Pereboom himself, Plum’s
action of killing is not completely determined by the programming.
In Case 3, the rigorous training is only part of the reason why Plum committed
murder. The other part is his own choice. When he grew out of the training phase and
left the community, it is unreasonable to say that he absolutely had no opportunity to alter
(at least part of) his egoistic character. Again, there are no details available which could
show that Plum 3 is successfully brainwashed.
Since Pereboom is being vague about case 2 and case 3, they can be interpreted in
two ways. One is to see Plum 2 and Plum 3 as Jimmler who is covertly manipulated and
the other is to see them as Timmler who is not manipulated at all. This is why the Four
Case Argument is susceptible to either the soft-line reply or the hard-line reply. The soft-
liners read Pereboom in the first way and the hard-liners read Pereboom in the second

91
Ibid., p.114.

92
Ibid.

102
way. In order to establish his thesis, Pereboom certainly needs to show that Plum 2 and
Plum 3 are covertly manipulated, otherwise, the analogy between his first three cases and
Case 4 will collapse.
Even if it is granted that both Plum 2 and 3 are manipulated and thus are not
responsible for their actions, the huge gap between Plum 4 and Plum 1-3 still cannot be
bridged because like Timmler Plum 4 is by no means manipulated. In Case 4 everything
is perfectly normal with Plum, there would be no reason to think that the egoistic reasons
and the background circumstances “deterministically” result in Plum’s act of murder. It
would be more reasonable to say that Plum freely chose to murder White under that
particular circumstance.
Pereboom may respond to this objection by pointing out that as part of his thought
experiment it has been stipulated that the egoistic reasons and the background
circumstances deterministically result in Plum’s act of murder. I hope my Timmler case
has shown that some stipulations in philosopher’s thought experiments are illegitimate
because they contain logical defect. While the T device does not play any role in forming
Timmler’s desires, choices and behaviors, the Martian scientists stipulate that it does.
This is like constructing a thought experiment where you describe a seven-year-old boy
who goes to school, participates in class, plays with friends, watches TV and goes to bed
the same way any other normal and healthy kid goes through a day but meanwhile
stipulate that this boy suffers unbearable toothache. When told this way, the story
contains an internal conflict which makes it unintelligible. Similarly, we cannot have a
meaningful thought experiment where a person works hard everyday but is stipulated as

103
lazy or a chair is described as green all over but meanwhile stipulated as red all over.
The problem with these thought experiments is not empirical but logical, that is, they all
contain an internal contradiction which renders them unintelligible.
Pereboom does not seem to be aware of the difference between the Himmler case
and the Timmler case, he remarks,
One distinguishing feature of Case 4 is that the causal determination of
Plum’s crime is not, in the last analysis, brought about by other agents.
However, the claim that this is a relevant difference is implausible.
Imagine a further case that is exactly the same as, say, Case 1 or Case 2,
except that Plum’s states are induced by a machine that is generated
spontaneously, without intelligent design. Would he then be morally
responsible? The compatibilist might agree that this sort of machine
induction is responsibility-undermining as well, and then devise a
condition that stipulates that agents are not responsible for actions
manipulated by agents or machines. But this move is patently ad hoc.
93

On the contrary to what Pereboom believes, it is certainly important to note that in
Case 4 Plum’s crime is not brought about either by other agents or machines because this
fact is not only relevant but also makes important differences and not in a ad hoc way. If
Plum’s states are induced by a machine in Case 1 or 2, we are still able to recognize that
Plum 1 or 2 is different from free agents, so the contrast between manipulation and
freedom remains. But when Pereboom claims that in a deterministic world no one is
different in any relevant way from someone who is manipulated, the concept of
manipulation has lost its contrast and thus does not mean what it ordinarily means. As a
result, there will be no such thing as acting freely. But this is not the case in Case 1, 2 or
3. In the first three cases, there is such thing as having free will or acting freely and it
makes sense to say that Plum is manipulated because a life with freedom is conceivably

93
Ibid, pp.115-116.

104
different. In other words, the distinction between free and manipulated actions remains
untouched. But in Case 4, the distinction collapses.
In response to this, the incompatibilist may contend that it does not follow from
Plum 4 that the concept of manipulation has lost its contrast because we might have been
using an incompatibilist (i.e. libertarian) standard as the contrast. We might have
assumed that we sometimes act with free will because we think that determinism is false.
If we think carefully along this line, we will find that whenever we assume that
determinism is false, we are dealing with an act that is free in the ordinary sense. In other
words, the incompatibilist standard understood as such coincides with our ordinary
standard and therefore has nothing to do with the actual truth or falsity of determinism.
Both van Inwagen’s Martian Device Argument and Pereboom’s Four Case
Argument invite us to imagine our current life to be a manipulated life controlled by
some hidden manipulators. But this picture is deeply confused because there is no such
thing. If there really are such manipulators, then we are either as free as they are or we
are free in a different sense from the sense in which they are.
The Four Case Argument fails because it mistakenly assumes that it makes sense
to stipulate that an agent is manipulated when he thinks and acts the same way a free
agent does or that every normal human being under normal circumstances is manipulated.
This is where the idea of the special global manipulation should be found problematic.

105
4. Mele’s Criticism
Let me repeat my reconstruction of Pereboom’s Four Case Argument:
(1) Plum1-3 are not responsible for killing White.
(2) The only thing Plum 1-3 have in common is that their actions are
manipulated actions which are caused by a deterministic causal process that
traces back to factors beyond Plum’s control.
(3) If determinism is true, then every act, including paradigmatically free and
responsible acts (i.e. Plum 4) are no different in any relevant respect from a
manipulated act, that is, are caused by deterministic factors beyond S’s
control.
(4) Therefore, if determinism is true, no one is ever responsible for performing
a paradigmatically free and responsible act.
Mele believes that Pereboom’s argument fails because it has misidentified the
reason why one’s intuition may be that Plum is not responsible for killing White.
94

Pereboom believes the reason is that the action results from a deterministic causal process
that traces back to factors beyond Plum’s control. Mele tests it by inviting us to imagine
indeterministic analogues of Cases 1 to 3. The indeterministic control exercised by the
neuroscientists over Plum works just like the deterministic control “except that there is a
tiny chance every few seconds that the program will incapacitate Plum”. The result of the
test is that the deterministic aspect of the cases is dispensable and it is the manipulation
that renders Plum’s act of killing unfree. As no manipulation can be found in Case 4,

94
Alfred Mele, Free Will and Luck, p.140.

106
Pereboom fails to show that (3) is true given the falsity of (2), that is, when the
manipulated action does not result from a deterministic but an indeterministic control.
Mele’s argument does point out a problem in Pereboom’s argument, but it seems
to me that the problem can be easily fixed. If we include indeterminism in (1) and (2),
the conclusion still follows:
(5) Plum1-3 are not responsible for killing White.
(6) The only thing Plum 1-3 have in common is that their actions are
manipulated actions which are caused by a deterministic or indeterministic
causal process that traces back to factors beyond Plum’s control.
(7) If determinism or indeterminism is true, then every act, including
paradigmatically free and responsible acts (i.e. Plum 4) are no different in
any relevant respect from a manipulated act, that is, are caused by
deterministic factors beyond S’s control.
(8) Therefore, if determinism is true, no one is ever responsible for performing
a paradigmatically free and responsible act.
This is probably the most obvious way to respond to Mele’s criticism on behalf of
Pereboom, but the new version of the argument makes a much stronger claim than the
original one because it supports the impossibilist view that both determinism and
indeterminism make free will and moral responsibility impossible.

107
5. Confusions Involved in Kane’s Idea of CNC Control
As discussed earlier, a regular kind of CNC control can be distinguished from a
special kind of it. If an agent is subject to a regular CNC control, he may be
brainwashed. For example, the members of the People’s Temple were manipulated by
their leader Jim Jones to the extent that they were willing to do anything Jim Jones
ordered them to do including killing their own children and then taking their own lives.
But when an agent is subject to a special CNC control, he is as normal as anyone else.
Kane believes that the idea of CNC control “poses problems for all compatibilist
views of free agency”
95
. I will try to show that what “poses problems” for compatibilism
is not the regular CNC control, but the special CNC control. The special CNC control
may appear to be a threat to compatibilism because the idea itself contains a logical
defect. I will try to expose the confusions involved in it.
Kane distinguishes between hard compatibilism and soft compatibilism. The hard
compatibilist denies that CNC control takes away any freedom and the soft compatibilist
believes that a free action in a deterministic world is different from a manipulated action.
He uses Skinner and Hobbes as examples of hard compatibilists. Skinner thinks that we
should not worry “about being covertly controlled or behaviorally engineered if it makes
us happy” and Hobbes believes that even though “God has predestined us” it does not
take away any freedom “that we could otherwise have had in any case”.
96

It seems to me that Kane overlooks an important difference between Skinner’s
view and Hobbes’. In the case of Skinner’s, some human beings are CNC controlled by

95
Robert Kane, The Significance of Free Will, p.67.

96
Ibid., p.67.

108
some other human beings, so a controlled person can be distinguished from a free person.
When we consider a paradigm case of brainwashing, say, the People’s Temple, we may
make the judgment that most members of the People’s Temple were brainwashed by Jim
Jones because it is easy to see how much their choices and actions deviate from a normal
person’s. In this case, the concept of manipulation is used to draw the distinction
between a normal person’s life and a cult member’s life. However, in the case of
Hobbes’, no such distinction can be located. We may wonder if it is morally acceptable to
manipulate others or if we should fight against other’s manipulation exercised on us, but
it does not make any sense to complain that we are manipulated by God or try to stop
God from doing it. In order to make sense of it, we need to be able to figure out what a
human life would be like without God’s manipulation and in what way our current life
deviates from it. Apparently, the concept of manipulation in the claim that all human
beings are manipulated by God has lost its contrast and thus no longer retains its ordinary
meaning.
Moreover, Kane’s discussion of soft compatibilism involves some confusion. He
writes,
They [the soft compatibilists] must emphasize the distinction between
CNC control by other purposeful agents and mere determination by
natural causes (without purposeful control by other agents); and they
must argue that while CNC control takes away freedom in a significant
sense, mere determination by natural causes does not do so. The
problem is to locate the relevant difference between the two that makes
one of them (CNC control) objectionable and the other (mere
determination) not.
97

97
Ibid., p.68.

109
…our powers or freedoms are equally impaired by CNC control as by
mere determination. Your power or freedom to run or dance is no less
impaired if you are paralyzed by natural causes than if some other agent
is holding you down. In each case, the significant thing is that you
cannot do something you want to do.
98

The soft compatibilist finds the CNC control caused by other purposeful agents
objectionable but mere determination caused by determinism not. Kane argues that this
attitude is arbitrary because the victim whose freedom (i.e. power to walk) is taken away
suffers the same effect (i.e. unable to walk) in these two cases even though the causes of
the suffering are different. In other words, so far as freedom and moral responsibility are
concerned, what matters is what a person is like at the time of action, not how he got to
be that way.
I will try to show that one’s freedom (i.e. power to walk) can be taken away by
purposeful agents or natural causes, but cannot be taken away by the truth of
determinism. The reason why the soft compatibilist finds the CNC control caused by
other purposeful agents or natural causes objectionable but mere determination caused by
determinism not is because a person can never lose his freedom to the truth of
determinism.
Let us consider an example. Three beautiful sisters Nadia, Olga and Perla were
all talented ballet dancers. One day their parents received terrible news about Nadia and
Olga. The city where they were suffered a major earthquake and a terrorist attack within
a few hours. Nadia was hurt by a falling tree and lost her ability to walk. Olga was
kidnapped by the terrorists and then rescued in a few hours by the police, but she was
paralyzed as a result of the torture by the terrorists. When the terrified parents flew to the

98
Ibid.

110
city, both Nadia and Olga were being treated in the hospital. When they opened the door
of the girls’ room, they could not hold back their tears. Their two pale daughters were
both helplessly lying in bed, once shinning stars dancing on the stage and now neither
could walk. At this point, Perla called. She told her parents that she was paralyzed too.
The poor parents could hardly breathe until she explained, “Unlike my sisters, my
paralysis was not caused by someone else or a natural disaster. I was paralyzed by
determinism.” The puzzling parents were still worried, “What do you mean?” Perla
continued, “It is in the news that scientists have proved that determinism is true. So no
one is able to act freely any more. My ability to walk or dance is taken away by
determinism!” “What are you dong right now?” asked the parents who became less
anxious. “I am in a shopping mall trying to find some nice gifts for Nadia and Olga”
answered Perla, “I will go to see them next week.”
The example was designed to compare three cases of paralysis caused by three
different causes. Nadia’s paralysis was caused by a natural cause; Olga’s was caused by
some purposeful agents and Perla’s was caused by determinism. I hope the example has
shown that when Perla said she was paralyzed, she did not use the word “paralyze” in its
ordinary sense. If we use the same word to mean the same thing, then both Nadia and
Olga were paralyzed but Perla was not because one can be paralyzed by purposeful
agents or natural causes but cannot be paralyzed by determinism. A mere determination
by natural causes is not an instance of the mere determination by determinism. But Kane
has confused the two by thinking that the soft compatibilist draws the distinction between
the “CNC control by other purposeful agents and mere determination by natural causes

111
(without purposeful control by other agents)” when the distinction is actually between the
“CNC control by other purposeful agents” and determinism.
The following example further shows Kane’s confusion:
If the distinction between CC control and mere determination by natural
causes does not make a difference in our powers, when the results are
the same, why should the distinction between CNC control and mere
determination by natural causes make a difference in our powers, when
the results are also the same? Imagine a possible world W
1
in which
every aspect of persons’ lives is controlled by invisible gods or spirits.
Imagine another world W
2
in which these persons’ lives are exactly the
same as in W
1
in every detail of thought, belief, and circumstance from
birth to death except that in W
2
everything is produced by natural causes
rather than by gods or spirits. … In order to find a relevant difference
between CNC control and mere determination, we have to look beyond
the powers that each might contingently take away, since any contingent
power the one might take away, the other might take away as well.
99

In this passage Kane assumes that our power (i.e. ability to walk) can be equally
destroyed by invisible gods or spirits, determinism, or natural causes. I hope the story of
the three sisters has shown that the mere determination by natural causes and the mere
determination by determinism cannot have the same impact on one’s freedom, that is, the
“results” can never be “the same” even though Kane has stipulated that they are.
His example of the possible world W
1
and the possible world W
2
makes a
comparison between the impact of the invisible gods or spirits on one’s freedom and the
impact of determinism on one’s freedom. The fact that persons’ lives in W
1
are the same
as persons’ lives in W
2
shows that the control exercised by the invisible gods or spirits is
a special CNC control, namely, as in the case of Timmler, it is not a control at all. If we
imagine that people in W
1
are wanton-like and can be discriminated from people in W
2
,
then their lives will be different from persons’ lives in W
2
. In other words, if the CNC

99
Ibid.

112
control in W
1
is a regular control, then the life style in W
1
cannot be the same as in W
2
; if
the CNC control in W
1
is a special control, the life style in W
1
is the same as in W
2
, but
then a control as such is not a control at all. Kane attempts to combine a regular CNC
control with a normal human life without realizing that this is a mission logically
impossible.

6. Conclusion
The Manipulation Argument is used by van Inwagen to refute compatibilism and
by Pereboom to justify incompatibilism. The compatibilists have advanced two types of
replies to it – the soft-line reply and the hard-line reply. Although each reply has
correctly identified a weakness of the argument, neither is able to expose the central
problem with the argument. I have tried to argue that the central problem with the
Manipulation Argument is that it relies on the idea of the special kind of global
manipulation which is an idea containing serious logical defects.

113

Chapter Five: The Paradigm Case Argument for Compatibilism

Paradigm Case Argument (PCA) is a form of argument that was popular in the
40’s and 50’s of the 20th century in analytic philosophy. It was constructed by
philosophers influenced by G. E. Moore and Wittgenstein. They tried to use PCA to
refute philosophical skepticism and defend concepts and expressions in ordinary
language.
Philosophical skepticism has been an important theme in the tradition of Western
philosophy. The skeptic typically argues that there is no way for us to know for certain
that the external world, other mind, physical objects, free will, solidity, knowledge,
certainty, clarity, color, induction, or time exists. Philosophers who wish to defend
common sense have come up with different ways to answer or refute these skeptical
doubts or denials. PCA is one of the achievements found in the contemporary
philosophical landscape. It received much attention in about two decades before it was
thought to be damaged by its opponents’ criticisms beyond repair. PCA disappeared ever
since and the philosophers who are determined to refute skepticism began to look for
other ways to accomplish the task. However, it seems to me that the advocates and
friends of PCA gave up too quickly. Like other arguments, PCA can easily be
misunderstood. The reason that many criticisms of it sound convincing is because the
critics picked a straw man to attack.
In this chapter, I shall discuss two versions of PCA – PCA as Ostensive Teaching
and PCA as Elucidation – and try to defend the second version. When properly

114
constructed and understood, PCA is not only useful but also very effective in refuting
skepticism with its various concerns, especially in the debate of the free will problem.

1. PCA as Ostensive Teaching
The most popular and straightforward version of PCA is to consider it as closely
related to teaching concepts through standard examples. This is also what most people
have in mind whenever Paradigm Case Argument is mentioned today.
Back in 1950’s, Robert Richman claimed that PCA is an inconclusive argument
and “of little philosophical importance”
100
based on his characterization of PCA as an
argument that
…appears to be closely related to the doctrine of ostensive definition. Its
nature may be indicated as follows: The meaning of some term “X” is
learned by being shown things t1, t2, etc., which have the property X.
These things, t1, t2, etc., are then standard examples, or paradigm cases of
X. Since this is so, it is absurd to question whether, say, t1 is an X, or
whether anything is an X.
101

According to Richman’s characterization, PCA is an argument that since we use
standard examples t1, t2, etc. to learn the meaning of term “X”, it would be absurd to
question or deny that anything is an X.
J. W. N. Watkins explains PCA in a similar way:

100
Robert Richman, “On the Argument of the Paradigm Case”, in Australasian Journal of Philosophy, Vol.
39, 1961, p.75.

101
Ibid., p.76.

115
Suppose that a metaphysician advances the thesis that nothing is really p,
where “p” is some expression in ordinary language, such as “acting of his
own free-will.” Our metaphysician obviously regards his thesis, and
therefore “p” itself, as meaningful. But, the linguistic philosopher will
say, “p” has a meaning precisely because there are standard situations to
which it is applied, because paradigm cases do exist which give it its
meaning.
102

According to Watkins’ explanation, the advocates of PCA hold that the meaning
of some expression “p” is given by standard situations to which “p” is applied, so there
must exist something that is p.
PCA as Ostensive Teaching can be reconstructed as follows:
(1) “X” is meaningful.
(2) The meaning of “X” must be learned by watching the paradigm examples of
X.
(3) Therefore, “X” must be instantiated.
For example, if “X” is “red”, in order to show that the skeptic’s claim “Nothing is
really red” is false, the advocate of PCA argues that since “we teach children the meaning
of ‘red’ by pointing to pillar-boxes”
103
, it is absurd to deny that there are red things. In
other words, if the skeptic agrees that “red” is meaningful and the meaning of it cannot be
learned unless through looking at red objects such as pillar-boxes, “Nothing is red” must
be false.

102
J. W. N. Watkins, “Farewell to the Paradigm-Case Argument”, Analysis, Vol.18, p.26.

103
Ibid., p.25.

116
Inspired by the construction of PCA and its promise to refute skepticism, Antony
Flew remarks that “The clue to the whole business now seems to lie in mastering what
has recently been usefully named, The Argument of the Paradigm Case”
104
. It seems to
him that PCA can be used to solve all kinds of philosophical problems.
Flew is the first to apply PCA to the free will problem in order to prove the
existence of free will. The skeptical thesis in the free will debate is that no action is ever
free if the world is deterministic. Flew contends that we learn the words and phrases
such as “acted freely”, “did it of his own free will”, “could have done otherwise”, “had a
choice about what she did”, “had alternatives” and “could have helped doing what he
did” “by watching people apply them in concrete situations in everyday life, just as we
learn, for example, colour words”.
105
The concrete situation where someone acts freely
may be something like this: a young man married of his own free will in the sense that he
was in love with his bride and was under no constraint or duress when decided to marry
her. Flew presents this example as a clear and uncontroversial case to which the concept
of free will is correctly applied. If these words apply to something, they must apply to
some standard cases. The central thesis of the first version of PCA can be summarized as
that the meaningfulness of an expression implies its instantiation in the world.

104
Antony Flew, edited, Essays in Conceptual Analysis, 1956, p.19.

105
This is Flew’s argument reconstructed by Peter Van Inwagen. See Flew’s “Divine Omnipotence and
Human Freedom” in New Essays in Philosophical Theology and Van Inwagen’s An Essay on Free Will,
p.107.

117
(1) Objection: Examples May Appear to Be X
A challenge to premise (2) of PCA is to question whether we can learn the
meaning of “X” through examples that are not X but only appear to be X. For example, I
point to an orange balloon and tell a small child “This is red” and then point to an orange
block and repeat “This is red”. Is it possible that one day the child will learn how to use
the word “red” correctly without ever seeing any red objects? If it is possible, it would
not be hard to imagine that most competent speakers of English language are in the same
situation. If so, PCA only proves that there must exist something that appears to be X
rather than is X.
The friend of PCA may rebut that to learn how to use “red” this way is impossible
because the child will certainly mistake orange objects as red objects. Anyone who
consistently makes such mistakes does not understand what “red” means.
Moreover, when we say something appears to be red, we imply that there are
things that are genuinely red. Otherwise, the concept of appearance in the expression “It
appears to be red” would not have an ordinary meaning.

(2) Objection: the Applicability of PCA
Watkins criticizes PCA as an argument that “appears to be a grandiose extension”
of “ontological arguments” which purports to infer from the “meaningfulness of a
concept to the existence of instances of it.”
106
For example, PCA can be used to prove

106
J. W. N. Watkins, “Farewell to the Paradigm-Case Argument”, Analysis, Vol.18, p.28.

118
the existence of instances of “miracle”, “unicorn” or “mermaid” given the
meaningfulness of these words.
Richman raises a similar objection. According to him, it makes perfect sense to
say that “There are no red things”, so the meaningfulness of “red” cannot entail the
existence of red things.
107
But what if we must learn the meaning of “red” by looking at
red objects? PCA requires that the terms or expressions in question must be learned by
ostensive definition. But are there such expressions at all? If there are, what are they?
In response to the objection, some advocates of PCA put restrictions on its use –
PCA can only be used to prove the existence of the expressions that must be learned
ostensively.
Flew tried to identify which concepts belong to the group of terms that are learned
ostensively only. He believes that simple concepts such as “yellow”, “wet” or “acting of
his own free will”, unlike “miracle” or “vacuum”
108
, is learned by watching yellow or wet
objects or an instance of acting of one’s own free will. By contrast, compound
descriptions are learned differently. For example, in order to understand the expression
“the first man to run a four-minute mile”
109
, we don’t have to look at that person first.
Rather, we learn the meaning of the expression first and then look around the world to
see whether anyone fits the description.
Watkins finds Flew’s distinctions arbitrary. He argues,

107
Robert Richman, “Still More on the Argument of the Paradigm Case”, Australasian Journal of
Philosophy, Vol.40, p.204.

108
Flew, Ibid., p.29.

109
Flew, New Essays in Philosophical Theology, p.150.

119
Why should we accept Flew’s unargued contention that “acting of his
own freewill” is like “yellow” and “wet” and unlike “miracle” and
“vacuum” in having to be defined ostensively by reference to
paradigm cases?
110

Watkins’ criticism seems to me right on the point. There is no obvious reason
why “acting of his own free will” is like “yellow” and “wet” but unlike “miracle” and
“vacuum”. Neither is it clear why “acting of his own free will” is a simple concept rather
than a compound description. Moreover, it is not obviously true that “yellow” or “wet”
has to be learned ostensively. Why cannot they be learned through descriptions? If they
cannot, how is it possible that blind people (born as blind) could master color concepts.
It seems to me that to choose to go down this path PCA is heading a wrong direction.
If it can be shown that the terms in question cannot be taught through paradigm
instances, PCA will turn out to be powerless to refute skepticism. It seems that, as
Watkins points out, many descriptive words cannot be taught merely ostensively, such as
“freewill”, “cause”, “law”, “unpredictable”, “novelty”, or “creativity”.
111

(3) Objection: The Meaning of “X” May Be Innate
Richman challenges the second premise of PCA that “the meaning of ‘X’ must be
learned by watching the standard examples of X”. He argues that PCA appeals to a
psychological fact rather than a logical fact because it tries to prove the existence of X by
an appeal to the fact of how we learn the meaning of “X”. It is logically possible that the

110
Watkins, Ibid., p.30.

111
Watkins, Ibid., p.32.

120
meanings of words are innate rather than learned. For example, it is logically possible
that “we know the meaning of ‘red’ (say, that we have the ability to apply the term
correctly) without learning the meaning (the ability being innate), or that we have the
ability as a result of alterations in our nervous system brought about, e.g., by intravenous
injections of certain chemicals.”
112
If, as Richman suggests, the learning process is not
necessary, then from the fact that “X” is meaningful we cannot infer that “X” must be
instantiated.
Richman’s objection to premise (2) seems to me valid. If we don’t have to teach
or learn the meaning of “X” by pointing to instantiations of “X”, then the fact that we
understand “X” does not necessarily imply that “X” is instantiated in the world. In other
words, it is possible that nothing is X.

(4) Normative vs. Empirical
When constructing a PCA, we point to something and utter “This is X”, are we
defining “X” ostensively or making an empirical judgment that this object is X? The
former makes “This is X” a normative statement and the latter renders it empirical.
Since Watkins, as well as many others, takes the skeptical thesis as denying the
existence of the instantiations of “X”, a successful refutation of skepticism ought to show
that “X” is necessarily instantiated in the world. So only the latter understanding of “This
is X” is able to respond to the skeptical thesis.

112
Robert Richman, “Still More on the Argument of the Paradigm Case”, in Australasian Journal of
Philosophy, Vol.40, pp.204-205.

121
For example, if someone thinks that dinosaurs do not exist and you try to prove
their existence by pointing to a crocodile in the pool and defining it as “dinosaur”, then
your utterance “This is a dinosaur” is begging the question and thus cannot prove
anything.
A PCA defender may rebut that it is one thing to define a concept in accordance
with our daily linguistic practice, and quite another to define it at whim. The philosopher
who tries to use PCA to refute skepticism is not like the person who defines “dinosaur”
by pointing to a crocodile; he is like someone who defines “red” by pointing to a red
balloon. The first person proves nothing, but the second proves the instantiation of “red”.
Watkins believes that when “This is X” is considered as an empirical statement
that describes what an object is, PCA will face another difficulty, that is, the example
pointed to may or may not be X, therefore “X” is not necessarily instantiated. But “the
paradigm case argument requires that they should be unquestionable (necessary)
exemplifications and not merely prima facie illustrations.”
113

Richman expresses a similar worry. If PCA makes an empirical description that
something is X, then it involves a “valuational leap” from an actual application of “X” to
the correct application of “X”. The example is described as an X, but it is not necessarily
X.
In his response to Watkins, Flew admits that “I obscured an important distinction:
between defining ostensively in terms of particular specific instances; and explaining by

113
Watkins, Ibid., p.31.

122
reference to examples which have to satisfy a certain general specification.”
114
The
options that Flew believes that he has conflated are seeing “This is X” as normative and
as empirical. He decides to abandon the former and adopt the latter. In making such a
move, Watkins comments, Flew has abandoned the fundamental feature of PCA, namely,
the logical priority of the paradigm cases. Watkins writes,
If we have the specification there is really no need to interpose
references to examples of it: we can refer directly to the specification
itself. This is important. Instead of beginning with paradigm objects
we now begin with a theoretical formula or specification.
115

The idea is that if examples of X are applications of “X” which may be correct or
incorrect according to the specification of the meaning of “X”, the examples will be
rendered obsolete. When a “theoretical formula or specification” of a concept is logically
prior to “paradigm objects”, the instantiations of the expressions in question are not
necessary, therefore, PCA as Ostensive Teaching fails in its attempt to refute skepticism.

2. PCA as Elucidation
All opponents and most proponents of PCA consider the first version of PCA,
namely, PCA as Ostensive Teaching to be a failure because it fails to establish that the
concepts under the skeptical pressure necessarily have instantiations in the world. When
the skeptic challenges our common sense by claiming that “Nothing is X”, we cannot use
PCA to show that something must be X. At this point, it looks like PCA is both useless
and hopeless. But I shall try to argue that it is too quick to jump to this conclusion

114
Flew, “ ‘Farewell to the Paradigm-Case Argument’: A Comment”, in Analysis, Vol.18, p.37.

115
Watkins, “A Reply to Professor Flew’s Comment”, in Analysis, Vol.18, p.42.

123
because the goal of refuting skepticism cannot be reached if the skeptical challenge is not
properly understood or the argument of PCA is not correctly formulated.
Oswald Hanfling has made a great contribution to defend PCA. He agrees with
the critics of PCA that from the mere fact that “X” is meaningful we cannot infer that
“X” is instantiated in the world, but, as he points out, “the essential aim of the PCA is not
to deduce the existence of something from facts about learning; it is to establish the
correct description of situations whose existence is not in dispute.”
116

Hanfling rejects the version of PCA as Ostensive Teaching. He believes that PCA
thus understood is indefensible because “the kind of teaching that is relevant to the
Paradigm Case Argument is not by showing but by description”. The meaning of “X” is
explained by a description of X as X is a standard example to which “X” is
paradigmatically applied. The function of PCA is to draw the skeptic’s attention to these
examples and force them to see the fact that “if the word is not applicable here, then it is
not applicable anywhere”
117
. In other words, PCA describes a situation to which “X” is
applied and the description of the situation is standard and not in dispute.
I shall argue that PCA as Elucidation is the correct version of PCA because it
avoids all the problems with the first version and is a powerful weapon against
skepticism.

(1) Empirical Dispute vs. Linguistic Dispute

116
Oswald Hanfling, “What is Wrong with the Paradigm Case Argument?”, p.25.

117
Ibid., p.24.

124
Before one can refute skepticism, one must understand their challenge correctly.
When the skeptic claims that “Free will does not exist”, “Solidity does not exist” or
“Nothing is red”, they certainly do not mean to deny that “free will”, “solidity” or “red”
are meaningful terms. What they deny, according to them, are instantiations of these
terms in the world. Even though everyone except the skeptic agrees that we have free
will, there are solid objects and the balloon is red, could it be the case that we are all
wrong and only skeptics get things right? After all, some years ago everybody was
wrong in believing that the earth is flat. We could be in the same situation regarding free
will, solidity or redness.
To address this worry, Norman Malcolm draws a distinction between two ways of
being mistaken.
118
One way of being mistaken is to be “mistaken about the facts” and the
other way of being mistaken is to use language incorrectly.
When people were mistaken about the shape of the earth, they made a mistake
about empirical facts. So the dispute between one person who thinks that the earth is flat
and the other who thinks that the earth is round is an empirical dispute. In other words,
they disagree on facts. But the dispute between the skeptic and an ordinary person is not
empirical. What the skeptic discovers about the objects in question are exactly what the
ordinary person finds about them, however, they disagree on how to describe them.
Malcolm offers an example to illustrate the distinction:
Two people are looking at an animal; they have a clear, close-up view
of it. Their descriptions of the animal are in perfect agreement. Yet
one of them says it is a fox, the other says it is a wolf. Their
disagreement could be called linguistic.

118
See Norman Malcolm, “Moore and Ordinary Language”.

125
Now suppose that one who says it is a wolf, not only agrees with the
other man as to what the characteristics of the animals are, but
furthermore agrees that that sort of animal is ordinarily called a fox. If
he were to continue to insist that it is a wolf, we can see how absurd
would be his position. He would be saying that, although the other
man was using an expression to describe a certain situation which was
the expression ordinarily employed to describe that sort of situation,
nevertheless the other man was using incorrect language. What makes
his statement absurd is that ordinary language is correct language.
119

This is a good example to illustrate what a linguistic dispute is, but it is not
strictly analogous to the disagreement between the skeptic and the ordinary man. I shall
revise it into a better analogy: the skeptic and the ordinary man see the same facts about
the same animal and agree that the animal is usually called “fox”, but when the ordinary
man says “There is a fox in front of us”, the skeptic corrects him, “I knew you would call
it ‘fox’, but it is not really a fox. Fox does not exist because foxes must reproduce three
times a year.” The ordinary man replies, “But the meaning of ‘fox’ never contains that
requirement.” The skeptic answers, “I know, but it should.”
The skeptic may continue to reject my revised version of the example because he
may find dissimilarities between the fox case and the free will debate. He maintains that
“Free will does not exist” is a statement about an empirical fact. If determinism is true
every event has a complete causal explanation and this fact makes it false that one can act
with free will, so the problem is empirical rather than linguistic or conceptual.
Assuming the skeptic is right, I shall consider a simple example to illustrate the
nature of the dispute between the ordinary man and the skeptic. Quincy raised his hand
to wave to Raissa when he saw her walking towards him on the street; Raissa’s hand was

119
Ibid., pp.117-118.

126
waved too but it was grabbed by her friend next to her and forced to wave. The ordinary
man will say that Quincy acted freely or waved of his own free will while Raissa did not.
However, the skeptic will say that neither Quincy nor Raissa waved freely if determinism
is true. Do the ordinary man and the skeptic agree on all the facts? Yes, they do. They
not only agree on what they saw and heard about Quincy and Raissa at the moment of
waving, but also agree on what can possibly be found out by any empirical investigation
including the truth or falsity of determinism. Assuming that determinism is found to be
true, just like the skeptic, the ordinary man understands that both Quincy and Raissa’s
bodily movements and emotions can be causally explained. And, just like the ordinary
man, the skeptic is perfectly aware of the external force that Raissa’s hand was subject to.
Apparently, their disagreement is not empirical. What they disagree on is whether to
apply the concept of free will to Quincy’s action of waving. The ordinary man insists on
doing what we ordinarily do, that is, describing Quincy’s act as free and Raissa’s unfree.
However, the skeptic refuses to follow the tradition because it seems to him that they are
equally unfree. Both Quincy’s and Raissa’s movements can be causally explained in
detail; neither movement was occurred due to an individual’s intention or purpose. But
wait! At this point, the skeptic may be reluctant to commit himself. He may clarify his
view by pointing out that he has no problem of seeing Quincy’s act of waving as brought
about by Quincy’s intention to please Raissa. But as soon as the skeptic refuses to give
up all the concepts surrounding the distinction he tries to remove, he will find himself in
need of a new word to replace the old expression “acting freely” or “of his own free will”

127
so the same distinction can be drawn. If so, his skeptical claim that no action is ever free
will turn out to be empty because he has smuggled the old concept of freedom in.
In the same spirit, Malcolm argues that when the skeptic claims that “all words
are really vague”, he gives up the concept that enables him to “call attention to those
differences of degree” between the words which we ordinarily consider as clear and
vague and when the “philosophizing biologist” is tempted to announce that “all matter is
really animate”, he gives up the tool that makes it possible for him to draw distinction
between animals and rocks. Moreover, the skeptic will soon find out that there is no way
to actually give them up because the distinctions he tries to remove are firmly embedded
in our conceptual network and daily practice.

(2) Disputes over Meaning
The Paradigm Case Argument helps us see that the skeptic is skeptical about a
linguistic fact rather than an empirical fact. As Keith S. Donnellan rightly points out,
“the skeptical position maintains its plausibility only through an unnoticed fluctuation
between the usual sense of the key expressions and some special sense.”
120
I shall
borrow Arthur C. Danto’s arguments in “The Paradigm Case Argument and the Free-Will
Problem” to explain why I think Donnellan’s thesis is both true and insightful.
Danto’s article was intended to be a criticism of PCA. He argued that PCA
cannot solve the free-will problem because the ordinary man and the skeptic are
interested in different problems. However, it seems to me that instead of seeing the

120
Keith S. Donnellan, “The Paradigm Case Argument”, in The Encyclopedia of Philosophy, edited by Paul
Edwards, 1967, p.40.

128
article as an objection to PCA, we may use it as a refutation of skepticism on behalf of
PCA.
Danto argues that regular people and metaphysicians are talking about different
things. The regular people are concerned with if Smith married of his own free will
while the metaphysician wonders is if anyone ever does anything of his own free will.
What we say in ordinary language would be irrelevant to the “celebrated problem of the
freedom of the will.” If it is true, by “acting freely” or “married of his own free will”, the
skeptic does not mean what the ordinary man means, then the skeptical thesis “X does not
exist” would lose its charm all together. As a result, the free will problem will be
dissolved.
According to Danto, whenever we employ the expression “acting freely” or “of
his own free will” in ordinary conversations, we always have in mind some specific
meaning of it. For instance, when Flew says that the young man was married of his own
free will, what he had in mind must be something like nobody forced him to get married
or he didn’t propose to his girlfriend under the influence of drugs or hypnosis. Danto
writes,
Thus, any sentence of the form (1) X did x of his own free-will –
which occurs in a normal, i.e., non-philosophical context – may be
translated into an exactly synonymous sentence of the form (2) X did x
in the absence of condition C. C is the specific condition which the
user of a (1)-like sentence means to exclude, and will vary, of course,
from context to context. But whenever a (1)-like sentence is true, it
will serve to contradict a sentence of the form (3) Nobody does x in
the absence of C.
121

121
Arthur C. Danto, ““The Paradigm Case Argument and the Free-Will Problem”, pp.121-122.

129
This is the case when the ordinary man talks but not the case with the skeptic (i.e.
hard determinist). Danto remarks, “the determinist can allow all ordinary uses of the
expression ‘of his own free-will’ and still retain his metaphysical equanimity.” The
skeptic who denies the pillar-boxes to be red would have no problem correctly
identifying red objects. Isn’t the skeptic contradicting himself? Danto believes that he is
not because “‘No one ever does anything of his own free-will’ simply fails, in the normal
contexts to which I refer, to mesh with, or oppose, ‘Smith married of his own free-
will’”.
122

The causal determination implied by the truth of determinism that has prevented
the skeptic from applying the expression of “married of his own free will” to Flew’s case
is usually considered different from the determination or manipulation resulted from
drugs, hypnosis, mental illness, gun threat, financial pressure, or even something like “It
was love at first sight, the moment we saw them together we knew it was meant to be, it
was inevitable they should be married”
123
, therefore
…the quarrel between the free-willist and his metaphysical enemy in
no way resembles an argument between two plain men who might
debate over whether or not Smith married of his own free-will. … The
disputants could then go on to argue metaphysics, or they could return
to the history of Smith’s matrimony. But not both together.
124

122
This is probably a misunderstanding of the determinist’s view because the incompatibilist never thinks
that he is talking about a different concept of free will than the ordinary one. For example, see van
Inwagen’s recent paper “How to Think about the Problem of Free Will?”.

123
Danto, Ibid., p.124.

124
Ibid.

130
Despite the misleading appearance, the skeptic and the plain man are not talking
about the same thing because by “married of his own free will” the metaphysician means
something different from what the plain man has in mind.
At this point, the skeptic might rebut that he does not intend his skeptical claims
to reflect ordinary usages of the relevant expressions. He has no problem to let ordinary
people keep their ordinary concepts and speak in an ordinary way; however, he must
maintain that according to a more rigorous standard, some of those concepts are not
applicable. For example, our concept of circle in mathematics is in accordance with a
much more rigorous standard than the ordinary usage. Why can’t we have an expression
of “acting freely” or “married of one’s own free will” in philosophy in the same fashion?
In response, Hanfling argues that if the expressions in dispute have a special
philosophical sense instead of an ordinary one, the skeptical claims would not be of any
interests. For example, if the cover story of Time Magazine is “No Cancer Will Be
Found on Earth Next Year”, I am sure everyone would regard it as breaking news. But
after reading the report, the readers realize that by “cancer” the author means “a cancer
that does not respond to any medical treatment”. I doubt if anyone would still be
interested in the content of the article. Hanfling rightly points out,
That the issue is not so simple is, however, evident from the fact that
the philosophers making the claims under criticism have, after all,
chosen to express them in ordinary language. When a philosopher
makes the statement “No bodies are solid”, it certainly looks as if what
he means is that no bodies are solid. If this is not so, why does he not
choose a frankly technical terminology? Or why, if his meaning is that
given by Passmore, does he not simply say that “no bodies are wholly
impenetrable?” This statement is certainly true as far as objects such
as table are concerned, as can be proved with the aid of an ordinary
drill; but it would not have the fascination or the quality of paradox of

131
the original statement. … It is just because the denials appear to be,
and are commonly taken to be, about solidity, certainty etc.,
themselves, that the claims command attention and provoke
discussion.
125

The skeptic who denies the existence of solid objects does not simply claim that
“No physical objects are wholly impenetrable” or “No physical objects are without
interstices” because the skeptic appears to, and is often taken to, say more than that.
Otherwise, it would not sound as provocative as it does because it may well be true that
nothing satisfies the above descriptions, but this is no news at all.
If it is correct to say that the nature of the skeptical challenge is not about the
instantiations of the concept in question but rather about its meaning, PCA as Elucidation
can be constructed as follows:
(1) “X” is meaningful.
(2) The meaning of “X” can be explained by describing paradigm cases to
which “X” is applied.
(3) To refuse to apply “X” to paradigm cases is to dispute over its meaning
rather than its instantiation.
(4) Therefore, “X” in the skeptical thesis “X does not exist” does not have an
ordinary meaning.
It is important to note that PCA as Ostensive Teaching is different from PCA as
Elucidation in the following aspects:

125
Hanfling, p.34.

132
(1) In order to refute skepticism, the first version aims at showing that “X” is
instantiated in the world; the second version aims at clarifying the meaning
of “X”.
(2) In the first version, the meaning of “X” must be learned by watching
paradigm situations or objects; in the second version, the knowledge of
employing “X” correctly can be innate.
(3) The first version assumes that the teaching process is nothing more than
pointing to an example and uttering “This is X”; in the second version, the
teaching process is an elucidation of “X”.

(3) Ordinary Concepts and Scientific Discoveries
MacIntyre raises an objection to PCA that since the advocates of PCA have taken
“a ‘high priori road’” to refute skepticism “by defining ‘free’ in terms of paradigm
cases”, they should have trouble explaining how it is possible that “we should revise our
opinion of a prima facie free act in the light of psychological or sociological evidence.”
126

If science informs us that determinism is true, then in light of this new piece of
knowledge we may realize that the paradigm cases of free will are not genuine cases of
free will. If so, we should stop using the concept of free will all together.
In response, Hanfling argues that science has nothing to do with our ordinary
concepts:
If certain paradigms are solid according to the normal conditions of
solidity (one cannot pass one’s hand through the object, etc.) then no

126
Alasdair MacIntyre, “Determinism” in Mind, vol. 66, 1957, p.32.

133
scientific advance could conceivably show that these paradigms are
not solid; and if the meaning of “free action” is to be found in
paradigms which include such features as rationality, then no scientific
advance could conceivably show that these are not really cases of free
action.
127

According to Hanfling, if by “solid” we mean what we ordinarily mean, namely,
my hand cannot pass through a solid object, then no scientific progress can show that the
paradigm cases of solidity is not really solid. Similarly, if it is a paradigm case of free
action when I raise my hand to wave to a friend, no scientific discoveries can change this
fact.
In everyday life, we often use concepts such as “up”, “down”, “above” and
“beneath”. The skeptic proposes that we should stop using them because scientific
advance shows that the earth is round rather than flat and the round shape of the earth
makes nothing really “above” or “beneath” something else. Should I stop using “above”
in my statement that “Toby’s mailbox is above mine”? Scientific knowledge may change
an application of a concept in empirical statements. For instance, as we become more
scientifically informed, we apply “round” instead of “flat” to the shape of the earth, but
the meaning of “round” or “flat” remains the same and the distinction between the two is
not changed at all. Similarly, the distinction between “above” and “beneath” or between
“free action” and “forced action” should not be changed by scientific advance either.

(4) Concepts and Explanations

127
Hanfling, Ibid., p.37.

134
The primary goal of PCA is to show that the skeptical thesis “X does not exist” is
nothing surprising because if the skeptic refuses to apply “X” to paradigm cases, “X”
does not have an ordinary meaning. The opponents of PCA wish to show that they have
a good reason not to apply “X” to paradigm cases because these cases may be misleading.
A. J. Ayer offers an example to illustrate this point.
128
He argues that when asked to
consider a paradigm case of “demonic possession”, the advocate of PCA may point to
someone who is “foaming at the mouth or uttering strange sounds”, but the existence of
the paradigm case of “demonic possession” does not imply that “demons are really at
work”
129
. Similarly, the fact that someone is exhibiting pain behavior does not imply that
person is really in pain. In the case of free will, the “symptom” of enjoying free will may
be “clearly marked” but the occurrence of the symptoms does not imply that free will
really exists. Since the paradigm cases of relevant expressions are misleading in the first
two cases, it can be misleading in the third as well.
Friends of PCA would certainly agree that it is possible that someone may pretend
to be in pain by displaying pain behaviors. But does such a case count as a paradigm
case of “being in pain”? The answer is no. When we elucidate a standard example of
someone in pain, a mere description of pain behaviors is not enough because the situation
where the person is found in pain should also be described. If the circumstance under
which the pain behavior occurs is properly explained, we will clearly see whether it is a
paradigm case of “pretending to be in pain” or “being in pain”. Actually it is not hard to

128
A. J. Ayer, The Concept of a Person, p.18.

129
Hanfling, p.28.

135
think of paradigm cases of someone in pain. For example, a spy caught by his enemy is
being tortured; a screaming child with bad teeth is holding his cheek; or a pregnant
woman is in labor. The absurdity of the skeptical position is that, according to it, there is
no such thing as a genuine paradigm case of “being in pain”.
The example of “demonic possession” can be responded in a similar way. In
order to give a paradigm case of “demonic possession”, it is not enough to know that
someone is foaming at mouth and uttering strange sounds. A “further and essential
point” needs to add is that “these things are due to the presence of a demon”
130
.
Similarly, that the case of the young bridegroom is considered as a paradigm case
of “married of his own free will” is not because of the “symptoms” of “acting freely” but
because of the elucidation of the situation as a whole. Simply to know that the young
man acted like a free man is not enough. We also need to know if he is under any
abnormal influence such as drug, hypnosis or a political threat etc. But the skeptic claims
that no one can ever act freely or do something of one’s own free will if a complete
causal explanation of the world can be given. He may insist that an accurate description
of the situation that the young man finds himself in should include “he could have chosen
otherwise.” Actually, PCA would have no problem to include the expression “could have
done otherwise” in his young bridegroom case. For example, it could be said that the
young man married the girl of his own free will because he could have chosen otherwise
if he had never met the girl or fell in love with her.

130
Hanfling, p.29.

136
Passmore and Lucas believe that something else should be added to the
description of the paradigm case of “married of his own free will”. Passmore thinks that
an act is free only when the act “proceeds from an act of will and when that act of will
has the metaphysical peculiarity of being uncaused”
131
. Similarly, Lucas holds that if
someone does something of his own free will then that action has to be an action to which
“no complete causal account can be given” of it.
132
But are these accurate elucidations or
explanations of our existing concept of free will? They are not. Hanfling rightly points
out that in Passmore or Lucas’ account of free will, extra conditions have been added
artificially to free will, which turns it into a different concept. However, Passmore or
Lucas may rebut that these conditions ought to be added to the expression if they are not
already part of it. But a suggestion as such would make it clear that they are
recommending a new concept rather than describing the concept that we have. So when
they make the claim that “free will does not exist” they are not talking about our concept
of free will but rather a concept of free will that is encoded with a special sense.
Ayer believes that the free will case is “similar to that of witches, where the
essential further condition was at one time thought to be satisfied in many cases, but,
again due to the advance of science, this is no longer so.”
133
Obviously, he overlooks an
important difference between the case of demonic possession and the case of acting
freely. The former is concerned with a causal explanation while the latter is not.

131
John Passmore, Philosophical Reasoning, 1961, p.118.

132
J. R. Lucas, The Freedom of the Will, 1970, p.12.

133
Hanfling, Ibid., p.29.

137
In the case of free will or solidity, we do not try to explain anything. Rather, we
are applying concepts. We use these concepts to draw distinctions between one’s raising
one’s hand to wave to a friend and one’s hand is grabbed by someone else and forced to
wave, or between a wall which we may lean against and a wall which we may not. These
distinctions have nothing to do with the hidden mechanisms that make these phenomena
possible. Suppose that a child is lying in bed in a hospital foaming at the mouth and
uttering strange sounds. Her father remarks, “My child is possessed by a demon. This is
a typical case of demonic possession because all the symptoms are caused by the work of
a demon.” Upon hearing this, the doctor rebuts, “No, the cause of the symptoms is not a
demon but a brain injury.” What is in dispute here is the causal explanation of the same
symptoms, which is not the case with “solidity” or “acting freely”. So Ayer’s example of
“demonic possession” cannot show that PCA is incorrect, but only that a particular causal
explanation is incorrect.

(5) Form of Life
Advocates of PCA endeavor to show that the skeptical thesis “X does not exist” is
the result of a conceptual confusion. I shall further support this thesis by analyzing the
waving example I mentioned earlier.
Let us assume that scientists have positively proved that our world is
deterministic. The ordinary man and the skeptic are disputing over whether free will
exists. The ordinary man continues to employ words and expressions such as “acting
freely” or “of his own free will” whenever he finds them applicable. However, the

138
skeptic stops using them all together because he believes that free will is an illusion in a
deterministic world.
One day, the ordinary man and the skeptic got together to have a cup of coffee.
One of the topics they chatted about is Quincy and Raissa.
O: I don’t think Raissa likes Quincy anymore because she didn’t even want to say
hello to him. But of course, Quincy still likes Raissa.
S: How do you know?
O: Quincy waved to her with a big smile on his face. Obviously, he did it of his
own free will. But Raissa didn’t. She was forced by her friend to wave back.
S: You may think so, but you’re wrong. Quincy’s act wasn’t free either.
O: Really? Did he put on a phony smile? Tell me what’s going on.
S: Don’t get me wrong. I didn’t mean his smile wasn’t genuine. What I meant
was that he had to do what he did at that moment and he could not have done otherwise
because his act was determined long before he was born. So neither Quincy nor Raissa’s
waving was a free act.
O: But do you think that Quincy wanted to please her by waving and smiling to
her while Raissa didn’t want to wave back if she had a choice?
S: Yes, of course. We may say so, but it is inaccurate. A more scientific way to
put it is that neither Quincy nor Raissa had a choice. Quincy could not have had a
different desire but to please her because that desire was determined long time ago.
O: Do you see the different way in which Quincy didn’t have a choice and Raissa
didn’t have a choice?

139
S: Yes, I do. Quincy only pretended to run into her. He actually followed her all
day long. But Raissa had absolutely no idea about Quincy’s plan.
O: When you say Quincy had a plan, you seemed to think that one can still make
a plan about future in a deterministic world.
S: No, I don’t think so. I was just being lazy when I said so. Actually, strictly
speaking, Quincy felt certain desire and had certain motive given the situation he found
himself in at that point in his life, so his mind worked out some ideas and thoughts as a
result of a long history of the network of causal chains. It certainly wasn’t a plan in the
sense that it was completely originated by him. But you may roughly call it a plan in the
sense that he spent three days and nights to figure out and write down all the details of it.
S: But this is exactly what I mean by “plan”. Why can’t we just talk in this
simple way? I don’t see any difference between how we think about or react to things.
The only difference between you and me is that you prefer a more complicated way of
talking ever since you heard that determinism is true.
It seems to me that the ordinary man is right that the skeptic’s claim “Free will
does not exist” is an idle wheel. Nothing comes from it and nothing follows it. In other
words, nothing has been changed about his way of thinking, judging, acting or reacting.
This is so because
Even if determinism came to be universally accepted, it would leave
this part of ordinary language quite unmodified. For people would
still have inclinations, would still sometimes be forced to act against
those inclinations, would very likely still seek extenuation, etc. Or
they would sometimes be released from certain pressures, restrictions,
and obligations. So we should still require the expressions we now
employ, e.g., “he did it of his own free-will” or “He is free” (i.e., “no

140
longer in conference,” “no longer engaged to the girl from Vassar,”
“has broken the habit,” “is out of jail”).
134

As long as we still live in the same way, interested in the same distinctions, such
as that between acting under pressure and acting without pressure, we would “have no
reason to drop perfectly innocuous, perfectly clear ways of making needed distinctions.
We will always have to have some way of distinguishing shot-gun marriages from other
kinds.”
135
This does not mean that our form of life never changes. It certainly does, but
the change is never a result of a whim, decision, or metaphysical argument. It can be a
result of a scientific and technological advance, but it does not happen in a way that the
skeptic has in mind. For example, if one day no woman will give birth or breastfeed
anymore due to the easy access to the service of in vitro fertilization and infant care, the
concept of “mother”, “give birth” or “breastfeeding” may vastly change.

(6) A Moorean Argument
Now I shall consider an argument that may be seen as similar to PCA. The
argument was originally constructed by G. E. Moore in his attempt to refute the skeptical
thesis that “There are no material objects”. Moore held out his hands and rebutted “Here
is one hand and here is another”. Since hands are material objects, the proposition that
there are two hands entails that there are at least two material objects. This is Moore’s
famous refutation of skepticism and proof of an external world.

134
Danto, Ibid., p.123.

135
Ibid.

141
Philosophers such as Lehrer and Barry Stroud find Moore’s argument arbitrary
and question-begging because it does not seem to take skeptical arguments seriously. In
response to these objections, William G. Lycan argues forcefully to defend Moore and a
Moorean argument against skepticism. He holds that
no deductive ‘proof’ can be anything more than an invitation to
compare plausibility: Of the propositions P
1
, …P
n
, and ~C’, which is
the least plausible or credible? The proof affords no deeper
investigation.
136

The main thesis of Lycan’s argument is that when two sets of premises support
contradicting conclusions in deductive arguments, we should compare the two sets of
premises and try to determine which set is more plausible. For example, we may adapt
Lycan’s construal of Moore’s argument to the Manipulation Argument as follows:
(1) A manipulated act is beyond one’s control and a free act is not beyond one’s
control.
(2) If determinism is true, then every act including paradigmatically free acts
are no different in any relevant respect from a manipulated act.
(3) Therefore, if determinism is true, then every act including paradigmatically
free acts are beyond one’s control.
This argument is competing with the following argument:
(4) A manipulated act is beyond one’s control and a free act is not beyond one’s
control.
(5) My act of waving to my friend is a free act regardless whether determinism
is true.

136
William G. Lycan, “Free Will and the Burden of Proof” in Minds and Persons, p.115.

142
(6) Therefore, if determinism is true, there is at least one act that is not beyond
one’s control.
The conclusion (3) of the first argument contradicts the conclusion (6) of the
second argument, so they cannot both be true. But since both arguments are valid, we
need to compare the plausibility of the two sets of premises. As (1) and (4) are the same,
the only difference is between (2) and (5). Which is more plausible? Lycan believes that
at any given time, some of my knowledge claims will be more
plausible, and rationally more credible, than are the purely
philosophical premise(s) of any skeptical argument.
137

Upon comparison, (5) is found to be more plausible than (2). Therefore, the
second argument is more plausible than the first. This way the skeptical argument is
refuted.
Even though the premises “Here is one hand and here is another” and “My act of
waving to my friend is a free act” are obviously true according to our common sense and
hence can be seen as paradigm cases of material objects and free act respectively, the
Moorean Argument is different from the Paradigm Case Argument that I try to defend.
The advocate of Moorean Argument simply presents two ways of using the concept in
question – a skeptical way and a paradigmatic way (or, commonsensical way) and points
out that the paradigmatic way is more plausible than the skeptical way without explaining
by what standard the comparison has been made. Lycan thinks that “no particular
criterion of plausibility, credibility or comparative certainty is or need be invoked”.138

137
Lycan, “Moore Against the New Skeptics”, p.44.

138
Ibid, p.40.

143
But this seems to me incorrect and it is the main weakness of the Moorean Argument.
The argument sounds question-begging and unconvincing is exactly because it does not
offer such a criterion.
PCA does not rely on a comparison as the Moorean Argument does because no
comparison can be possibly made. When we compare which claim is more plausible or
credible, we assume that they are stated within the same conceptual framework, that is,
they are using the same concept in the same way, but this is not the case with the
disagreement between the Moorean premise and the skeptical premise. The skeptic is
able to see hands but refuses to call them “material objects”. He is not doing anything
implausible, but only challenging the ordinary way of using the term “material objects.”
Similarly, the incompatibilist is able to recognize that my act of waving to a friend is a
free act and hence under my control in an ordinary sense, but he is challenging this sense.
It seems to me that this challenge should not be characterized as implausible. At least,
just label it “implausible” will not help untie the knots of skeptical worries.
PCA in its second version aims at elucidating or explaining the conceptual
connections between the concept at issue and its surrounding concepts that are internally
related to it. This is an elucidation rather than a definition with respect to the free will
problem because an analysis of the concept of free will cannot be unified into necessary
and sufficient conditions for the application of it. The conceptual connections of free will
can be clarified by an elucidation, namely, drawing a conceptual map of various uses of
free will. Once the meaning of free will is clarified, we will see that the nature of the
skeptical thesis is challenging the meaning of the concept rather than its instantiation.

144
Does the skeptic try to propose a new concept? The skeptic is not even trying to do this
because both his challenge and proposal involve logical problems as it was shown in
chapter 4.

3. Conclusion
In this chapter, I described the debate between the advocates of PCA and the
opponents of PCA. I distinguished PCA as Ostensive Teaching from PCA as Elucidation
and tried to argue that the second version of PCA is correct and can be used as a powerful
weapon against skepticism.
I also tried to show that the dispute between skepticism and PCA is not over an
empirical problem but a conceptual problem. What skepticism denies is not the
instantiation of a meaningful expression but the applicability of it, that is, the very
meaningfulness of the expression. So, when correctly understood, PCA is not an
argument that deduces the instantiation of a concept from its meaningfulness but rather
an argument that clarifies the meaning of the concept in question. The dispute over the
instantiation of a word is a disagreement about an empirical fact while the dispute over
the meaning of a word is a disagreement about a linguistic convention.

The PCA compatibilist tries to offer a correct conceptual analysis of free will by
taking a different strategy (the paradigm case strategy) from the most popular strategy
(the analysis strategy) that has been taken by both classical compatibilists (i.e. Hume) and
contemporary compatibilists (i.e. Frankfurt). The analysis compatibilist wishes to

145
discover conditions that are both necessary and sufficient for applying the concept of free
will while the PCA compatibilist believes that there is no such thing for free will because
not all correct uses of free will have something in common. The PCA compatibilist
should also be distinguished from those who understand paradigm cases of free will as
instances rather than conceptual elucidations of free will (i.e. Watson). According to the
PCA compatibilist, the correct account of free will is a conceptual analysis of free will
consisted of the explanations of paradigmatic examples to which free will and its
surrounding concepts are correctly applied.
The PCA compatibilist takes the incompatibilist challenge seriously. He believes
that simply offering a correct account of free will is not enough to refute incompatibilism
even though it is the crucial first step. What must be done further is to uncover the
hidden problems with the thesis of incompatibilism. As soon as we look carefully at the
logical structure of the concepts (i.e. global manipulation) which the incompatibilist
theory relies on, we will see that incompatibilism cannot stand.

146

Bibliography

Alexander, H. G., 1957, “More About the Paradigm-Case Argument”, Analysis, 18: 117-
120.

Anscombe, G. E. M., 1957, Intention, Oxford: Blackwell.
Armstrong, David, 1999, “A Naturalist Program: Epistemology and Ontology”, The
Eleventh Annual Romanell Lecture. Reprinted in Proceedings and Addresses of the
APA, 73(2): 77-89.

Austin, J. L., 1953, “Other Minds”, in Flew 1953. Reprinted in Austin 1961.
-------, 1961, Philosophical Papers, Oxford: Oxford University Press.
Ayer, A. J., 1963, The Concept of a Person, London: MacMillan Press.
-------, “Freedom and Necessity”. Reprinted in Watson 1982.
Baker, G. P. and Hacker, P. M. S., 1980, Wittgenstein, Understanding and Meaning,
Oxford: Blackwell.

-------, 1985, Rules, Grammar and Necessity, Oxford: Blackwell.
-------, 1986, Language, Sense & Nonsense, Oxford: Blackwell.
Berofsky, Bernard, 1966, Free Will and Determinism, New York and London: Harper &
Row, Publishers.

Bishop, John, 1989, Natural Agency: An Essay on the Causal theory of Action,
Cambridge: Cambridge University Press.

Bok, Hilary, 1998, Freedom and Responsibility, Princeton: Princeton University Press.

Bouwsma, O. K., 1949, “Descartes’ Evil Genius”, The Philosophical Review, 58: 141-
151.

Bradley, Raymond, and Swartz, Norman, Possible Worlds: An Introduction to Logic and
Its Philosophy, Oxford: Blackwell.

147
Bratman, Michael E., 1987, Intention, Plans and Practical Reason, Cambridge, Mass:
Harvard University Press.

-------, 1999, Faces of Intention, Cambridge: Cambridge University Press.
Campbell, Joseph Keim, O’Rourke, Michael and Shier, David, 2004, (ed.), Freedom and
Determinism, Cambridge: MIT Press.

Campbell, Joseph Keim, O’Rourke, Michael and Silverstein, Harry, 2007, (ed.),
Causation and Explanation, Cambridge: MIT Press.

Casullo, Albert, 2003, A Priori Justification, New York: Oxford University Press.

Chappell, V. C., 1961, “Malcolm on Moore”, Mind, 70: 417-425.
Chisholm, Roderick, 1973, The Problem of the Criterion, Milwaukee, WI: Marquette
University Press.

-------, “Human Freedom and the Self”. Reprinted in Watson 1982.
Danto, Arthur C., 1959, “The Paradigm Case Argument and the Free-Will Problem”,
Ethics, 69: 120-124.

Davidson, Donald, 1963, “Actions, Reasons, and Causes”, Journal of Philosophy, 60.
Reprinted in Davidson 1980.

-------, 1980, Essays on Actions and Events, Oxford: Clarendon Press.
Dennett, Daniel, 1984, Elbow Room: The Varieties of Free Will Worth Wanting
Cambridge, MA: MIT Press.

Dilman, Ilham, 1999, Free Will: An Historical and Philosophical Introduction, New
York: Routledge.

-------, 2002, Wittgenstein’s Copernican Revolution: The Question of Linguistic Idealism,
New York: Palgrave.

Donellan, Keith, 1967, “The Paradigm Case Argument”, in Paul Edwards (ed.) The
Encyclopedia of Philosophy, New York: MacMillan, 39-44.

Ellenbogen, Sara, 2003, Wittgenstein’s Account of Truth, New York: State University of
New York Press.

148
Eveling, H. S. and Leith, G. O. M., 1957, “When to Use the Paradigm-Case Argument”,
Analysis, 18: 25-33.

Findlay, J. N., 1951, “Time: A Treatment of Some Puzzles”, in Flew 1951.
Fischer, John Martin, 1994, The Metaphysics of Free Will, Oxford: Blackwell.
-------, 2004, “Responsibility and Manipulation”, Journal of Ethics, 8: 145-177.
-------, 2005, (ed.), Free Will: Critical Concepts in Philosophy, New York: Routledge.

Fischer, John Martin and Ravizza, Mark, 1998, Responsibility and Control: A Theory of
Moral Responsibility, Cambridge: Cambridge University Press.

Fischer, John Martin, Kane, Robert, Pereboom, Derk, and Vargas, Manuel, 2007, Four
Views on Free Will, Malden MA: Blackwell Publishing.

Flew, Antony G. N., 1951, (ed.), Essays on Logic and Language, 1st series, Oxford.

-------, 1953, Essays on Logic and Language, 2
nd
series, Oxford.
-------, 1955, “Divine Omnipotence and Human Freedom”, in New Essays in
Philosophical Theology, New York: The MacMillan Company.

-------, 1956, (ed.), Essays in Conceptual Analysis, London: MacMillan.
-------, 1956, “Philosophy and Language”, in Flew 1956.
-------, 1957, “‘Farewell to the Paradigm-Case Argument’: A Comment”, Analysis, 18:
34-40.

Flew, Antony and Vesey, Godfrey, 1987, Agency and Necessity, Oxford: Basil
Blackwell.

Forster, Michael N., 2004, Wittgenstein on the Arbitrariness of Grammar, Princeton:
Princeton University Press.

Frankfurt, Harry, 1969, “Alternate Possibilities and Moral Responsibility”, Journal of
Philosophy, 66:5-20. Reprinted in Frankfurt 1988.

-------, 1971, “Freedom of the Will and the Concept of a Person”, Journal of Philosophy,
68: 5-20. Reprinted in Watson 1982 and Frankfurt 1988.

149
-------, 1988, The Importance of What we Care About, Cambridge: Cambridge University
Press.

Gallois, Andre, 1977, “Van Inwagen on Free Will and Determinism”, in Philosophical
Studies, 32: 99-105.

Gellner, E. A., 1951, “Use and Meaning”, The Cambridge Journal, September 1951.

Ginet, Carl, 1966, “Might We Have No Choice?” in Lehrer 1966: 87-104.
-------, 1990, On Action, Cambridge: Cambridge University Press.
Grant, C. K., 1955, “Polar Concepts and Metaphysical Arguments”, Proceedings of
Aristotelian Society, 1955-6.

Hacker, P. M. S., 1987, Insight and Illusion, Oxford: Oxford University Press.
-------, 1990, Wittgenstein: Meaning and Mind, Oxford: Blackwell.
-------, 1996a, Wittgenstein: Mind and Will, Oxford: Blackwell.
-------, 1996b, Wittgenstein’s Place in Twentieth-Century Analytic Philosophy, Oxford:
Blackwell Publishers.

-------, 2007, “Thought and Action”, in Leist 2007.
Hanfling, Oswald, 1980, “Does Language Need Rules?”, The Philosophical Quarterly,
July: 193-205.

-------, 1987, “How is Scepticism Possible?”, in Philosophy, 62: 435-453.

-------, 1989, Wittgenstein’s Later Philosophy, Albany: State University of New York
Press.

-------, 1991, “What is Wrong with the Paradigm Case Argument?” in Proceedings of the
Aristotelian Society, 91: 21-38.

-------, 2000, Philosophy and Ordinary Language, London and New York: Routledge.

-------, 2002, Wittgenstein and the Human Form of Life, London: Routledge.
Haji, Ishtiyaque, 1998, Moral Appraisability, New York: Oxford University Press.

150
-------, 2006, “Hard- and Soft-Line Responses to Pereboom’s Four-Case Manipulation
argument”, Acta Analytica, 21:19-35.

Hardie, W. F. R., 1957, “My Own Free Will”, in Philosophy, 32, 120: 21.
Harre, Rom., 1957, “Tautologies and the Paradigm-Case Argument”, Analysis, 18: 95-96.

Hirsch, Eli, 2002, “Against Revisionary Ontology”, Philosophical Topics, Vol.30, No.1.

Hobbes, Thomas, 1651/1996, Leviathan. (ed.) Richard Tuck, New York: Cambridge
University Press.

Hook, Sidney, 1961, (ed.), Determinism and Freedom, New York: Collier Books.

Hume, David, 1739/1978, A Treatise of Human Nature, New York: Oxford University
Press.

James, William, The Principles of Psychology, in the series of The Britannica Great
Books Of The Western World, 53: 767-835.

Kane, Robert, 1996, The Significance of Free Will, New York: Oxford University Press.

-------, 2002, Free Will, Malden: Blackwell.
-------, 2005, A Contemporary Introduction to Free Will, New York: Oxford University
Press.

Kant, Immanuel, 1781/2007, The Critique of Pure Reason, New York: Penguin.
Kelly, Thomas, 2005, “Moorean Facts and Belief Revision, or Can the Skeptic Win?”, in
Philosophical Perspectives, 19: 179-209.

Lehrer, Keith, 1966, “An Empirical Disproof of Determinism?”, in Freedom and
Determinism.

-------, 1976, Free Will and Determinism, Humanities Press.

-------, 1960, “Can We Know that We Have Free Will by Introspection?”, in The Journal
of Philosophy, Vol. 57, No.5.

Leist, Anton, 2007, (ed.), Action in Context, Berlin: Walter de Gruyter GmbH & Co. KG.

Lewis, David, “Are We Free To Break The Laws?”, in Watson 2003.

151
Loux, Michael J., 1998, Metaphysics: A Contemporary Introduction, London: Routledge.

------- and Zimmerman, Dean, 2005, The Oxford handbook of Metaphysics, Oxford:
Oxford University Press.

Lucas, J. R., 1970, The Freedom of the Will, Oxford: Oxford University Press.
Lycan, William, 2003, “Free Will and the Burden of Proof”, in Antony O’Hear, (ed.),
Minds and Persons, Cambridge: Cambridge University Press, 107-122.

-------, 2001, “Moore against the New Skeptics”, Philosophical Studies, 103(1): 35-53.

MacIntyre, A. C., 1957, “Determinism”, Mind, 66: 28-41.
Mackie, Joan, 1985, Logic and Knowledge, Oxford: Clarendon Press.
Malcolm, Norman, 1942, “Moore and Ordinary Language’, in Richard Rorty, (ed.), The
Linguistic Turn: Recent Essays in Philosophical Method, Chicago and London: The
University of Chicago Press.

-------, 1963, Knowledge and Certainty: Essays and Lectures, London: Cornell University
Press.

-------, 1971, Problems of the Mind: Descartes to Wittgenstein, New York: Harper &
Row.

-------, 1977, Thought and Knowledge, Ithaca and London: Cornell University Press.

-------, 1986, Nothing is Hidden, Oxford: Blackwell.
McKenna, Michael, 2001, “Source Incompatibilism, Ultimacy, and the Transfer of Non-
Responsibility”, American Philosophical Quarterly, 38: 37-52.

-------, 2008, “A Hard-line Reply to Pereboom's Four-Case Manipulation Argument”,
Philosophy and Phenomenological Research, 77: 142-159.

McKenna, Michael and Russell, Paul, 2008, (ed.) Free Will and Reactive Attitudes:
Perspectives on P. F. Strawson’s “Freedom and Resentment”, Burlington: Ashgate
Publishing Limited.

Mele, Alfred, 2006, Free Will and Luck, Oxford: Oxford University Press.

Moore, G. E., 1959, Philosophical Papers, London: George Allen & Unwin Ltd., New
York: Humanities Press, Inc.

152

Moser, Paul K., 1987, A Priori Knowledge, New York: Oxford University Press.

Narveson, Jan, 1977, “Compatibilism Defended”, Philosophical Studies, 32: 83-88.

Nielsen, Kai and Phillips, D. Z., 2005, Wittgensteinian Fideism?, London: SCM Press.

Passmore, John, 1961, Philosophical Reasoning, New York: Basic Books, Inc.,
Publishers.

Pereboom, Derk, 1997, Free Will, Indianapolis: Hackett Publishing Company.
-------, 2001, Living Without Free Will, Cambridge: Cambridge University Press.

Phillips, D. Z., 1982, Through a Darkening Glass, Notre Dame: University of Notre
Dame Press.

Pink, Thomas, 2004, Free Will: A Very Short Introduction, Oxford: Oxford University
Press.

Pollock, John and Cruz, Joseph, 1999, Contemporary Theories of Knowledge, Oxford:
Rowman and Littlefield.

Rhees, Rush, 2003, Wittgenstein’s On Certainty: There, Like Our Life, Oxford:
Blackwell.

-------, 2006, Wittgenstein and the Possibility of Discourse, Oxford: Blackwell.
Richman, Robert, 1961, “On the Argument of the Paradigm Case”, Australasian Journal
of Philosophy, 39: 75-81.

-------, “Still More on the Argument of the Paradigm Case”, Australasian Journal of
Philosophy, 40: 204-207.

Rorty, Richard, 1967, The Linguistic Turn, Chicago and London: The University of
Chicago Press.

Ryle, Gilbert, 1949, The Concept of Mind, Chicago: The University of Chicago Press.

Scanlon, T. M., “The Significance of Choice” in Watson 2003.
Sider, Theodore, Hawthorne, John and Zimmerman, Dean W., 2008, (ed.), Contemporary
Debates in Metaphysics, Malden: Blackwell.

153
Smart, J. J. C., 1961, “Free Will, Praise and Blame”, in Mind, Vol. 70, No.279.

Soames, Scott, 2003, “Malcolm’s Paradigm Case Argument”, in Philosophical Analysis
in the Twentieth Century, Vol. 2, Princeton, NJ: Princeton University Press.

Strawson, P. F., 1962, “Freedom and Resentment”, Proceedings of the British Academy,
68: 187-211. Reprinted in Watson 1982.

Stroud, Barry, 1984, The Significance of Philosophical Scepticism, Oxford: Clarendon
Press.

Taylor, Richard, 1976, “Responsibility for Self”, Reprinted in Watson 1982.
Tomberlin, James, 2000, (ed.), Philosophical Perspectives, 14 (Action and Freedom),
Malden: Blackwell.

Van Inwagen, Peter, 1972, “Lehrer on Determinism, Free Will, and Evidence”, Vol. 23.

------- 1975, “The Incompatibility of Free Will and Determinism”, Philosophical Studies,
25: 185-99, Reprinted in Kane 2002.

-------, 1977, “Reply to Narveson”, in Philosophical Studies, Vol. 32.
-------, 1983, An Essay on Free Will, Oxford: Clarendon Press.
-------, 2008, “How to Think about the Problem of Free Will?”, in The Journal of Ethics,
12: 327-341.

Van Inwagen, Peter and Zimmerman, Dean W., 1998, (ed.), Metaphysics: The Big
Questions, Malden: Blackwell.

Vihvelin, Kadri, 2000a, “Libertarian Compatibilism”, in Tomberlin 2000, 14:139-166.

-------, 2000b, “Freedom, Foreknowledge, and the Principle of Alternate Possibilities”,
Canadian Journal of Philosophy, 30: 1-24.

-------, 2008a, “Compatibilism, Incompatibilism, and Impossibilism”, in Contemporary
Debates in Metaphysics, 303-318.

-------, 2008b, “Arguments for Incompatibilism”, on Website Stanford Encyclopedia of
Philosophy.

Wallace, Jay, 1998, Responsibility and the Moral Sentiments, Harvard: Harvard
University Press.

154

Watkins, J. W. N., 1957, “Farewell to the Paradigm-Case Argument”, Analysis, 18: 25-
33.

-------, 1957, “A Reply to Professor Flew’s Comment”, Analysis, 18: 41-42.
Watson, Gary, 1975, “Free Agency”, in The Journal of Philosophy, 72: 205-220,
Reprinted in Watson, ed. 1982.

-------, 1982, (ed.), Free Will, New York: Oxford University Press.
-------, 1987, “Responsibility and the Limits of Evil: Variations on a Strawsonian
Theme”, Reprinted in Fischer 2005.

-------, 2003, (ed.), Free Will (2
nd
edition), New York: Oxford University Press.
-------, 2004, Agency and Answerability, New York: Oxford University Press.
William, C. J. F., 1961, “More on the Paradigm Case Argument”, Australasian Journal of
Philosophy, 39: 276-278.

Wiggins, David, 1973, “Towards a Reasonable Libertarianism”, in Ted Honderich (ed.),
Essays on Freedom of Action, London: Routledge, 31-61.

Wittgenstein, Ludwig, 1953/2001, Philosophical Investigations, Malden: Blackwell.

-------, 1972, On Certainty, edited by G. E. M. Anscombe and G. H. von Wright, New
York: Harper & Row.

Wolf, Susan, 1993, Freedom within Reason, Oxford: Oxford University Press.
Wright, G. H. von, 1971, Explanation and Understanding, New York: Cornell University
Press.

Abstract (if available)

Abstract The objective of this study is to solve the free will/determinism problem by defending a new form of compatibilism. Unlike most classical and contemporary compatibilists, the author takes the Paradigm Case Argument (PCA) strategy instead of the analysis strategy to defend compatibilism. She endeavors to offer a correct conceptual analysis of free will as well as an adequate response to the incompatibilist challenge.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

How to use intuitions in philosophy

PDF

The effect of massed versus spaced arguments on persuasion

PDF

Semantic web technologies for event analysis in reservoir engineering

PDF

An argument for the criminal hoax

PDF

On top to random shuffles, no feedback card guessing, and fixed points of permutations

PDF

The amygdala and conditioned taste aversion

PDF

The minimal approval view of attributional-responsibility

PDF

Just imagining things: Hume's conception-based account of cognition

PDF

Assessing therapeutic change mechanisms in motivational interviewing using the articulated thoughts in simulated situations paradigm

PDF

In-flight turbulence: an articulated thoughts in simulated situations (ATSS) paradigm investigation of air travelers ' experiences

PDF

Reducing gang involvement through employment: a pilot intervention

PDF

Positivist realism

PDF

Molyneux's question answered!

PDF

Isoniazid population pharmacokinetics in HIV perinatally exposed infants

PDF

Using organized objectives to structure arguments for collaborative negotiation of group decisions in software design

PDF

A deontological explanation of accessibilism

PDF

Beyond paradise and power. Contending arguments on the future of transatlantic relations and the West (1991-2001)

PDF

Genomic mapping: a statistical and algorithmic analysis of the optical mapping system

PDF

Promoting student achievement: a case study of change actions employed by an urban school superintendent

PDF

Troubling the boundaries: "blacknesses," performance, and the African American freedom struggle of the 1960s

Asset Metadata

Creator Zhu, Xiaoyu (author)

Core Title Defending compatibilism: a paradigm case argument

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Publication Date 11/02/2009

Defense Date 10/01/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag compatibilism,determinism,free will,incompatibilism,OAI-PMH Harvest,paradigm case argument

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Vihvelin, Kadri (committee chair), Albertson, David (committee member), Levin, Janet (committee member)

Creator Email sherri.zhu@gmail.com,sherri_zhu@hotmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2704

Unique identifier UC1330287

Identifier etd-Zhu-3326 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-271951 (legacy record id),usctheses-m2704 (legacy record id)

Legacy Identifier etd-Zhu-3326.pdf

Dmrecord 271951

Document Type Dissertation

Rights Zhu, Xiaoyu

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu