-- the Determinism and Freedom Philosophy Website --

When you are making up your mind, deciding what  to do, you have the idea that you are free in what you are doing. It is hard  to shake. You are going to do the one thing, but you can certainly do the  other. That is what you think. Rational deliberators, as they can be called,  have an inescapable sense of freedom. Dana Nelkin, in the following clear-headed  paper, asks if this sense of freedom establishes that determinism is not true. Read on for her answer. She also has things to say about another understanding  of the claim that we know we are free when we are making up our minds. Whether or not you agree, you will learn things. Prof. Nelkin is at the University of California at San Diego.


I. Introduction1

We appear to have an inescapable sense that we are free, a sense that we cannot abandon even in the face of powerful arguments that this sense is illusory.  It has often been suggested, perhaps most notably by Kant, that we have such a sense of freedom in virtue of being rational deliberators.  More precisely,

(R) Rational deliberators, in virtue of their nature as rational deliberators, necessarily have a sense that they are free.

This claim has a great deal of intuitive plausibility: it is when we engage in rational deliberation about what to do that we are most likely to become aware that we have a sense of ourselves as free.  Even skeptics about free will, who are convinced that they are not free, often admit to a sense that they are free when they deliberate about what to do.2

(R) constitutes a rare point of agreement among philosophers who are deeply divided on the question of whether we really are free.  But while there are skeptics who accept (R) and lament the fact that we are thereby condemned to live under an illusion, there are others who have much higher hopes for (R): they take (R) to be the first major step in arguments that we are in fact free.  For example, after claiming that we have a belief in our own freedom "so necessary in most of our rational operations", Thomas Reid writes that,

This natural conviction of our acting freely, which is acknowledged by many who hold the doctrine of necessity, ought to throw the whole burden of proof upon that side...3

and Kant writes,

Now I say that every being which cannot act in any way other than under the idea of freedom is for this very reason free from a practical point of view.4

It is important to determine whether (R) is true, then, for at least two reasons.  First, if (R) is true and we are not free, then one of our most central conceptions of ourselves is both false and inescapable.  On the other hand, if Reid and Kant are correct, and (R) can be shown to be true, then a major premise can be established in an argument that we are in fact free.5      

Is (R) true?  In this paper, I will explore two lines of reasoning supporting the conclusion that it is.  I believe that one of these lines is promising, and will suggest some different way in which it might be developed.  I begin by offering a more detailed account of the sense of freedom attributed to all rational deliberators.  As might be expected, philosophers have understood (R) in different ways.  In section II, I examine what I call the "Indeterministic" reading, according to which the sense of freedom in (R) is the belief that one can choose from a set of undetermined alternative courses of action.  Although I argue that, on the Indeterministic reading, (R) is false, I believe the ways in which it has been defended contain important insights.  In section III, I argue that we should adopt what I call the "Belief-Concept" reading of (R), according to which the sense of freedom in (R) is the belief that one's actions are up to one in such a way that one is accountable for them.

II. The Indeterministic Reading of (R)

According to the Indeterministic reading of (R), all rational deliberators necessarily believe that they can choose from multiple future courses of action, none of which is determined.  Perhaps the earliest adherent of the Indeterministic reading is Aristotle.  In the Eudemian Ethics, Aristotle claims that about other [things] not only existence and non-existence is possible, but also human deliberation; these are things the doing and not doing of which is in our power.6

As stated in this passage and elsewhere in Aristotle's work, it initially appears that Aristotle is making the strong claim that indeterminism is required for deliberation rather than the weaker claim that a belief in indeterminism is required for deliberation.  However, in several places, Aristotle qualifies his claims about the proper objects of deliberation in a way that suggests that deliberation requires the belief that the object of deliberation possesses certain properties rather than that the object of deliberation actually possesses those properties.7  What is essential to deliberation are the subject's epistemic attitudes about the object of deliberation, and, in particular, the belief that any object of one's deliberation is possible but not determined. 

While the inspiration for the Indeterministic reading can thus be traced to Aristotle, the view has many contemporary adherents, and some have thought Kant to be another advocate.8  Despite its long history and current popularity, I claim that the Indeterministic reading is false.

Let me begin by setting out one line of reasoning that might be used to support the Indeterministic reading of (R).9

(i)If one deliberates about an action, A, then one must believe that it is in one's power to do and to forego A.


(ii)If one believes that it is in one's power to do/forego A, then one must believe that there exist no conditions sufficient to render inevitable either A or not-A.


(iii)If one deliberates about A, then one must believe that there exist no conditions sufficient to render inevitable either A or not-A.10

Premise (i) has seemed to many to be an attractive thesis, but I will present several thought-experiments designed to undermine it.  Then I will go on to challenge what are perhaps even more basic assumptions on which the indeterministic view must ultimately rest: assumptions about the very nature and purpose of rational deliberation.

First, consider the following case: imagine that you know that a brilliant scientist has the ability to fiddle with your brain in a way which causes you to act as she wishes you to.  You know that she wants you to vote for Gore over Bush in the upcoming presidential race, and that if you do not decide to vote as she wishes, she will cause you to vote that way.  So, for instance, you know that if you were to prepare to vote for Bush or otherwise fail to decide to vote for Gore, the brilliant scientist would cause you to vote for Gore.  It seems to me that you could still evaluate the reasons for voting for each candidate and decide to vote for Gore on the basis of those reasons.  In this case, contrary to (iii), you know that conditions exist which are sufficient for your voting for Gore, while you nevertheless deliberate and decide to vote for him.  Further, contrary to (i), you do not believe you could forego the action upon which you decide. 11

Now one might object that your voting for Gore as a result of deliberation and decision differs in kind from the action of voting for Gore as a result of the brilliant scientist's interference with your brain.  One might think that the actions are of different types, especially if one is convinced that types of actions are individuated in part by the types of causes they have.12  If this is so, the objection goes, then you do deliberate about something which you believe you could do or forego: a particular type of voting for Gore.

Although this objection has some force, I think that it can be met.  For if you accept the initial description of the thought experiment, what you evaluate with a view to acting are the reasons for voting for Gore.  The objection, by contrast, has you evaluating reasons for voting for Gore on the basis of deliberation and decision, where this action is something that you could do or forego.  Even if this sort of deliberation is possible, it remains a coherent part of the thought experiment that you deliberate and decide to vote for Gore.  Thus, additional objects of deliberation which might be regarded as actions to be done or foregone within the thought experiment do not alter the fact that there is something about which one deliberates and decides, despite not believing that one could either do or forego it.

The objection might be pressed in the following way, however: consider a modified version of the original voting case (the "inevitability" case) in which you believe it to be inevitable that, no matter what, the scientist will fiddle with your brain in such a way that you vote for Gore.  To ensure that the case is one in which it is inevitable that your action will have the same cause in counterfactual circumstances, let it be built into the case that there is no possibility of over-determination: you believe that, no matter what, the changes in your brain induced by the scientist will be the sole cause of your voting for Gore.  In this case, the objection proceeds, surely you cannot deliberate, and the explanation is that you fail to believe that the action in question-voting for Gore in a particular way-is one that you could either do or forego.  If the lack of such a belief is the correct explanation of the failure of deliberation in this case, then such a belief must be a necessary condition for deliberation.  Further, this conclusion shows that there must be a mistake in the description of the original voting case.  In particular, the objection continues, the conclusion of the inevitability case reveals that the supposed lesson of the original case rests on a mistaken assumption about action individuation, a mistake that is successfully avoided in the inevitability case.

The conclusion drawn about the inevitability case is questionable, however.  It is clearly a case in which you believe that you will perform a particular action in a particular way and so lack the belief that you can either do or forego an action.  It may also be that you cannot deliberate in this case.  But it is not the former point that explains the latter.  Rather, the explanation for your inability to deliberate in such a case is that you do not believe that the deliberation and the resulting decision on your part could be causally efficacious.  Possession of the belief that one's deliberation and decision could be causally efficacious is a necessary condition for deliberation, and the lack of such a belief in this case is the explanation for any inability to deliberate.13

These two beliefs-that one's deliberation could be causally efficacious and that one's future actions are undetermined-might easily be conflated.  The idea that one's deliberation and decision caused one to perform an action might be confused with the idea that one's deliberation and decision resulted in one's making an undetermined possibility actual.  But if we are careful to distinguish between these two ideas, then we can resist the present challenge to the original voting case. 

At this juncture, the proponent of the Indeterministic reading of (R) might point out that even if the original voting example is effective in raising questions about (i) and so about (iii), it remains the case that it is not an example in which the agent is said to believe that determinism is true.  While it is perhaps true that, in one sense, the example is a case in which the agent believes that conditions exist which are sufficient for his or her future actions, it is not true if by "conditions" we mean "a previous state of the world together with natural laws."14

Yet there are cases in which individuals can rationally deliberate despite lacking a belief in indeterminism.  For example, consider the fact that there are determinists who deliberate.  More particularly, suppose that a person believes that scientists have succeeded in creating a computer that predicts her actions on the basis of past states of the world and natural laws.  This case is clearly one in which someone believes that there are conditions sufficient for her performing the actual action she will perform and in which she believes that her action is physically determined. 15 The question is, can such a person deliberate if she knows that such a computer is busy churning out its predictions?  The answer seems to me to be "yes".  I can imagine that such a computer exists right now somewhere in Siberia, printing out predictions about my future actions.  Even if I believed this to be true, and so did not believe it possible for me to perform each of a number of alternative actions, I would go right on deliberating about all sorts of things.

To the case of the deliberating determinists, the following reply might be offered. Such people do exist, and while they believe that determinism is true, they also hold the contradictory belief that determinism is false. 16   Thus, although there are those who believe that determinism is true, as long as they continue to deliberate their behavior manifests the belief that they can do or forego the actions about which they deliberate, and thus manifests the belief that determinism is false.  In fact, it appears that the proponent of the Indeterministic reading could account for every purported counter-example to (i) or (iii) simply by attributing contradictory beliefs to every deliberator who claims to have a belief contrary to that attributed in (i) or (iii).

However, the idea that deliberation manifests either of these beliefs can be seen to rest on still more fundamental assumptions which can themselves be questioned, assumptions about the nature and point of rational deliberation.  For we can ask: what are the conditions under which we take a piece of behavior to manifest a belief?  The answer is that, in general, we do so when we think that the behavior is rationalized by a belief, that is, when we think that the presence of the belief in question is part of the correct explanation of the behavior that makes it intelligible in the context of the agent's beliefs, desires and actions.17  For example, consider the case of a store owner who hires a private detective to watch and report on his cashier's behavior.  In the absence of a belief that it is likely that the cashier is embezzling money, the store owner's action does not make sense in light of his other actions and attitudes.  Thus, this belief rationalizes his hiring of a detective.

Given this general understanding of what it would be for an activity to manifest a belief, let's consider why it might be thought that the beliefs attributed to deliberators by the indeterministic view are manifested in the activity of deliberation.  I will focus directly on the belief attributed in (iii) for the sake of simplicity, although the reasoning can easily be transposed to account for the belief attributed in (i), as well.  Here, then, is the question: in what way can the belief attributed in (iii) be said to rationalize deliberation? 

Either of two answers might be given, one having to do with the essential nature of deliberation, the other with its point, or purpose.  Richard Taylor provides material for the first sort of answer.  He writes that if an agent, Adam, believes that conditions already existed sufficient for his performing the action he will perform in the future, then he can no longer deliberate about the matter because... he believes it is not up to him what he does; the matter has already been "decided," one way or the other, and there is no decision for Adam to make.18

On one natural reading of this passage, the purpose of deliberation is to decide upon an action, where "decision" is constituted by an agent's closing all but one of several previously "open" possibilities.  It follows that if conditions exist sufficient for a particular action's being performed, and hence that there are no previously open possibilities to be closed or to be chosen to remain open by the agent, then there can be no true decisions on the part of anyone.  It could then be argued that if an individual did not believe that his deliberation could succeed in bringing about a decision, then his deliberating would not make sense.  Therefore, the belief that he could make a decision would rationalize his deliberation, and so that belief would be manifested in the activity of deliberation.  Finally, given the present understanding of "decision," it follows that one can only deliberate if one believes that there are open possibilities for one to close. 

The problem with this suggestion is that the understanding of "decision" on which it is based ought to be rejected.  For on the view that a decision is the closing of previously open possibilities, it is not strictly speaking true to say that someone has reached a decision if, for some reason, it cannot be carried out.  Yet the naturalness of saying that someone made a decision which could not be carried out suggests that we do not normally think of a decision as being defined by its effects.  Rather, it is more natural to say that a decision is something like the formation of an intention which, depending on the circumstances, may or may not be carried out.  According to this latter understanding of "decision" we can successfully decide to do something even if the decision does not succeed in effecting anything, let alone in the closing of all but one of several previously open possibilities.  

The second way in which the belief attributed in (iii) might be said to rationalize deliberation involves a somewhat more modest view of the connection between determinism on the one hand, and deliberation and decision on the other.  The idea is that if deliberation and decision are to have a point at all, then they must succeed in closing all but one of several previously open possibilities.  This view is suggested by Aristotle, who, in arguing that it would be absurd to think that everything that happens, happens of necessity, claims that in that case "there would be no need to deliberate or to take trouble."19

On this view, one could be said to make a decision even if determinism were true, although in that case the decision would lack a point.  It could then be argued that if an agent failed to believe that there was a point to making a decision, then it would not make sense for her to deliberate.  Therefore, deliberation manifests the belief that there is a point to making a decision, and thus, that there are open possibilities to be closed by the agent as a result of deliberation and decision.

But is it in fact true that rational deliberators can only take rational deliberation to have a point if they take it that their decisions result in the closing of all but one of several previously open possibilities?   Of course, I take it that the original voting case provides a negative answer here.  For I find it perfectly coherent that the agent in that case does not believe that her purpose in deliberating is to change the course of events from what it would otherwise be, or even to find out what she will do, and yet the possibility that she deliberates is perfectly intelligible.  Yet merely appealing to this sort of case will not convince those who would attribute contradictory beliefs to the voter.  I must say what it is that the case reveals about our attitudes toward the point of deliberation.  And that is that reasons can motivate us to act in a certain way even if we know that we will act in that way no matter what.  The case shows us that the possibility of acting on the basis of good reasons can itself be seen to have value, and so, one can take the activity of rational deliberation to have a point if one takes it that the activity can result in one's acting for good reasons.  And the possibility that rational deliberators could take deliberation to have a point in the absence of indeterminism undermines this kind of defense of the Indeterministic reading.  Thus, we have not been given sufficient reason to conclude that the belief attributed in (iii) is manifested by rational deliberation.

Each of the two attempts to provide an argument for the conclusion that rational deliberation manifests the belief attributed in (iii) relies on a problematic premise: the first rests on a dubious assumption about the nature of a decision, the second on a questionable assumption about the point of deliberation.  Of course, there may be reasons other than the two that I have considered here for accepting the conclusion.  However, the burden rests with those who advocate the Indeterministic reading; for the fact that the purported counter-examples to it can only be resisted at the cost of attributing contradictory beliefs to otherwise rational people provides some prima facie reason to reject the view.20

Having made it clear which aspects of the reasoning to the Indeterministic reading I wish to reject, let me emphasize an aspect of the reasoning that I have not rejected.  I believe that the general strategy of showing that the sense of freedom rationalizes the very activity of rational deliberation is a promising one, and I will pursue it in section III.  In particular, I will return to the idea that for rational deliberation itself to be an intelligible activity, deliberators must take such activity to have a point or purpose.  For even if deliberation can be seen to have a point in the absence of indeterminism, it does not follow that it has a point in the absence of freedom.  And if we refrain from prejudging the question of whether freedom requires indeterminism, it is possible that, in an important sense of "free," we cannot take our deliberation to have a point if we do not have a sense that we are free. 

III. The Belief-Concept Reading

A. The Sense of "Freedom"

To explain the sense of "freedom" in the type of reading of (R) I favor, it will be useful to borrow some terminology from John Rawls, and distinguish between the concept of freedom and particular conceptions of freedom. 21  First, there is a concept of freedom that all (or at least many) people grasp, even if they disagree as to the necessary and sufficient conditions for its instantiation.  There must be such a concept if compatibilists, incompatibilists, and those who believe freedom impossible are rightly said to disagree.  Of course, it is possible to argue that the whole problem with the free will debate is that the participants are indeed speaking different languages, and that agreement could be reached immediately if this fact were only acknowledged.  I think that this view gives too little credit to the participants in the debate, however, and that there is at least one important concept about which there is disagreement. 22

Second, there are also different analyses of freedom offered (such as, the ability to choose from among various undetermined actions or the ability to act in the absence of constraint).  These are conceptions of freedom, attempts to spell out in detail the conditions under which actions are free.

It is easy to describe various conceptions of freedom, then.  But what is the concept of freedom?  In my view, the concept of free agency is the notion of one's actions being up to one in such a way that one is, in a basic sense, responsible or accountable for them.

This account must be clarified in various ways.  First, let me emphasize that in identifying the concept of freedom in this way, I mean to suggest that the mutual implication between 'free' and 'being up to one in such a way that one is accountable' should be recognized by anyone in possession of the concept of freedom.  Second, "responsibility" and "accountability," like "freedom," each have many different senses.  Perhaps "accountability" has fewer meanings, and so there is reason to favor this term over "responsibility."  Third, it is also very important that "responsible" and "accountable" not be read as "morally responsible" and "morally accountable," despite the fact that the former locutions are often used as shorthand for the latter.  I think that we have a more basic notion of responsibility (or accountability) which underlies various kinds of responsibility: responsibility for one's productivity and creativity, moral responsibility, and so on.  Responsibility in this basic sense is a necessary, but not sufficient, condition for moral responsibility.  The reason is that moral responsibility requires certain sorts of knowledge, both general and particular.  But this does not mean that the concept of freedom I have characterized is a weak notion.  It captures what is traditionally known as the "freedom" condition for moral responsibility, while leaving open the question of how to characterize the "knowledge" condition.  And it provides the condition for the normativity required by moral responsibility, while leaving open the possibility that one who is free is not morally obligated.

I believe that this concept remains neutral as among various particular conceptions of freedom.  To see this, consider that an agent-causationist might argue that if one's actions are to be up to one in this way, they must be agent-caused.23  Alternatively, one might argue that if one's actions are to be up to one in this way, they must be determined by one's valuational judgments.24  The first of these views requires the truth of indeterminism, while the second does not. 

Now some might argue that the true concept of freedom is not in fact neutral, and that the concept of freedom just is the notion that 'one can choose from between alternative courses of action' or 'that one could have done otherwise' or that 'one is the undetermined source of one's actions.'  Yet, there have been a large number of attempts to provide such conditions that do not allude to the choosing between alternatives, having the ability to do otherwise, or being undetermined.  I do not mean to suggest that any such attempts have met with complete success, only that they have been genuine attempts.  If the concept of freedom were the notion that one can do otherwise, then it would seem that agreement as to the entailment between freedom and the ability to do otherwise would be readily forthcoming.  The fact that many who have attempted to provide necessary and sufficient conditions for freedom have not granted the point suggests that the concept of freedom is not simply the notion that one can do otherwise.  Further, even if we were to accept the point of the objection for the sake of argument, we can still accept that the notion I have described is an important condition for moral and other sorts of responsibility.25  Thus, if it can be shown that we must represent ourselves as free in the sense that I have described, that would be a significant result.  Further, if Kant and Reid are correct, and (R) is the first step in a sound argument that we are free (in the sense specified in (R)), then that, too, would be a significant result. 

Finally, the concept I have described satisfies widely accepted constraints on the notion of freedom which is at the heart of the free will debate.  For example, philosophers often claim that the notion of freedom in which they are interested is a notion that applies to persons, but not to many non-human animals who are thought not to be persons.26  That is, there is a sense in which the actions of humans can be up to them, or be their own, in a way that the actions of a spider cannot.  Since the notion of accountability might be thought to be limited in its applicability in such a way as to exclude many non-human animals, the concept that I have described matches this widely accepted feature of freedom.  The readings I will consider in what follows take "free" in (R) to pick out the concept of freedom.  Hence, they are "conception-neutral."27

B. The Sense of "Sense"

There are a number of different ways one might understand "sense" in the "sense of freedom," including "feeling", "appearance", "belief," and "rational commitment."  I begin in section IIIC by arguing that the sense of freedom in (R) is an epistemic commitment, or belief, that one is free.  It does not follow from this reading that all rational deliberators can articulate their commitment without considerable probing and reflection; but it does follow that they actually represent themselves as free. In this respect, it is like the Indeterministic reading of (R).  In what follows, I call the conception-neutral reading that understand "sense" in this way, the "Belief-Concept" reading.  In section IIID, I consider a weaker reading of (R) according to which the sense of freedom is a rational commitment that one is free.  Roughly, one is rationally committed to a proposition when reflection and recognition of features of one's own mental states and reasoning is sufficient for one's believing that proposition.  But one need not actually believe that one is free in order to be rationally committed to the proposition.  I call the reading of (R) that incorporates both the conception-neutral reading of "free" and the rational commitment reading of "sense", the "Commitment-Concept" reading.  Although I believe that there is good reason to accept the Belief-Concept reading, it is true that even weaker premises are needed to support the Commitment-Concept reading, and thus, that the latter requires less in the way of defense.  At the same time, as I will explain, both readings can do powerful explanatory and justificatory work.         

C. An Argument for the Belief-Concept Reading

We can begin by focusing once again on the nature of rational deliberation.  Recall that one rationally deliberates when one considers and evaluates reasons with a view to deciding to act.  The culmination of rational deliberation is the adoption, on the basis of one's evaluation, of certain reasons as one's reasons for performing an action, and the consequent decision so to act.  In other words, the deliberator decides to act on the basis of those reasons.

Given the nature of rational deliberation just described, deliberators are the sorts of beings who have a guiding conception of their purpose when they perform intentional, goal-directed actions.  Further, it is important to note that they have a sense of their purpose whether or not they explicitly deliberate about what to do.  For example, if a rational deliberator is engaged in seeking the latest international news, then she has a conception of the goal that defines her activity.  If she does not conceive herself to be seeking the latest news, then whatever activity in which she is engaged, she is not seeking the news.  This point applies to rational deliberation itself.  Here, too, one need not be able to articulate one's purpose in deliberating without considerable reflection, but one must have a conception of one's purpose when one deliberates, namely, that of finding, adopting and acting on good reasons.

Accepting this point depends crucially on an understanding of the nature of rational deliberation as an activity directed by the agent toward a particular goal.  The fact that a set of mental states constitutes a genuine piece of rational deliberation is determined not merely by the content of those states (e.g., "R is a reason to do action A"), and not merely by the fact that they tend to issue in action.  It is also essential that the deliberator be thereby attempting to accomplish something, namely, the adoption of good reasons as her own.  This requires that the deliberator have a guiding sense that her activity is aimed at this goal.28 

Now if rational deliberators have a conception of themselves as seeking to adopt and act on good reasons, then they must take themselves to be capable, in general, of finding such reasons.  This claim is independently plausible, but I believe it derives additional support from its being explained.  Implicit in the sense of purpose that guides their intentional, goal-directed actions is a commitment to a plan (however incomplete) to achieve it.  Yet if one does not view one's purpose as attainable, then one cannot envision a plan for its implementation.  In other words, in order to have a guiding conception about how one can achieve one's purpose, one must believe that one can.29  It is not necessary for deliberators to believe that they must succeed on each occasion of deliberation; but they must believe that they have the ability to succeed in at least some typical situations.  In other words, they must believe that they are the sort of being who, by engaging in the activity of rational deliberation, can succeed in finding and adopting good reasons for acting.  Otherwise, it would not be possible to conceive of their activity as constituting a way of achieving their purpose.  Thus, rational deliberators must take themselves to be the sorts of beings who can and do sometimes find, adopt, and act on good reasons.

If this is the case, then rational deliberators must believe that there are good reasons to be found, adopted, and acted upon.  In other words, they must believe that there are reasons they ought to act upon.  But if one is the sort of being to whom such "ought-statements" apply, then one's actions can be rationally justified or unjustified.  And that one's actions are potential objects of this sort of justification is the way in which one can be accountable for one's actions. 

At this point, the question arises whether rational deliberators must believe that they are accountable simply because they believe that there are reasons upon which they ought to act.  Of course, it is not true in general that an agent must believe the consequent of a true entailment when she believes the antecedent.  But this is a special case.  There is a tight conceptual connection between the idea of there being reasons upon which one ought to act and the idea of one's being accountable; it seems that in order to see oneself as the sort of being to whom such reasons apply, one must see oneself as the sort of being who is accountable for his or her actions. 30

Finally, then, we can see how one must believe that one's actions are up to one in such a way that one is accountable for them: one believes that one's actions can be performed as a result of one's own adopted reasons and, further, that they are potential objects of rational justification.31   Thus, by identifying essential features of rational deliberation, we can see why deliberators must have a sense of themselves as free (in the conception-neutral sense).

D. Objections, Replies, and Elaboration

At this point, a number of questions might be raised about this line of reasoning.  One natural question is whether rational deliberators can take their activity to have a different purpose from the one I identify.  In particular, can rational deliberators who come to believe that they are not free continue to deliberate while taking themselves to have a different purpose in doing so?  For example, they might take the purpose of their deliberation to be to achieve the maximal satisfaction of their desires, or to increase the beauty in the world by being a passive receptor of reasons.

Although this objection has some force, I believe that its force can be defused by focusing once again on the nature of deliberation.  It is compatible with the reasoning above that rational deliberators take their activity to have more than one point or purpose.  The claim that those engaged in rational deliberation take their activity to have as its purpose their acting for good reasons is perfectly consistent with the claim that they also take that activity to have an additional, or further, point.  For example, a rational deliberator might be asked to deliberate and decide on a course of action as part of her participation in a researcher's psychological experiment.  In deliberating, she might have as her purpose to please the researcher, but if she follows the researcher's instructions to the letter, she must also take as her purpose to act for good reasons.  In order to do so, she must take it that her deliberation can be efficacious and that there are reasons on the basis of which she should act.  Hence, as I argued earlier, she must have a sense that she is free.  The fact that one can imagine a rational deliberator taking her activity to have any of a number of different points does not show that rational deliberators can avoid taking their activity to have the point of acting on the basis of good reasons.32

Further, in addition to consulting our intuitions about the nature of rational deliberation in the abstract (as I have so far been urging us to do), it is useful to consider in more concrete terms the question of whether rational deliberators must take deliberation to have a particular purpose.  To this end, consider our own behavior when we believe that there are no good reasons for acting in a certain way.  For example, suppose you are at an ice cream shop, having decided to buy an ice cream cone.  You believe that all the flavors are equally good.  You don't generally deliberate in such circumstances, but simply pick a flavor at random.  On my view, the reason for your lack of deliberation in such a case is that you see no point to deliberating.  We see no point in engaging in the evaluation of reasons for acting when we think that there is no reason that we should act in a particular way.  We are not in a position in which we could act for good reasons in choosing a flavor of ice cream, hence we see no point to deliberating, and so do not engage in it. 

This thought experiment-together with the investigation into the nature of goal directed intentional activity I sketched earlier-strongly suggests that the very activity of rational deliberation manifests the sense that there is a point or purpose to the activity.  Further, these considerations suggest that rational deliberation manifests the sense that the point of one's activity is to decide and ultimately to act on the basis of good reasons.

At this point, a second objection might naturally arise: even if one has a conception of the purpose of one's activity, one need not believe that one's purpose can be achieved in order to be rational in engaging in it.  In what follows, I defend the idea that one must believe that one has the general ability to succeed in order to have a conception of one's purpose in performing goal-directed intentional actions.  There is a long history of controversy on this and related questions, and I cannot do full justice to it here.33  But I will suggest one approach to it in the context of defending a key premise in the reasoning for the belief-concept reading.  Let me begin by emphasizing that I am not relying on the claim that one must believe one can succeed on every occasion.  Rather the claim is that one must believe one can succeed in typical circumstances when one attempts the relevant type of action.  This point might be sufficient to defuse the present worry about the belief-concept reading.  For as I hope to show, those on both sides of the debate over the connection between intentional action and belief might be able to find agreement once this point is noted.       

In thinking about the relation between intentional action and belief, intuitions about examples play an important role.  Here is an example that provides support for the idea that one must believe one can achieve one's purpose, when one has a conception of it in acting intentionally: I cannot intentionally engage in any activity that might be described as either flying or trying to fly, no matter how much I desire to and how much I flap my arms.  The natural explanation seems to be that I do not believe it possible for me to fly.  In fact, I believe that I lack the general ability to fly.  I can imagine circumstances in which someone else tries to fly.  But these circumstances include delusions on the part of the person making the attempt.  If someone were to believe that his physical abilities were very different from what they are, or that the laws of aerodynamics or gravity were different from what they are, then he could try to fly.  (Of course, this would require a great number of changes in one's belief system and probably a great deal of irrationality.)   Here, it seems that what explains the difference between this person and me is the difference between our beliefs about what our flapping our arms could possibly accomplish.

Those who argue against a necessary connection between belief and intentional action offer examples on the other side.  Kirk Ludwig offers this one, among others34: A friend insists that I can make a basket from half-court.  I believe it impossible for me to make a basket from that distance, and set out to show him that even if I try as hard as I can, I still can't do it.  I try, and, amazingly, make the basket.  This appears to be a case in which I intentionally make a basket without believing that it is possible.  According to Ludwig, my intentionally making the basket in this case shows that I also intended and tried to make it before surprising myself with my success.   Since I have argued that rational deliberators engaged in intentional goal-directed activity have a sense of their purpose, the case seems to be one in which the sense of purpose is not accompanied by a belief that it can be achieved. 

Reasons might be offered to resist this example and others like it, but what is more important for our purposes here is the kind of example it is (and is not).  For even those who argue that intentional action does not require the belief that success is possible agree that there are some things one cannot intentionally do (or even try to do), like flying or making a basket from ten miles away.  Some explanation for this fact is required.  

Ludwig, for example, acknowledges that you could not try to hit a home run by holding the bat in a "bunt" position.35  Why not?  The reason, according to Ludwig, is that this stance and the limited motion it permits are not designed to bring about the end of hitting home runs.  In cases like this in which one intentionally performs an action without believing that one can succeed, one might recognize that circumstances are special in such a way that one's activity cannot succeed.  This is a cashing out of the idea that in order to act intentionally, one must conceive one's actions as of a type designed to bring about a certain end. 

If this is right, then we have a way to distinguish between the cases of making a basket from half-court and swinging "all-out" for a home-run on the one hand, and making a basket from Mars and a home-run from a "bunt" position on the other.  The view also suggests that we assume that when we act intentionally our actions are of a type that, under at least some typical circumstances, can succeed.  Further, there is a rationale for this: if one views a certain kind of activity as one some of whose typical instances are successful in achieving their purposes, then one can have a conception of how they are successful.  This makes it possible for us to envision our own activity of this type as a way of implementing a plan to achieve its purpose.36

Admittedly, it is difficult to give criteria for "type" of action here.  But I think it is possible to rely on an intuitive idea.  I can't try to fly, or make a home-run from the bunt position, because these are not the kinds of actions at which I could succeed under anything like normal circumstances.  I must believe that I have the general ability to succeed, if I have a sense of my purpose in acting.37

This conclusion is all that is needed in order to defend the belief-concept reading of (R) from the present objection.  For even if one need not believe that one can succeed on a particular occasion of deliberation, one must believe that one has the general ability to succeed.  One must take it that one's activity is successful under at least some typical circumstances.  Otherwise, one won't be able to view one's engagement in the activity as counting as a way of achieving the purpose of adopting good reasons for acting.  Thus, one must believe that one is free in the conception-neutral sense. 

As I have argued, this reasoning should persuade those on a wide spectrum of views positing particular connections between intentional action and belief.  However, others might worry that, according to this reasoning, rational deliberation requires too much in the way of conceptual development and self-reflection.  Young children and non-human animals would appear to be counter-examples to the reasoning for the Belief-Concept Reading of (R) since they include rational deliberators who lack mental state and related concepts (e.g., 'action', 'reasons,' and 'responsibility') and those who have not reflected on the purpose of their activity.  This sort of worry deserves to be taken seriously.

However, I believe that the appeal of the alleged counter-examples is undermined by reflection on the nature of the central concepts in the reasoning for the Belief-Concept reasoning.38 First, let me emphasize that rational deliberation as I understand it is itself a very sophisticated cognitive activity: the consideration and evaluation of reasons with a view to deciding to act, where one's decision is based on one's evaluation and adoption of reasons as one's reasons for performing an action.  Once this is understood, it becomes difficult to maintain that young children and non-human animals provide counter-examples to the claim that all rational deliberators must have the concept of reasons, for example.  For the cognitive sophistication required to engage in rational deliberation itself would seem to rule out at least some members of these groups, and, in particular, the very same members who are excluded from possession of sophisticated concepts such as reasons.

At the same time, it is important to emphasize once again that the concepts in question do not include the concepts of moral reasons and moral responsibility, but the concepts of reasons and responsibility in a basic sense.  Thus, the argument for the Belief-Concept reading of (R) does not presuppose the possession of any moral concepts by rational deliberators.  Recognition of the robust understanding of rational deliberation articulated above together with the relatively basic nature of the notions of reasons and responsibility should dispel the worry raised by the alleged counter-examples that there must be a faulty step somewhere in the reasoning for the Belief-Concept reading.  Finally, even if lingering doubts remain about the strength of the Belief-Concept reading of (R), it is possible to adopt a weaker conception-neutral reading without giving up much of the explanatory and justificatory role of (R).  According to the Commitment-Concept reading, rational deliberators are rationally committed, in virtue of being rational deliberators, to their being free.  They need not actually believe that they are free, but they are in a state such that mere reflection and recognition of features of their own mental states and activity would suffice for the beliefs in question.  To reach this weaker conclusion, each premise of the reasoning for the Belief-Concept reading of (R) might be weakened in such a way as to incorporate rational commitment in place of belief, and thus, to require less in the way of self-reflection than the reasoning for the Belief-Concept reading.  Equally importantly, the Commitment-Concept reading of (R) can do much of the work that the Belief-Concept reading can do.  For it can explain the centrality of the belief that one is free for those (like us) who have reflected on the matter.  And it can also play a powerful role in arguments like Reid's and Kant's that we are in fact free.  To take the simplest example, Reid's burden-of-proof argument set out earlier does not seem to lose much of its force when we substitute "rational commitment" for "conviction".  Thus, even if we adopt the Commitment-Concept reading of (R) over the Belief-Concept reading, (R) remains an important thesis.                     

Before concluding, it remains to consider one final objection.  Unlike the others, this one does not question the reasoning for the Belief-Concept reading of (R), but instead questions the meaning and significance of the conclusion.  The objection is that once we see that freedom in the conception-neutral sense is closely tied to the ability to act for good reasons, it is not clear that the notion of freedom in the conception-neutral readings of (R) is really the concept of freedom after all.  The quality of "being chosen" seems to fade into the background while the notion of rational capacity can be seen to undergird the conception-neutral notion of freedom described. 39

In response, it is important to note that although the concept of freedom can be seen to be closely associated with reasons and accountability, this fact does not detract from its conception-neutral character.  For example, nothing that I have said rules out the possibility that satisfaction of the concept of freedom as I have characterized it requires the falsity of determinism.  Further, the concept, 'freedom', is not simply to be equated with the concept, 'ability to respond to good reasons to act.'  The concept of freedom, on my view, is that one's actions are up to one in a particular and special way, namely, such that one is accountable for one's actions.  It is true that at least one way of one's actions being up to one such that one is accountable for them depends partly on the agent's ability to respond to good reasons for acting.  But I believe that this is a connection that we should find attractive, particularly once the connection between obligation by reasons and accountability is noted.  For if one's actions are up to one in such a way that one is (in a non-moral sense) accountable for them, then we would seem to have just what is wanted in a "freedom" condition for moral responsibility. 

Finally, even if we set aside the question of whether the concept of freedom is the concept of one's actions being up to one such that one is accountable for them, the latter remains a significant one.  For it is one that provides an important condition for moral and other sorts of responsibility, and one that we care deeply about. 

In sum, the Belief-Concept reading of (R) faces important challenges.   Yet an intuitively plausible line of reasoning gives it considerable resources with which to respond.  As a result, the Belief-Concept reading (together with the weaker Commitment-Concept reading) remains a promising interpretation of the widely accepted idea that in virtue of being rational deliberators we cannot escape the sense that we are free.

IV. Conclusion

Both the Indeterministic reading and the Belief-Concept reading of (R) rest on reasoning which takes the sense of freedom to be a belief manifested by the very activity of rational deliberation.  They diverge in the content of the beliefs each attributes to rational deliberators.  Equally important is the difference between the notions of rational deliberation employed by each reading, for it is on this difference that the difference in beliefs ultimately rests. Unlike the Indeterministic reading, the Belief-Concept reading takes the essence of rational deliberation to be the attempt to find and adopt good reasons for acting.  By building on the reason-seeking features of rational deliberation, it is possible to explain why rational deliberators must have a sense of their actions being up to them in such a way that they are accountable for them.  At the same time, this focus on the reason-seeking features of rational deliberation makes the Belief-Concept reading of (R) particularly well suited to an anti-skeptical argument in the spirit of Kant and Reid.  For if rational deliberation is essentially a reason-seeking faculty, it is tempting to conclude that the simple possession of such a rational faculty could not be responsible for a false belief.  Of course, turning this thought from a tempting idea into the conclusion of a sound argument that rational deliberators are free is a project of its own.  Yet if the Belief-Concept reading of (R) is correct, then we are at least entitled to a key premise in such an argument.


Adams, F. (1995). "Trying: You've Got to Believe," Journal of Philosophical Research 20, 563-570.
Albritton, R. (1985).  "Freedom of the Will and Freedom of Action," Presidential Address, Proceedings of the American Philosophical Association 59, 239-251.
Anscombe, G.E.M. (1957). Intention. Oxford: Blackwell.
Aristotle. (1984a). The Complete Works of Aristotle. J. Barnes, ed. Princeton: Princeton University Press.
Aristotle. (1984b). Eudemian Ethics. J. Solomon, tr. In Aristotle (1984a), 1922-1981.
Aristotle. (1984c). Nicomachean Ethics. W.D. Ross, tr., revised by J.O. Urmson. In
Aristotle (1984a), 1729-1867.
Aristotle. (1984d). De Interpretatione. J.L. Ackrill, tr. In Aristotle (1984a), 25-38.
Bok, H. (1998).  Freedom and Responsibility. Princeton: Princeton University Press.
Bratman, M. (1986).  Intention, Plans, and Practical Reason. Cambridge: Harvard University Press.
Burge, T. (1979).  "Individualism and the Mental," Midwest Studies in Philosophy 4, 73-121.
Burge, T. (1998).  "Reason and the First Person," in Knowing Our Own Minds, C. Wright, B. Smith, and C. MacDonald, eds. Oxford: Clarendon Press.
Castañeda, H. (1975). Thinking and Doing. Dordrecht: Reidel.
Chisholm, R. (1964). "Human Freedom and the Self," The Lindley Lecture. In Watson (1982), 24-35.
Clarke, R. (1992). "Deliberation and Beliefs About One's Abilities," Pacific Philosophical Quarterly 73, 101-113.
Davidson, D. (1963). "Actions, Reasons, and Causes," Journal of Philosophy 60. Reprinted in Davidson (1980), 3-20.
Davidson, D. (1969). "The Individuation of Events," in Essays in Honor of Carl G. Hempel, N. Rescher, ed. Dordrecht: Reidel. Reprinted in Davidson (1980), 163-180.
Davidson, D. (1980). Essays on Actions and Events. New York: Oxford University Press.
Davidson, D. (1985). "Reply to Pears" in Essays on Davidson Actions and Events B. Vermazen and M. Hintikka, eds. Oxford: Clarendon Press.
Dennett, D. (1984). Elbow Room. Cambridge: MIT Press.
Fine, G. (1981). "Aristotle on Determinism: a Review of Richard Sorabji's 'Necessity, Cause, and Blame'," The Philosophical Review 90, 561-580.
Frankfurt, H. (1969). "Alternate Possibilities and Moral Responsibility," Journal of Philosophy 66. Reprinted in Frankfurt (1988), 1-12.
Frankfurt, H. (1971). "Free Will and the Concept of a Person," Journal of Philosophy 68. Reprinted in Frankfurt (1988), 11-25.
Frankfurt, H. (1988). The Importance of What We Care About. Cambridge: Cambridge University Press.
Grice, P. (1971). "Intention and Uncertainty," Proceedings of the British Academy. London: Oxford University Press.
Harman, G. (1976). "Practical Reasoning," Review of Metaphysics 29, 431-463.
Hill, T. (1985). "Kant's Argument for the Rationality of Moral Conduct," Pacific Philosophical Quarterly 66, 3-23.
Hobbes, T. (1668/1994). Leviathan. E. Curley, ed. Indianapolis: Hackett.
Honderich, T. (1988). The Consequences of Determinism 2.  Oxford: Clarendon Press.
Jackson, F. (1998).  From Metaphysics to Ethics. Oxford: Oxford University Press.
Kant, I. (1781/1965). The Critique of Pure Reason. N. Kemp Smith, tr. New York: St. Martin's Press.
Kant, I. (1785/1981). Grounding for the Metaphysics of Morals. J. Ellington, tr.  Indianapolis: Hackett.
Kapitan, T. (1986). "Deliberation and the Presumption of Open Alternatives," The Philosophical Quarterly 36, 230-51.
Ludwig, K. (1992). "Impossible Doings," Philosophical Studies 65, 257-281.
Ludwig, K. (1995).  "Trying the Impossible: A Reply to Adams," Journal of Philosophical Research 20, 563-570.
Malle, B. and Knobe, J., (2001). "The Distinction Between Desire and Intention: A Folk-Conceptual Analysis," in Malle, B., Moses, L. and Baldwin, D. (2001).
Malle, B., Moses, L., and Baldwin, D., eds. (2001). Intentions and Intentionality Cambridge: MIT Press.
McCann, H. (1986). "Rationality and the Range of Intentions," Midwest Studies in Philosophy 10, 191-211.
Mele, A. (1995). Autonomous Agents: From Self Control to Autonomy. New York: Oxford University Press.
Mele, A. (2001). "Acting Intentionally: Probing Folk Notions," in Malle, B., Moses, L. and Baldwin, D. (2001).
Meltzoff, A., Gopnik, A., and Repacholi, B. (1999) "Toddlers' Understanding of Intentions, Desires and Emotions: Explorations of the Dark Ages," in Developing Theories of Intention, Zelazo, P., Astington, J., and Olson, D., eds. London: Lawrence Erlbaum Associates, Publishers. 
Nelkin, D. (in-preparation). Rational Deliberation and the Sense of Freedom.
Pereboom, D. (1995). "Determinism Al Dente," Nous 29, 21-45.
Pereboom (2001). Living Without Free Will.  Cambridge: Cambridge University Press.
Pettit, P. (1989). "Determinism with Deliberation," Analysis 49, 42-44.
Reid, T. (1788/1983). Essays on the Active Powers of Man, in Inquiry and Essays, R.E. Beanblossom and K. Lehrer, eds. Indianapolis: Hackett.
Sorabji, R. (1980). Necessity, Cause, and Blame. Ithaca: Cornell University Press.
Strawson, G. (1986). Freedom and Belief. Oxford: Oxford University Press.
Taylor, R. (1964). "Deliberation and Foreknowledge," American Philosophical Quarterly 1, 73-80.
Thalberg, I. (1962). "Intending the Impossible," Australasian Journal of Philosophy 40, 49-56
Thomas, G. (unpublished manuscript). "Deliberation and Determinism."
Van Inwagen, P. (1983). An Essay on Free Will. Oxford: Oxford University Press.
Velleman, J.D. (1989). "Epistemic Freedom," Pacific Philosophical Quarterly 70, 73-97.
Velleman, J.D. (1985) "Practical Reflection," The Philosophical Review 94, 33-61.
Watson, G. (1975). "Free Agency," Journal of Philosophy 72. Reprinted in Watson (1982), 96-110.
Watson, G., ed. (1982). Free Will. Oxford: Oxford University Press.
Wellman, H. and Inagaki, K. (1997). The Emergence of Core Domains of Thought: Children's Reasoning About Physical, Psychological, and Biological Phenomena. San Francisco: Jossey-Bass Publishers.
Wellman, H. and Phillips, A. (2001). "Developing Intentional Understandings," in Malle, B., Moses, L. and Baldwin, D. (2001).


1 This paper is a longer version of a paper presented at the Inland Northwest Philosophy Conference 2001.  I am very grateful to the organizers of the conference and the editors of this volume, Joseph Keim Campbell, Michael O'Rourke, and David Shier, for the opportunity to be part of such a stimulating and constructive exchange of ideas.  I would also like to thank my commentator, Ron Wilburn, and to the audience for their helpful comments and for an excellent discussion.  I presented an earlier version of the paper at the University of Virginia in February 2001, and I thank the audience for valuable discussion on that occasion.  A number of people read versions of this paper and offered comments that caused me to improve the paper.  For this, I am very grateful to Tal Brewer, Tyler Burge, Jim Cargile, Peter Hanowell, Barbara Herman, Tina Huggins, Andrew Hsu, Pat Matthews, Peter Murphy, Michael Otsuka, Derk Pereboom, Sam Rickless, George Thomas, and two anonymous referees for this volume.

2. For example, Galen Strawson writes that a free will skeptic who concentrates on abandoning his "ordinary conception of freedom" may temporarily experience "a total paralysis of all purposive thought as it is ordinarily conceived and experienced." (1986), p. 102.  It is when we are trying to make decisions (moral and otherwise) that our belief in our own freedom is most acutely felt.  (See p. viii, and chapter 3).  At the same time, Strawson argues that freedom is impossible and that our belief is false.  It must be noted that Strawson does not accept (R), but rather the weaker claim that humans who rationally deliberate necessarily believe themselves free. Castañeda accepts something like (R), and at the same time accepts the possibility that the skeptic is right: if so, "the universe is ugly; given the biological and psychological primacy of practical over contemplative thinking, we are, thus, condemned to presuppose a falsehood in order to do what we must think practically." (1975), p. 134.

3. Reid 1788/1983, 344.

4. Kant 1785/1981, p. 50 [GW 448].  Controversy abounds over just how Kant should be interpreted here.  I will not enter into that controversy here.

5. I explore different ways this argument can be developed in Nelkin (in-preparation).

6. Aristotle (1984b), p. 1942 [EE 1226a, 26-27].  See also, Aristotle (1984c), p. 1798 [NE 1139a, 13-14], for a similar statement.

7. See, for example, Aristotle (1984b), p. 1941 [EE 1225b, 34-36]: "...nor does he even choose what is possible, generally, if he does not think it in his power to do or abstain from doing it."  See also, Aristotle (1984b), p. 1942 [EE 1226a, 25-26]: "...about these [things the production of which is not in our power] none would attempt to deliberate except in ignorance".  These qualifications have suggested to some that what Aristotle meant in expressing the apparently stronger claim that deliberation requires indeterminism was actually the weaker claim that deliberation requires a belief in indeterminism.  There is some dispute about this, however.  See Gail Fine (1981), p. 572 and note 10, for a statement of this view, as against Sorabji (1980), p. 228 and pp. 245-6, who credits Aristotle with the stronger claim and argues only that Aristotle should have offered the weaker. See note 17 for further textual evidence in support of the view that Aristotle held the weaker thesis.

8. See, for example, Castañeda (1975) who accepts the Indeterministic reading.  Van Inwagen accepts something like the Indeterministic reading of (R). [See Van Inwagen (1983)].  He argues that deliberators must believe that multiple alternative possibilities are each within one's power.  However, Van Inwagen stops short of claiming that deliberators must believe that these alternative possibilities are undetermined.  Those who have understood Kant as advocating the Indeterministic reading of (R) include Castañeda, (1975), p. 134, and Thomas Hill (1985), pp. 16-17.

9. See Richard Taylor (1964) and Van Inwagen (1983), especially pp. 152-161, for similar lines of argument.  Taylor argues that a deliberator must not believe that she cannot choose among undetermined alternatives, rather than that such a deliberator must believe that she can so choose.  Van Inwagen adapts Taylor's argument in order to argue for the presence of a belief.  As mentioned in note 7, Van Inwagen argues that all deliberators necessarily believe that multiple alternative possibilities are within their power.  

10. This is not the only line of reasoning that might be used in support of the Indeterministic reading of (R).  One might begin with the claim that one can deliberate only if one believes oneself to be the ultimate source of one's actions, and infer that one must believe determinism to be false in order to do so.  Although I believe the line of reasoning set out in the text to be the one most often deployed, much of what I go on to say addresses this second line, as well.

11. My claim here is analogous in certain respects to Harry Frankfurt's claim to have found a counter-example to the "Principle of Alternate Possibilities" (see Frankfurt (1969)).  That principle states that one cannot be responsible for an action if one cannot do otherwise.  I claim here to have a counter-example to what might be called the "Principle of Belief in Alternate Possibilities," the claim that one cannot deliberate if one believes that one cannot do otherwise.

12. See, for example, Davidson (1969), p. 179.

13. See Randolph Clarke (1992), for a similar line of reasoning against the Indeterministic reading. 

14. It might also be objected that while the original voting case shows that (i) and even (iii) are false, it is nevertheless a case in which the agent believes that there is something she can do or forego (e.g., voting for Gore in a certain way), even if this is not the (primary) object of deliberation.  Thus, the case leaves open the possibility of finding reasons other than (i) through (iii) for accepting (R), including the reasoning described in note 9.  The case to follow in the text can be used to address this objection, as well.  

15. An added virtue of this case is that it is immune to the sort of objection we saw earlier concerning action individuation.  For if the agent believes that her future actions are determined by past physical states together with natural laws, then she believes that conditions are sufficient for her performing a particular action in a particular way (i.e. with certain causal antecedents).

16. A similar sort of reply is suggested by Van Inwagen (1983).  However, he offers the reply on behalf of free will skeptics who deliberate rather than on behalf of determinists who deliberate.  See pp. 157-158.

17. This notion of rationalization is somewhat broader than that introduced by Davidson (1963).  In that paper, Davidson uses "rationalization" to refer to a reason consisting of a desire (or other pro-attitude) towards actions of a certain type together with the specific belief that one's action is of that type.  My formulation allows for additional "background" beliefs, which are necessary for seeing the behavior as intelligible, to count as rationalizing.  The example which follows in the text is one in which the belief rationalizes the action in my sense even though it is not a belief of the specific form: 'my action is of a certain type.'

18. Taylor (1964), p. 77.

19. Aristotle (1984d), p. 29 [DI 9, 18b31-2].  This passage also provides further support for the claim that Aristotle did not hold the view that deliberation requires indeterminism.  For if he had held that view, then it would have been natural for him to identify the claim that we could not deliberate as an absurd conseqence of indeterminism.  But in fact, he identifies the weaker claim that deliberation would have no point as the absurd consequence of the determinism.

20. There are ways to resist the Indeterministic reading other than the one I have set out here.  Some have described the phenomenology of deliberation and decision in a way that undermines the Indeterministic reading (see, for example, Mele (1995),  pp. 133-36 and Strawson (1986), p. 115, note 30), and others have argued against the Indeterministic reading by providing diagnoses of why we mistakenly take the view to be true.  One such diagnosis is that the Indeterministic reading of (R) is confused with the true claim that we human beings have a sense that our actions are undetermined.  J. David Velleman (1989), for example, accepts this weaker thesis, and offers a psychological account of the phenomenon.  A second kind of diagnosis is that the belief in the contingency of one's future actions has been confused with a belief in the epistemic or, alternatively, the doxastic contingency of those actions.  (See Dennett (1984), pp. 112-13, Pereboom (1995), pp. 32-33, Kapitan (1986), and Pereboom (2001), pp. 136-37.)  A related diagnosis has been offered by Philip Pettit (1989), who argues that the Indeterministic reading has been confused with the following true claim: for each of a set of alternative courses of action under deliberation, a deliberator must not believe that it is not possible.  Hilary Bok (1998) argues that while we have reason to regard ourselves as having genuine alternatives, this means that we must see ourselves as having alternatives in the sense that there are multiple actions we would perform if we chose (pp. 110-114).  George Thomas (unpublished manuscript) defends what I take to be the most promising of these accounts: rational deliberators must believe that they can choose among alternate possibilities where the alternate possibilities are those actions that are not made impossible by anything that does not proceed via the process of deliberation itself.  (See also Dennett (1984), p. 115-122 for a related suggestion.)  I discuss all of these suggestions in Nelkin (in-preparation).

21. Rawls (1971), p. 5.

22. See Honderich (1988) for the view that both compatibilists and incompatibilists make the mistake of focusing on one family of important human attitudes to the exclusion of another with which it is inconsistent.  I cannot do justice to Honderich's rich discussion here.  For our purposes here, it is worth noting that it is consistent with there being a single concept of freedom, and even one to which all rational deliberators are in some way committed, that human beings often possess inconsistent attitudes concerning particular conceptions of freedom.            

23. See, for example, Roderick Chisholm (1964).

24. See Watson (1975).

25. In this connection, see Frank Jackson (1998) who writes: "I find compelling Peter Van Inwagen's argument that [...] determinism is inconsistent with free will.  What compatibilist arguments show, or so it seems to me, is [...] that free action on a conception near enough to the folk's [i.e., common-sense] to be regarded as a natural extension of it, and which does the theoretical job we folk give the concept of free action in adjudicating questions of moral responsibility and punishment, and in governing our attitudes to the actions of those around us, is compatible with determinism.  (pp. 44-45).  Although the concept that I offer is neutral as between compatibilism and incompatibilism, the spirit of Jackson's point applies to it, as well.  Even if one doubts that the concept I offer is the concept of freedom, one can still accept that it can do the theoretical work we want it to do, including supporting our attributions of moral responsibility.

26. See, for example, Frankfurt (1971).

27. Bok (1998) has recently defended what might naturally be thought of as an explicitly compatibilist reading of (R).  In particular, she argues that being practical reasoners gives us reason to regard ourselves as free in a compatibilist sense.  We are free in the relevant sense when we can determine our conduct through practical reasoning, and we have genuine alternatives among which we can choose, where genuine alternatives are those actions we would perform if we chose to perform them.  See especially pp. 118-119.    

28. Strawson's rejection of (R) may ultimately rest on an understanding of rational deliberation that differs from mine on just this point.  For example, in (1986), he describes the rational deliberation of an imaginary being, the Spectator, who is "experientially detached from her desires-from her motivations generally-in some curious way."  (p. 234).  For her, rational deliberation is a series of "practical-rational calculations going on in" a person in such a way that the person need have no sense that she is the decider and rational planner of action. (p. 235).  Although Strawson sees this as an anomolous case of rational deliberation, it reveals a possible divergence from his conception of rational deliberation and my own.  For a less recent, but very explicit, example of a conception of rational deliberation very different from mine on this point, see Hobbes, who describes deliberation in the following way: "When in the mind of man appetites and aversions, hopes and fears, concerning one and the same thing arise alternately, and diverse good and evil consequences of the doing or omitting the thing propounded come successively into our thoughts, so that sometimes we have an appetite to it, sometimes an aversion from it...the whole sum of desires, aversions, hopes and fears, continued till the thing be either done or thought impossible, is that we call DELIBERATION."  (1668/1994), p. 33. 

29. Frederick Adams develops a similar idea in Adams (1995), p. 552.  Adams there argues that all intentional action requires an attempt, and that trying to perform an action requires the lack of a belief that success is impossible.  Thus, it would seem to follow that intentional action requires the lack of a belief that success is impossible.  Sometimes Adams also seems to endorse the stronger claim that intentional action requires the belief in the possibility of success (see pp. 553-554, for example).  And the stronger claim fits well with his reasoning that intentional action requires that one have beliefs about how to achieve one's end.

30. As Burge has pointed out, one might have "incomplete mastery" of a concept, have false beliefs about even some of the essential properties of its instances, and yet have genuine beliefs employing the concept, nevertheless.  For example, one might believe that one suffers from arthritis, even if one believes it is not a disease of the joints.  (see Burge (1979)).  However, it may be that there are certain true beliefs that one cannot lack, and still be said to have the concept.  The case at hand appears to be of this kind: it is constitutive of having the concept of having reasons that one believe one ought to act in certain ways if one has reasons to act.

31. This idea, too, recalls Kant (1781/1965): "That our reason has causality, or that we represent it to ourselves as having causality, is evident from the imperatives which in all matters of conduct we impose as rules upon our active powers.  'Ought' expresses a kind of necessity and connection with grounds which is found nowhere else in the whole of nature..." (pp. 472-473 [A457/B575]).

32. Along lines similar to the objection in the text, it might be argued that in rationally deliberating, one is sometimes guided only by the purpose of finding the best thing to.  Yet, as I argued earlier, if one's activity toward the goal of finding the best thing to is to constitute genuine rational deliberation (as opposed to other activities that might aim at that goal, such as making a sacrifice to the gods), one must also be guided by the purpose of finding good reasons for acting. 

33. For a small sampling of the literature on the connections among intentional action, intention, and belief, see Davidson (1985) who argues that intentional action requires the belief that one can succeed, Grice (1971), Harman (1976), and Velleman (1985) who argue that having an intention requires the stronger belief (or acceptance, in the case of Grice) that one will succeed, and Bratman (1986) who argues that (normally) having an intention and being rational requires that one not believe that one will not succeed.  Many have argued against one or more of these claims.  For example, McCann (1986) argues against all of these claims.  Ludwig (1992) and (1995) also argues against all of these claims, and goes one step further.  He defends the claim that one can be rational in both intending and acting intentionally even though one believes that one cannot succeed.  I am grateful to Kirk Ludwig for an e-mail correspondence about his view.            

34. See Ludwig (1992), p. 263.

35. See Ludwig (1995), p. 566-568.

36. Albritton (1985) argues that one can try to do what one believes is impossible, but, like Ludwig, also recognizes the need to account for cases like the flying case.  For example, he agrees that there are certain things he cannot try to do, including trying to jump over a building and even trying to do fifty push-up.  His explanation for this fact is that "in his present cognitive position and state of mind that description of him [as trying] would be inept whatever he did," and "It's that nothing I can think of to do this evening would be rightly described as trying to jump over this building, in a straightforward sense, unless, for example, my beliefs were to alter or go very dim." (1985), p. 245.  It seems to me that something important must be contained in the phrases, "present cognitive position" and "my beliefs."  It is tempting to take them to include the lack of belief that success is possible or the lack of belief that one has the general ability to perform these kinds of actions.

37. One consequence of this reasoning is that there is a certain kind of irrationality in simultaneously being a rational deliberator and a "practical reasons-nihilist" or even a skeptic about practical reasons.  Against this, it might be argued that a skeptic about reasons could rationally deliberate (and be perfectly rational) simply by seeking reasons if they happen to exist.  In reply, as argued earlier, one could not deliberate if one lacked the belief that one can succeed in at least some typical situations; otherwise, one would lack a conception of what one was doing in deliberating.  Since deliberation requires such a conception, one must believe that one can sometimes succeed in finding reasons for acting.  Importantly, this leaves open the possibility of a rational skepticism about the possibility of finding reasons on a particular occasion.  See also Burge (1998) for a different argument that finds a similar, albeit more general, target in a broad reasons skepticism. Burge argues persuasively that in order for one to fully understand reasons, "one must be susceptible to reasons," and "one must recognize" the effect of reasons on one's judgements and inferences (p. 250).

38. The initial appeal of the examples also fades upon examination of the growing body of work on infant and child development.  Developmental psychologists are divided on the question of exactly when many of the relevant concepts, such as 'self, 'goal', 'desires', and 'desires as reasons for acting' emerge in the human infant and child.  And while psychologists see the field as one that is need of a great deal more research (see, for example, Wellman and Iagaki (1997), p. 2 and Meltzoff et al. (1999), p. 19), there is an increasing consensus that such concepts appear much earlier in human development than was previously thought. (see, for example, Wellman and Iagaki (1997) and the essays in Part II of Malle, Moses, and Baldwin (2001).  For example, it is argued that by the age of 18 months, many children not only have the mental state concepts of 'desire' and 'intention' but also understand that others have desires different from their own and that others can intend to perform actions even though they are prevented from performing them (Meltzoff et al. (1999)).  Thus, there is good evidence that very young children have rich mental state concepts.  Of course, it is true that even if this evidence were conclusive, it does not provide a positive argument for the Belief-Concept reading of (R).  At the same time, the available evidence detracts considerably from the initial appeal of the alleged counterexamples.  

39. The objection might be pressed in the following way: Consider theoretical deliberation.   When engaged in it, we can suppose that we have a sense that we can believe or judge for good reasons.  But there is nothing like a sense of freedom associated with theoretical deliberation; to the contrary, we do not choose our judgements as we do many of our actions.  So perhaps the sense of freedom as I have characterized it is not really a sense of freedom either.  This way of pressing the objection raises a number of interesting issues regarding the relationship between rational deliberation and its theoretical parallel.  For the objection makes a number of presuppositions including these: (1) that rational deliberation and its parallel in the realm of judgement differ in that freedom and "chosenness" are associated with the former but not the latter, and (2) that the two forms of deliberation do not differ in any relevant way in the implications of their respective requirements that agents represent themselves as capable of responding well to good reasons.  Both presuppositions might be challenged, and I discuss both options in Nelkin (in-preparation).


HOME to Det & Free Website front page
HOME to T.H. website front page