Ultimacy and Sweet Jane
by Michael McKenna

There has long been a line of reflection on determinism and freedom that involves the idea that if determinism is true then your choice now to go on reading this sentence is just the effect of a sequence of events stretching into the distant past, back to before you came into existence -- and hence that you could only be responsible for going on reading this sentence if you were responsible for things that happened before you came into existence. Michael McKenna's essay is effective and persistent reflection on such lines of thinking. It seeks to support something like the traditional view that determinism is compatible with freedom and responsibility -- despite a doubt as to our having a single and stable sense of responsibility.

Some people, they like to go out dancing
And other peoples, they have to work
And there’s even some evil mothers
Well they’re gonna tell you that everything is just dirt
You know, that women, never really faint
And that villains always blink their eyes
And that, children are the only ones who really blush
And that, life is just to die.
And, everyone who had a heart,
They wouldn’t turn around and break it
And that everyone who played a part
Oh wouldn’t turn around and hate it.
– Lou Reed, “Sweet Jane”

Lou Reed’s “Sweet Jane” is one of my favourite rock and roll songs from one of my favourite albums, the Velvet Underground’s Loaded released in 1970. The narrator, who is in a rock and roll band, tells us of Jack and Jane, who save their monies and sit down by the fire when they come home from work. I am not sure what Lou Reed wanted us to take from this song, but when I hear it I think about how people find meaning in their lives. Sure, some people are in a rock and roll band, and others just go to work, saving their monies and listening to classical music on the radio when they come home at night. But however they do it, many people (if not most) find something that is for them their passion. Some people like to go out dancing, but Jack and Jane have something different.  Why? Because heavenly wine and roses whisper to Jane when Jack smiles. And one can only imagine that this is good news for Jack as well as Jane [1].

Furthermore, the song has protagonists, who deny that there is anything of worth about going out and dancing, singing in a rock and roll band, or saving your money and listening to classical music on the radio. According to these sceptics, when sweet Jane hears heavenly wine and roses whisper when Jack smiles, she’s just deluded. These sceptics tell you that everything is just dirt. Life has no meaning. Nothing is lovely or tender. None but innocent children can have gentle, fragile feelings. Even villains are inauthentic. It turns out life is just to die. Yet, these sceptics explain, they would not turn around and break everyone’s heart. They do not mean to say that everyone who plays a part—who invests in their lives—would turn around and eventually hate their parts. But, as Reed’s sneering cynicism makes clear, he doesn’t buy their apologies. If life is just to die, that is heartbreaking; that would be reason to turn around and hate your life.

I do not want to fix upon meaning in life, though I suppose what I do want to consider—free and morally responsible agency—is closely related. How so? It is natural to find many of the activities and achievements in our lives meaningful in large part because we regard ourselves as having freely invested in them and as responsible for them (morally and otherwise). Furthermore, like the very idea of meaning in life, the notion of free and morally responsible agency has philosophical sceptics, and it is with them that I shall take issue. Unlike Reed, I do not regard the sceptics whose arguments I want to confront as evil. But there is something about one of their arguments that is like the standpoint that Reed confronts, one in which everything is just dirt, and in which, if it is, nothing that is apparently wonderful is quite as it seems to be. If so, it is heartbreaking, or at least damn disappointing. 

In this paper, I shall do three things. First, I will identify two variations on an argument known as the Ultimacy Argument, which invokes a form of scepticism that I shall compare to the one Lou Reed faces. I shall show why I think either formulation of the Ultimacy Argument is no good. Second, I will offer an account of ultimacy that I think normal persons like us can satisfy, even in the absence of grandiose presuppositions about our metaphysical status. Third and finally, I will offer a diagnosis of why some have hoped for a kind of ultimacy that is beyond reach. This will bring us back to the fear that we might have reason to hate the part in life that we have come to play.

    Before turning to the Ultimacy Argument, I begin with some definitions. Morally responsible agents are persons for whom moral demands are fitting. They are accountable for their behaviour and are fair targets of moral praise and blame, as well as reward and punishment. They are also apt targets of the morally reactive attitudes, which are attitudes revealing interpersonal expectations of due regard (moral and otherwise) for others. I understand free will as satisfying all that is required for the control condition for morally responsible agency. Persons are morally responsible only if they are able to control their conduct. Free will just is the capacity that gives persons the relevant sort of control required for morally responsible agency. A free action, as I shall understand it, is one that issues from an agent’s exercise of her free will [2].

Determinism is the thesis that the temporally non-relational facts at a time in conjunction with the laws of nature entail every truth about every later time. Thus, if determinism is true, the facts of the remote past, say just after the big bang, in conjunction with the laws of nature, entail every truth about every later time. Assuming determinism is true, if you ate cornflakes for breakfast, or waived hello to a friend this morning, or chose to sit down just when you did as you began to read this paper, each of these facts is entailed by some facts about the remote past and the laws of nature. I shall take naturalism to be a weaker thesis than determinism, though similar in relevant respects. Naturalism, as I shall understand it, is the thesis that everything, including the conduct of persons like us, is entirely the product (deterministic or indeterministic) of conditions that can in principle be characterized in terms of our most basic sciences [3]. Thus, even if the universe proceeds only according to indeterministic, probabilistic laws, everything in it is entirely the outgrowth of a natural order of which we are but a part. 

    Incompatibilists hold that, if determinism is true, no one has free will. If so, no one controls her conduct or her life’s path in a manner that is required for moral responsibility. Hence, if determinism is true, no one is morally responsible. Some incompatibilists are inclined to treat naturalism the same way [4]. If naturalism is true, there is no place for free will or moral responsibility in the world. At best, they are illusions. A salient feature of incompatibilism as I understand it is that the absence of free will at a deterministic world is due to the fact that determinism is true (and for some incompatibilists, a similar point applies to naturalism). Compatibilists disagree. They hold that the truth of determinism is not on its own a reason to think that no person has free will and that no person is ever morally responsible. And compatibilists reason similarly about naturalism. To stretch the metaphor in “Sweet Jane,” with regard to freedom and responsibility, incompatibilists hold, roughly, that if we persons arise exclusively from dirt (or the bumping together of cosmic dust), then human agency in one of its fullest forms, morally responsible agency, is just dirt, nothing more. Compatibilists, on the other hand, will grant that we might well arise from dirt (the blind interactions of cosmic dust), but from dirt alone comes considerably more: free persons who are morally responsible and who control their actions and the trajectory of their lives [5].

    I am a compatibilist and a naturalist. While I have arguments for my compatibilism, I have none for my naturalism. Naturalism is something I believe is true; I have no idea how to argue for it. Regardless, here I will not offer any arguments to advance my own theses. Rather, I will respond to one of my opponents’ arguments against compatibilism, the Ultimacy Argument. 

1. The Ultimacy Argument: An Impossibilist Version

Here is a first pass. Assume, as seems reasonable, a person acts as she does when she is (allegedly) free due to her state of mind at the time. She might act from passion, out of anger, on calm cool reason, but whatever it is, her agency is engaged by virtue of features of her mental economy. When she is ‘in control’ her mental life produces her actions in a non-deviant fashion. Just to simplify matters, suppose we think of reasons on the classic Humean model as belief-desire pairs. If an agent, Ann, acts to steal a loaf of bread, there is some combination of belief and desire that is Ann’s reason why she so acts, a reason we can abbreviate as BD. If she is morally responsible for stealing the bread, and if she acted freely in doing so, then it seems that both her responsibility and her freedom are due to her reason, BD, that is, are due to the way she is mentally. But if she is not morally responsible for the way she is mentally at the time at which she acts, then she is not morally responsible for her action, since her action is due to the way she is mentally. What would be required for her to be morally responsible for BD? Well, she would have to be responsible for her freely acquiring BD. So the way she was mentally in the acquisition of BD would have to be the product of something she did freely and for which she is morally responsible. But then there must have been some other belief-desire pair, BD′, that non-deviantly led to her acquisition of BD, and for her to be morally responsible for BD, she must be morally responsible for BD′. And so on, ad infinitum. 

It turns out that, on this version of the Ultimacy Argument, a person is morally responsible for her actions only if she freely chose and brought about her own self. But of course, no one, at least no person short of God, is a self-cause. This is impossible; therefore, so is morally responsible agency. Consider Galen Strawson’s formulation:

1. Nothing can be causa sui—nothing can be the cause of itself.
2. In order to be truly morally responsible for one’s actions one would have to be causa sui, at least in certain crucial mental respects.
3. Therefore, nothing can be truly morally responsible [6].

This version of the Ultimacy Argument is extremely strong [7]. Note that determinism does not play any role in the argument at all. In fact, it is not really an argument for incompatibilism in the traditional sense (and as I have defined incompatibilism above), since it does not say that determinism is what rules out freedom and moral responsibility.  Rather, the very concept of moral responsibility is incoherent. It has the seeds in it of a demand that is metaphysically impossible to achieve (for finite beings). So, let us label this version of the Ultimacy Argument an impossibilist version [8]. We can take this argument for impossibilism about freedom and responsibility to feature the thesis that free and morally responsible agency is impossible due to the very concept of responsibility (or freedom), and not because of any fact about the universe (such as that it is deterministic, or instead indeterministic only in ways consistent with naturalism). 

      In my estimation, this impossibilist version of the Ultimacy Argument is poor.  Its weakness is found in the second premise. Were we to construct a formal argument for it from the resources given above, we would find a questionable premise. It is:

P: You are morally responsible for what you do only if you are morally responsible for the way you are mentally (as it bears on what you do).

A related premise might be put in terms of freedom or control instead:

P′: You act freely in what you do only if you are free with respect to the way you are mentally (as it bears on what you do).

I think that neither P nor P′ is credible [9]. I’ll focus just on P′. Many philosophers analyze free will, at least in part, in terms of what is up to an agent. (I myself do not analyze it this way, but that is of no importance here.) Suppose, as in the case of Ann above, she steals the bread due to BD. BD captures how she was mentally at the time at which she acted. In order for it to be up to Ann that she steals the loaf of bread, must it be, as P′ requires, up to her that she is the way she is mentally in that she possesses BD? If it must be, an argument needs to be given, one that goes beyond the Ultimacy Argument. Cannot one argue that even if it is not up to Ann how she is mentally—that is, whether she possesses BD—it is up to her whether she acts on BD? This criticism is especially forceful for incompatibilists who hold that in the absence of determinism persons might be free and responsible. According to them, it could very well be that at a certain time, a person has no direct control over how she is mentally, say what beliefs and desires she has, but if it is causally open to her that she act on any belief-desire pair or not, her freedom can be found at these junctures. 

    This criticism of the argument is also available to the compatibilist. Mightn’t Ann be directly free with respect to acting on her belief and desire pair even if she is not directly free with respect to her possession of that belief and desire pair? On a compatibilist view of the sort I endorse, Ann might very well be free in so acting. Think of it this way: Amongst the many casually related events that will unfold in the history of the world, somewhere in the unfolding of them a newly developing person will be an effect of prior causes. She will come to have various states of mind that will be the product of factors that are patently beyond her control, genetic traits, parental influence, the vagaries of luck, and so on. At some point, in the interstices of various casually related events, this agent will act in such a manner that she will exercise control over her conduct. From conditions that lack control, control will arise. Or, to return to my strained metaphor, from mere dirt, something much more than dirt will arise. No compatibilist can deny that something like this must be how, at a determined world, free and responsible agency first appears on the scene [10].

Before turning to a different version of the Ultimacy Argument, consider the picture of moral responsibility that impossibilists like Galen Strawson hold. Strawson takes it that our concept of moral responsibility is one that has built right into it an exceedingly high standard [11]. Examining the concept so construed, Strawson then argues that nothing could possibly satisfy it. But why understand the concept this way if it turns out that nothing short of a deity could make the cut? Why not think that this way of understanding responsibility could not be the one that is ours, could not be the one that has a bearing on our concerns, since it sets the bar so very far beyond us? Shouldn’t we revisit our initial presuppositions that led us down this rabbit hole and see if there is a conception of freedom, responsibility, and ultimacy that has merit, speaks to our concerns as finite creatures, and is, at the very least, coherent? [12] To avoid misunderstanding, I do not regard these as rhetorical questions. Maybe there is reason to think that Strawson’s way of understanding moral responsibility does capture our folk concept. Maybe we have no license beyond theoretical bias for fashioning a different concept from the one the folk have fashioned. I revisit these questions in closing.   

2. The Ultimacy Argument: An Incompatibilist Version

I turn to an incompatibilist rather than an impossibilist version of the Ultimacy Argument, one that relies upon at least one premise having to do with determinism. This version is due to Saul Smilansky [13]. According to Smilansky, if determinism is true, then all of our conduct is just part of the “unfolding of the given” [14]. Assuming determinism, anything a person does is but an outgrowth of factors that are ultimately beyond her control, since the origin of her conduct is found in sufficient casual springs obtaining before she was even born [15] According to Smilansky, no one is the ultimate source of her actions if determinism is true. 

How should we understand ultimacy? Although he does not present the claim explicitly, Smilansky, it seems, is committed to the following:

U:  An agent, A, is the ultimate source of her action D only if she contributes some necessary condition, C, to D such that there are no sufficient conditions for C that obtain independently of A.

Call this the U condition [16]. Taking the U condition as a partial definition of ultimacy (partial because it identifies only a necessary and not a sufficient condition), we can now construct an incompatibilist version of the Ultimacy Argument:

1. A person acts freely in the sense required for true moral responsibility only if she is the ultimate source of her action.
2. If determinism is true, no one is the ultimate source of her actions.
3. Therefore, if determinism is true, no one acts freely in the sense required for true moral responsibility.

Notice that this version is not as demanding as Strawson’s impossibilist version. It leaves open that if determinism is false, perhaps in a way that reaches beyond naturalism, it might be possible to satisfy the requirement of ultimacy, and so for agents to act freely [17].

    Before stating where I think the argument fails, I pause to note further insights that Smilansky appends to his conclusion. According to him, if determinism is true and therefore no one is ultimately free and responsible, it does not follow that persons are not free or responsible in some sense. Compatibilist distinctions help us sort some persons as free and responsible in an anemic way, and others as not free and responsible even in an anemic way (such as young children, the mentally ill, and so forth). This has, Smilansky explains, great social value. Some persons can be swayed by our moral reproach; others can’t. Furthermore, there is a kind of dignity that can be accorded to determined persons within a moral responsibility community, and we need to mark those who do violence to our moral demands and those who support them. Compatibilist distinctions help make that cut and apply moral reproach properly, even if, at some deeper level, it all is merely the playing out of a moral order for which no one is ‘really’ free or morally responsible. 

Also, according to Smilansky, we operate under the illusion that we have more than compatibilist free will; we believe that we have freedom in a strong sense and are truly morally responsible. This helps us to face a life that we would otherwise find dark were we to face the fact that true, deep, ultimate moral responsibility is beyond our reach.  On Smilansky’s view, the freedom and responsibility compatibilism offers us is morally shallow [18]. It lacks depth. Who we are and how we act is the upshot of nothing but luck, and following Thomas Nagel [19], Smilansky argues, luck embedded in the creation of our own agency erodes control, and with it freedom and responsibility [20]. For each of us, by sheer luck, temporally distant forces set in motion long before any of us were born have settled for us who we will be and how we will act. 

We all would much prefer, Smilansky maintains, to be ultimate sources of our actions. Without ultimacy, the objective truth of the matter is that our lives do not matter nearly as much as we assume they do. So we live precariously relying upon our illusion of deep freedom and responsibility. Here Smilansky advises that we shield people from the ugly truth that, if determinism is true, no human achievement or failing is ever as it appears. To my ear, this is the analogue of saying that there is nothing but illusion to what Jane hears when Jack smiles.

    But is the Ultimacy Argument upon which Smilanksy builds these further theses sound? I do not think so. Each premise is open to assault. In the remainder of this section I will focus on the first premise. In the next section I will focus on the second. To begin, consider the first premise, that the strong sense of freedom relevant to genuine moral responsibility requires ultimacy. If we simply give the incompatibilist her definition of ultimacy as (partially) specified by U, then undoubtedly no compatibilist can hold that ultimacy is necessary for freedom and responsibility. U demands causally indeterministic breaks in the internal etiology of agency. Compatibilism holds that no such breaks are necessary. But now, if the compatibilist will simply allow the incompatibilist this definition of ultimacy as a matter of stipulation, then it is unclear why the premise should be regarded as true. The compatibilist will claim that something that lacks U is sufficient for freedom (and moral responsibility) and this is compatible with determinism. Let us call it CF for compatibilist freedom. If CF is sufficient for the strong sense of freedom required for true moral responsibility, then ultimacy (and U) is not necessary. So the first premise is false. 

    Of course, Smilansky and his incompatibilist compatriots will shoot back that he and they do not take U to be mere stipulation. Its truth captures the real nature of the freedom and responsibility that we all presuppose. But if Smilansky takes U to be more than a stipulative (partial) definition of ultimacy, then he needs an argument. Notice that the content of U is negative. The necessary condition that it demands is an absence—an absence of a certain sort of causal history. But if this is all the content there is to U, that amounts just to the demand that compatibilism is false. Here, it seems that U simply begs the question against the compatibilist [21]. If so, then it is dialectically inappropriate for the incompatibilist to put forward the Ultimacy Argument, since it does not amount to an argument at all but rather a mere expression of the incompatibilist thesis. Maybe the incompatibilist is right. But claiming that U is the reason that we should accept that she is right, and is the reason we should reject the sufficiency of CF, is tantamount to claiming that she is right because she is. 

    Because the challenge of begging the question is one of my central challenges to Smilansky and other incompatibilists who defend this version of the Ultimacy Argument, I pause to make explicit what I mean by the fallacy of begging the question. Recently, this expression has been bandied about loosely in the free will dispute, and some have understandably come to think that the challenge reduces to no more than an expression of philosophical disagreement [22]. I take the charge, however, to amount to a genuine fallacy of reasoning, one that involves a failure to appreciate properly the dialectical demands of proper argumentation as regards some disagreement. An argument begs the question just in case it imports without independent argumentation a premise that presupposes everything that is controversially at issue in a debate regarding the conclusion that is in contention. If this incompatibilist argument seeks the conclusion that determinism rules out free will or moral responsibility, then the premise that ultimacy is required for free will or moral responsibility had better have some intuitive clout beyond the thought that determinism makes it so that no one is free or morally responsible.  

    To convince us that U has argumentative clout, the negative incompatibilist condition that it asserts must be shown to flow from some positive aspect of agency that freedom and responsibility require. All of Smilansky’s entreaties to give substance to U are oriented around claims such as the shallowness that the lack of U would bring to our understanding of our moral status, and the depth that possession of U affords us. Our lives would be merely the products of luck, and we would find this disheartening, whereas with U it would be up to us and not mere luck how we acted and how our lives unfolded. Others, such as Robert Kane [23], have also advanced a similar U condition, and like Smilansky have offered positive accounts of free and morally responsible agency that invoke the requirement of ultimacy. 

I have never been able to sort out exactly why Kane, Smilansky and other likeminded theorists are convinced that the considerations they invoke give us a non-question begging reason to accept U (or something like it) as it is supposed to figure in an argument for incompatibilism [24]. In some of the considerations to which they appeal, such as depth, there is a compatibilist account in the offing, and so it is unclear that a compatibilist cannot have the depth that the incompatibilist says we must [25]. In other cases, such as luck, it seems that construed in a certain way, compatibilists will have to admit that luck does creep in, but then it turns out that the luck at issue is bad just because it makes it so that there are causally sufficient conditions for what we do that we do not control. This, however, is again just the question-begging demand for incompatibilist freedom.

Space does not permit an exhaustive examination of the reasons that might be advanced on behalf of Smilansky or Kane as a positive basis for U. I will consider just briefly one of Smilansky’s key considerations, which relies upon the notions of fairness, justice and desert. Consider this key passage from Smilansky:

We can make sense of the notion of autonomy or self-determination on the compatibilist level, but if there is no libertarian free will, no one can be ultimately in control, ultimately responsible, for this self and its determinations. All that takes place on the compatibilist level, irrespective of the local distinction in respect of control, becomes on the ultimate level ‘what was merely there’, ultimately deriving from causes beyond the control of the participants…
The difficulty is that there is an ethical basis for the libertarian requirements, and, even if it cannot be fulfilled, the idea of ‘simply dropping it’ masks how problematic the result may be in terms of fairness and justice. The fact remains that if there is no libertarian free will, a person being punished may suffer justly in compatibilist terms for what is ultimately her luck, for what follows from being what she is—ultimately without her control, a state which she had no real opportunity to alter, hence not her responsibility and fault [26].

There are at least three key assumptions here, two of which I shall focus upon. One, which I will not consider, is that an essential element of moral responsibility (and presumably blameworthiness) involves just suffering, presumably based on desert [27]. Second, ultimate control is missing on a compatibilist construal since at the “ultimate level” it all arises from what was “merely there”. But what is ultimate control? There is no positive account of it. What it turns out to be is this: control that does not issue from a deterministic history. That is just the question-begging negative condition again. Third, fairness and justice are unattainable on a compatibilist construal since what a person ultimately is and how she exercises her (compatibilist) control is a matter of luck. Why is this unfair and thus unjust? Because a person fully determined had no “real” opportunity to alter her future, and so on. The problem is that compatibilists have their own notions of opportunity, and of the luck that carries with it unfairness in contrast with the luck that does not. Some luck, as in the case of the truck driver who drives home drunk hitting no one in contrast with the one who kills a child, does involve (what seems to be) unfairness should one driver be held morally blameworthy through bad luck and another not held morally blameworthy due only to good luck. But the luck here can be teased out from the exercises of free agency that a compatibilist advances and that should be a proper basis for moral blameworthiness.  

    What Smilansky seems to have in mind by luck and justice is something like the unfitting relation between what one sometimes gets in a distribution of positive and negative goods and what it would be fair to get. A person who, through luck, has doled out to her a life’s trajectory in which she is (in compatibilist terms) blameworthy for many acts and praiseworthy for only a few is unlucky in that she gets lots of negative social and psychic goods and few positive ones in contrast with the person whose life trajectory takes a morally laudatory path. Now, on Smilansky’s picture, if there were a way for the former person to avoid this path, or if the path she took were ‘truly’ her doing, then an unequal distribution of the relevant goods might be fair and so not unjust. 

In reply, consider how I have fleshed out this concern regarding fairness and justice. On a compatibilist construal, according to Smilansky, a person cannot avoid the life path she is on, and the path she takes is not truly her doing. This is what gives rise to the complaint of unfairness. However, the compatibilist can simply deny these assertions and offer compatibilist versions of what an agent can avoid and (or) what is her doing.  Granted, Smilansky and other advocates of the Ultimacy Argument can counter that these compatibilist notions are insufficient. But then, why? Is the reason just this: All of these compatibilist notions permit causally sufficient conditions obtaining prior to an agent’s existence. If this is the reason, again we are left with a question-begging incompatibilist demand. 

An incompatibilist might fight to secure her conception of ultimacy by arguing for it. This is, for instance, how Derk Pereboom defends a condition very close to U, a condition he characterizes in terms of origination:

O:  If an agent is morally responsible for a decision, then she must be its source in the sense that she exercises control in making this decision, and she must also be its ultimate source—and then the action cannot be deterministically produced by factors beyond her control, it cannot be produced by nothing at all, and it cannot be such that factors beyond her control contribute to its production but do not deterministically produce it, while there is nothing that supplements the contribution of these factors to produce the decision [28].

But the difference between Pereboom’s stance towards O and the stance that Smilansky and Kane take towards U is that Pereboom maintains that O follows from an argument for incompatibilism. It is not itself to be understood as figuring in a premise of an argument for incompatibilism. As Pereboom sees it, a manipulation argument (his is labelled the Four-Case Argument) shows that O is true, and that determinism is incompatible with free will and moral responsibility. 

    A manipulation argument for incompatibilism proceeds by developing a case in which an agent is manipulated into performing an action. Then two claims are established: first, that the manipulation undermines the agent’s freedom or her moral responsibility; and second, that the agent was manipulated so that she satisfied all that is required by compatibilist standards and that the manipulation is no different than a mere deterministic process. I believe that one variant or another of a manipulation argument for incompatibilism offers a serious challenge to the compatibilist [29]. Elsewhere I have attempted to resist on compatibilist grounds [30]. Regardless, all that concerns me here is that if the incompatibilist wishes to establish U by appeal to a manipulation argument, then the Ultimacy Argument itself does no work. In this way, the incompatibilist can evade my prior challenge of attempting to establish U without begging the question, but only by shifting the argument to a different (and to my mind better) one.  

In closing this section, my challenge to Smilansky and other advocates of the Ultimacy Argument is to produce some positive account of the element of free action that ultimacy affords and that a compatibilist cannot provide. Until then, I remain sceptical that any argumentative force should be granted to the U condition as an aid in an argument for incompatibilism. 

3. A Compatibilist Account of Ultimacy

Another way to resist the incompatibilist’s Ultimacy Argument is to resist the second premise: Deny that determinism is incompatible with a person’s being the ultimate source of her actions. Of course, there is no way for the compatibilist to fit an account of ultimacy to U. The condition U simply cannot be satisfied at a determined world. I offer two points in the service of a compatibilist account of ultimacy. The first is a challenge to the presumption that ultimacy must be characterized in terms of causal origination. The second is a sketch of ultimacy that does not conflict with determinism. 

    As to the first point, note that there are various notions of origination that do not presuppose the initiation of a causal sequence. Consider the origin of species. One species, Homo sapiens for example, arose from variation in another, Homo erectus. At some point along the way, there was a ‘new kind’. Here we find an originating source, and yet there is no reason to think that there was a distinct causal break in the chain between species. Or consider the case of the famed Perrier sparkling water, whose source or origin is somewhere in France [31]. There is no reason to think that the true source of the water is not itself something that was caused (deterministically or indeterministically) due to a long history of geological events. Or think of such cases as the source of a river like the Mississippi or the Nile. Suppose that someone were to say that, if determinism is true, then the ultimate source of all of the examples mentioned here is, really, the Big Bang. The other sources offered were merely mediated sources. One could say this, but in any but a highly contrived context it would be an artifice. 

On the proposal I offer, sourcehood, even with the modifier ‘ultimate’, is a context sensitive notion. We want to know about—we seek an explanation for—one sort of thing holding fixed other natural facts. Of course, if ‘ultimacy’ is nothing but an incompatibilist’s term of art, then indeed the ultimate source of Perrier drinking water must be something like the Big Bang. But if our goal is to get a grip on what in ordinary thought one might have in mind in speaking of ultimacy—something that a theorist can then apply without bias to the notion of agency as Smilansky and others intend to do—then we can come to understand the notion by contrasting mediated as opposed to ultimate sources:

A:    Where did this water come from?
B:      You mean that bottle of water, the one in your hand? From the refrigerator.
A:  No, I mean what is the source of the water?
B:      You mean where did I buy it? The grocery store.
A:  No, I mean where in the world did the water in the bottle originate?
B:      In a bottling plant somewhere in Europe.
A:  No. I mean what is the geological source of this water? Where ultimately did the
     liquid in the bottle originate? Was it, for example, from Saratoga Springs, NY?
B:    Oh, from the famous spring in southern France.
A:    Right. Thank you. That was what I wanted to know.

It could be that in the exchange above, speaker A was a geology professor and B a student. And it might be that there is some further causal story about shifting paths of underground water. In this case, the professor’s inquiry would not be at its end and the ultimate source not yet discerned. But setting aside a possibility like this one, as the exchange unfolded, until B gave the last reply, the only sources he offered, given relevant explanatory interests regarding origins, were mediated sources. The last one, by contrast, was ultimate.

Why can’t the compatibilist claim this about free actions? Suppose it is assumed that determinism is true. Now imagine that I have a group of friends with me on one of my insanely long, mountainous and especially taxing bicycle rides. One of the crew cries out, ‘Whose hair-brained idea was this?’ It might have been that various members of the group discussed the matter with me after I proposed it, entertained alternative routes, and so on. But if I was the one who hatched it, I can, so far as I can tell, fess up and say that it ultimately originated with me when I was studying a map a week prior. I am the ultimate source of that plan and the fine mess that I have gotten the entire group into. Could I really reply, ‘Well, determinism being true and all, though it was my idea, ultimately, it really did not originate with me—it started with the big bang’? 

I propose that sourcehood as regards the actions of persons, ultimate originating sourcehood, does not require the notion of an initiating causal sequence. It does not require a suitably placed deterministic break. I grant that an incompatibilist can disagree.  But setting aside mere stipulation, an argument is needed. A mere assertion that ultimacy entails indeterministic breaks is not on. Having said that, it seems only dialectically fair that if the compatibilist wants to challenge the second premise of the argument in question, she should have some account of ultimacy. In the space remaining, I can only offer a sketch. But rather than offering one, I’ll offer two.

Before proceeding, I offer a way to frame an approach to the notion of ultimacy. What often comes to mind when considering the demand for ultimacy is the capacity of self-determination, not (at least at first) in the extreme sense Strawson had in mind—being a self-cause—but in the perfectly ordinary sense in which we hold that some persons have fashioned their characters so that they are ‘self-made’. Through a process of personal reflection and effort of will, some come to make themselves more understanding towards others, or better fathers or mothers, or less self-centered. This, however, is perfectly consistent with a compatibilist thesis [32]. The controversy arises, however, in what Kane has called self-forming actions (SFA-s), actions that are the initiating—the ultimate—source of the character setting features by which one shapes one’s developing character. A way to approach the question of ultimacy is to ask about the status of these actions, these SFA-s. Must they satisfy a condition like U? Or can a compatibilist offer a characterization of ultimacy that helps us to understand how these SFA-s provide a proper moral foundation for later attributing to an agent moral responsibility for her own self? I proceed to sketch two compatibilist approaches.

Some notable compatibilist figures have proposed an expressive theory of moral responsibility. John Martin Fischer offers one version of an expressive theory in which the value of our responsible agency is to be understood in terms of a narrative [33]. Our actions (some of them) have significance as, so to speak, contributing to the creative enterprises which are the stories of our lives, just like sentences have meaning in the context of a novel. On this theory, an important subset of an agent’s actions can be understood as ones that ultimately originate with her if they arise as distinct contributions to the narrative that is her life. For all I know, we might be able to find the very same sentence (type) in novels by Steinbeck, Hemmingway, and O’Conner. But each of those instances in the different novels would play different roles in relation to the unfolding narratives. What gives each the sort of significance that makes it count as a salient feature of each respective novel arises out of its relation to that novel. And so it is for human agency, when the life in which actions unfold is capable of the right narrative structure. 

On this narrative model, not all of our actions will have this feature of ultimacy. Some will be mundane, having little or no bearing on the broad narrative features of our lives. Sometimes, we’re just opening the refrigerator or trying to get to work. Regardless, we operate in a space of possible actions in which it is at least sometimes a live option for us that what we do could shape the narrative contours of our lives. In these cases, we can be and sometimes are ultimate sources of our selves and our lives. 

I turn to a second compatibilist approach to ultimacy. Gary Watson [34], taking inspiration from P.F. Strawson [35], has suggested an expressive theory that is communication-based. The significance of morally responsible agency and of holding agents morally responsible is a matter of the expression of moral demands and expectations we have for one another. I follow Watson in endorsing a communication-based theory. Watson’s proposal is focused on the standpoint of holding morally responsible. By way of the morally reactive attitudes, Watson holds, we express moral demands. And a condition of morally responsible agency is that an agent be able to appreciate the demands that we make in this way. But Watson does not extend his communication-based theory to the initial actions of morally responsible agents—to the actions that are themselves candidates for evaluation from the standpoint of holding morally responsible. He does not commit to the idea that these actions can themselves be expressive in a manner that can be understood as communicating with members of the moral community who play the role of holding morally responsible (though Watson says nothing to deny this as a live option).

On my view, the actions of morally responsible agents have meaning and are candidates for moral assessment in that they reveal the moral quality of an agent’s will.  This moral quality in turn is a function of the interpretive significance that an agent does assign or is able to assign to her actions. How so? A morally responsible agent has mastered a schema in which actions can be interpreted by members of a moral community as revealing an agent’s (any agent’s) moral stance towards others. When she acts, she acts within the context of this schema. For a competent moral agent, it is always a live option that her actions could be so interpreted. Therefore, in a straightforward way, agents can be the authors of their actions, and of the salience or meaning that can be assigned to them. Here, an agent can function as the ultimate source of her action to the extent that her action has moral significance as it issues from her moral outlook. 

    These two proposals for compatibilist ultimacy are not in conflict and so do not need to be thought of as competing. What is crucial is that each offers a credible way for compatibilists to make sense of how competent morally responsible agents could be the ultimate sources of their actions.

    In light of these two compatibilist sketches of ultimacy, consider Smilansky’s charge that compatibilist notions of freedom and responsibility lack depth. Depth and shallowness are spatial metaphors, and ones that can make sense only relative to a contrastive case. (A river that is only four feet deep is shallow only because some are twenty feet deep.) But strip away the metaphor, and what does incompatibilist ultimacy buy? I have left the incompatibilists with the challenge of stating what that is. I will instead offer two competing metaphors. One is a metaphor of being knitted into a fabric; another is of being embedded into a narrative. If one’s actions can be fitted into a fabric of interpersonal expectations in our moral interactions, or if they can be embedded into a richer story about the unfolding of her life, then her actions can be seen as ultimately originating in her contribution to the fabric or the novel. I suppose that you could call either of these compatibilist proposals shallow, or say that each lacks depth. But then, we need the contrastive case. I have yet to find a clear statement of what that case is.

4. The Craving for an Ultimacy that is Beyond Our Reach

I close by returning to Galen Strawson’s impossibilist demand, and to those sceptics Reed stares down, the ones who claim that they would not have it that everyone turn around and hate their lives. In arguing for his impossibilist thesis, Strawson insists that the concept of moral responsibility that he analyzes is to be found in ordinary thought. It is our ordinary folk concept and not some eccentric one that gets him the extreme result. I think that he goes wrong in supposing that there is any single, stable concept of moral responsibility that is ours [36]. I think that ordinary thought likely has various strands from whence we might construct various distinct concepts of moral responsibility. If so, it is not as problematic as some take it to be that our ordinary concept (one of them as I see it) has this incoherent demand of self-creation built into it. For this reason, I remain open to the possibility that Strawson is right to this extent: there is a strand of our ordinary thought, in one of the ways we understand the concept of moral responsibility, that makes this extreme impossibilist demand [37].

    Indeed, it is understandable that one way that we ordinarily think about moral responsibility and freedom is in the extreme way Strawson has in mind. The reason is this: What it is reasonable for a person to want need not line up with what is possible for her. Consider the sceptic about the meaning of life. Suppose that the meaning of life is simply bound up in activities like going out and dancing, or being in a rock and roll band, or loving your smiling Jack or your sweet Jane. If that is all it comes to, then what happens when your knees give out and you cannot dance, or your drummer becomes a junkie and your band disintegrates, or when smiling Jack turns out to be a philanderer? If there were some grander meaning than that, then we could be protected from the contingencies in which our lives could lose meaning. A big, grandiose purpose for the meaning of life would ensure that we could not, at least not rationally, turn around and hate the part that we were playing in life, since the really important part would remain, for Jane, for example, even if Jack betrays Jane. Surely this is something it could be reasonable to want, even if under careful scrutiny it is just not possible, causally or metaphysically. But the philosophical mistake, to my mind, is this: Failing to find something that could do duty for this big, grandiose thing, we conclude that life has no meaning at all. Rather, the meaning it has is simply not one that gives us the grand assurances we were hoping for. At least this way, we can make the normal distinctions that seem extremely reasonable between those lives that end meaninglessly in pointless frustration, like Captain Ahab’s, and those ennobling figures whose lives are sources of inspiration, like Socrates or Joan of Arc. 

    I think the same point applies to the demand for an extreme form of ultimacy. It is understandable why people want a kind of control in which they can choose their very selves. With it, there simply is no way that life could unfold so that, once a person finds herself there, she could not simply up and change who she is. Of course, it is an incoherent demand, and a bit of reflection reveals this. But I think some people entertain it. Regardless, for finite creatures like us, there is no guarantee that a person will wind up satisfied with who she is, just as there is no guarantee that she will wind up having a meaningful life. Some fail on both fronts. Through no process which they controlled, some come to be people they most distinctly do not want to be and yet are unable to do anything about it; some find their lives’ projects and dreams shattered. A notion of freedom or the meaning of life that would protect us against these outcomes is not crazy to want, and so it is little surprise that we find elements of it in our folk concept. But once we try to articulate what that would be, we find it is not a coherent alternative. And it is the coherent ones that should inform our philosophical theorizing. So how should we proceed? By a judicious revisionist parsing away while preserving what is coherent, what we can make sense of, what does speak to our concerns. 

Return to Jack and Jane, sitting down by the fire listening to the classical music on the radio. Sure enough, there is a standpoint on the meaning of life in which, no matter how lovely Jack’s smile is to Jane, all is just dirt. But if this standpoint on the meaning of life can’t distinguish between the richness of life that Jack and Jane find as opposed to the poor wrecked Joe whose lovely Jamie dies very young of cancer or slips irretrievably into madness, why should we find it credible? What is the coherently articulated contrastive case whereby life is not dirt so that what Jack and Jane have is really nothing, is really just dirt? Sounds to me like Jack and Jane have it pretty good. Why not find the meaning of life or instead the ‘ultimate’ sources of human agency in the very places people typically think they are? [38]


1. This essay first appeared in Essays on Free Will and Moral Responsibility, edited by Nick Trakakais and Daniel Cohen (Cambridge Schoalrs Series, 2008), pp.186-208.  Reprinted with kind permission of Cambridge Scholars Series. 

2. Unless otherwise indicated, I restrict discussion to directly free action and direct moral responsibility.  Derivatively free (as opposed to directly free) action derives from its casual dependence on some other action freely performed by the same agent. The same applies to moral responsibility. 

3. So as not to mislead, if some Y is entirely the product of conditions X, and X can in principle be characterized in terms of our most basic sciences, it does not follow that Y can be predicted on the basis of conditions X.

4. But not so for all incompatibilists. Notably, event-causal libertarians like Robert Kane hold that free will and moral responsibility are compatible with naturalism, but not with determinism. See Robert Kane, The Significance of Free Will (Oxford: Oxford University Press, 1996). So long as human action is not the product of a deterministic history, it is possible to satisfy the needed conditions. Causal indeterminacy located in the natural order will do the trick. 

5. Without due caution, I am liable to characterize incompatibilism unfairly. To continue use of the metaphor, it is open to incompatibilists to allow that, from dirt, such complex phenomena as consciousness, agency, deliberative reasoning, and so on might arise. What they deny is only that morally responsible agents arise endowed with the sorts of agential capacities that are required to underwrite such sophisticated kinds of agency. I am indebted to Derk Pereboom for encouraging me to clarify this point.

6. Galen Strawson, “The Impossibility of Moral Responsibility,” Philosophical Studies 75 (1994): 5.

7. Suppose one thought of oneself as a human animal, as I do. To take Strawson’s argument seriously, a slight modification is required. Moral responsibility surely does not require that one, so to speak, literally parent oneself (we can’t all be expected to be like characters in Heinlein stories). Rather, what Strawson can be taken to mean is that one is the cause of one’s moral self or personality. Spelling this out in the manner that Strawson is inclined to do would get something close to the notion of a causa sui, and would be enough for Strawson to say that this is impossible. I set these qualifications aside for ease of presentation. 

8. For an illuminating discussion of the free will problem that distinguishes between incompatibilism and impossibilism, see Kadri Vivhelin, “Compatibilism, Incompatibilism, and Impossibilism,” in John Hawthorne, Theodore Sider, and Dean Zimmerman (eds), Contemporary Debates in Metaphysics (Malden, MA: Blackwell, 2008), pp.303-18. While I disagree with Vivhelin’s claim that an incompatibilist cannot be an impossibilist, I do agree with her that at least some arguments for impossibilism are not arguments for incompatibilism. 

9. This criticism is also developed by Randolph Clarke, “On an Argument for the Impossibility of Moral Responsibility,” Midwest Studies in Philosophy 29 (2005): 13-24, and Alfred Mele, Autonomous Agents (New York: Oxford University Press, 1995), p.222.

10. Mele makes a similar point in Autonomous Agents, p.225.

11. Galen Strawson, Freedom and Belief (Oxford: Clarendon Press, 1986), p.30.

12. John Martin Fischer develops this point and he credits it to Joel Feinberg. See Fischer, “The Cards That You Are Dealt,” Journal of Ethics 10 (2006): 110, and Joel Feinberg, Harm to Self (Oxford: Oxford University Press, 1986), pp.33-34.

13. Robert Kane also develops a similar argument in The Significance of Free Will, pp.73-78.

14. Saul Smilansky, Free Will and Illusion (Oxford: Clarendon Press, 2000), p.284.

15. Ibid., p.45.

16. This condition is close enough to the condition Smilansky calls the Principle of Sole Attribution (PSA):

Any feature F due to which a person deserves something S in the libertarian free will-dependent sense must, in the normatively relevant respects, be solely attributable to the person or to the pertinent aspect A of the person.

Kane also thinks about ultimacy in terms similar to U. See Kane, The Significance of Free Will, pp.73-74.
In correspondence, Derk Pereboom has helpfully suggested a way to develop U further, giving more substance to this necessary condition. He proposes:

U2: An agent, A, is the ultimate source of her action D only if she voluntarily contributes some necessary condition, C, to D, such that she understands that C so contributes to D, and such that there are no sufficient conditions for C that obtain independently of A.

As a partial definition of ultimacy, U2 is certainly a more informative necessary condition, and perhaps would be welcomed by Smilansky and Kane. Nevertheless, I will work with U instead, as my argument would not be influenced by this amendment and U avoids worries that might arise with U2. In particular, if ‘voluntarily contributes’ requires free action, then we have the seeds of a regress of the sort that Strawson was concerned to highlight. 

17. Here I distinguish between Smilansky’s argument against the compatibilist and his overall position. His overall position includes the claim that libertarian free will (as he calls it) is incoherent for reasons the impossibilist Strawson enlists, reasons that have nothing to do with determinism (see Smilansky, Free Will and Illusion, p.65). Regardless, Smilansky does defend an argument that relies upon the notion of ultimacy and that moves from the assumption that determinism is true to the conclusion that, if it is, no one has (true) freedom or moral responsibility. See Smilanksy, Free Will and Illusion, p.45. 
    For an example of one who endorses an incompatibilist version of the Ultimacy Argument but who holds that in the absence of determinism there is a way to satisfy ultimacy, see Kane, The Significance of Free Will. Kane holds that ultimacy can be achieved using only the resources available to the naturalist (as I have defined naturalism).
    In Living Without Free Will (Cambridge: Cambridge University Press, 2001), Derk Pereboom defends a position in which neither determinism nor an exclusively event causal indeterminism (and thus naturalism) are compatible with ultimacy. Yet Pereboom contends that ultimacy is necessary for moral responsibility. Pereboom, however, thinks that free will and moral responsibility are at least conceptually coherent in that agent causation is coherent and would be adequate for persons to be the ultimate sources of their actions. Nevertheless, as I shall explain below, unlike either Smilansky or Kane, Pereboom does not advance an argument for incompatibilism by appeal to a premise that demands ultimacy. Rather, he sees the condition of ultimacy as something that can be concluded on the basis of an independent argument for incompatibilism, a manipulation argument.

18. See Smilanksy, Free Will and Illusion, p.285, and “Compatibilism: The Argument from Shallowness,” Philosophical Studies 115 (2003): 257-82.

19. Thomas Nagel, The View from Nowhere (New York: Oxford University Press, 1986).

20. Smilanksy, “Compatibilism: The Argument from Shallowness,” p.275.

21. John Martin Fischer makes a similar point in response to Smilansky in “The Cards That You Are Dealt,” p.123.

22. In correspondence, Saul Smilansky has thoughtfully raised this concern. 

23. See Robert Kane, The Significance of Free Will.

24. In Autonomous Agents, Mele has suggested that a certain sort of libertarian, a modest libertarian, might claim that one could have the following reason for accepting something like ultimacy: She prefers to be an indeterministic initiator of her actions. This would be rational in the minimal sense that it would not be irrational for her to prefer this. But this modest position would not suffice to show that, absent being an indeterministic initiator, an agent is not free and morally responsible (as U requires). So, on top of being modest, this sort of libertarian would also be soft in that she would not deny the truth of compatibilism.  Obviously, what philosophers like Strawson, Smilansky, and Kane are working toward is a stronger sort of reason for accepting ultimacy, one that secures a much stronger position than the modest soft libertarianism that Mele sketches as a live theoretical option. 

25. Consider, for example, ‘deep self’ views such as Harry Frankfurt’s in “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68 (1971): 5-20.
26. Smilansky, Free Will and Illusion, [emphases are Smilansky’s pp.47-48.

27. It seems to me that in most discussions of the free will problem, suffering is too closely associated with moral responsibility, blameworthiness and blaming. Suffering, if ever ‘justly’ inflicted, should be understood in terms of justified punishment. A distinct question is whether determinism is compatible with a form of punishment that requires justified infliction of suffering. One might be a compatibilist about moral responsibility (blameworthiness, blaming) and determinism, and an incompatibilist about punishment (construed as such) and determinism.

28. This formulation of O is proposed by Pereboom in correspondence, and is a development of a leaner version found in Pereboom, Living Without Free Will, p.4. The difference is that O as formulated here makes clear why Pereboom is an incompatibilist about naturalism and moral responsibility as well as determinism and moral responsibility.

29. The two best thus far are Pereboom’s and Mele’s. See Pereboom, Living without Free Will, and Alfred Mele, Free Will and Luck (New York: Oxford University Press, 2006). 

30. Michael McKenna, “A Hard-Line Reply to Pereboom’s Four-Case Argument,” Philosophy and Phenomenological Research (2008, forthcoming).

31. Thanks to Randy Clarke for this example.

32. In assessing the Ultimacy Argument, Vihvelin makes a similar point in “Compatibilism, Incompatibilism, and Impossibilism.” For compatibilist proposals that can account for the self-made agent, see John Martin Fischer and Mark Ravizza, Responsibility and Control (Cambridge: Cambridge University Press, 1998), Ishtiyaque Haji, Moral Appraisability (New York: Oxford University Press, 1998), and Mele’s Autonomous Agents and Free Will and Luck.

33. John Martin Fischer, “Responsibility and Self Expression,” Journal of Ethics 3 (1999): 277-97, and “The Cards That You Are Dealt.”

34. Gary Watson, “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme,” in Ferdinand Schoeman (ed.), Responsibility, Character, and the Emotions: New Essays in Moral Psychology (Cambridge: Cambridge University Press, 1987), pp.256-86.

35. P.F. Strawson, “Freedom and Resentment,” Proceedings of the British Academy 48 (1962): 187-211.

36. At this point, I part company with compatibilists who wish to argue that our concept of moral responsibility does not include an excessive demand like Strawson’s. It seems to me that in resisting Strawson on this point, some compatibilists share Strawson’s assumption that there is a unified, single concept of moral responsibility that is our ordinary one. Then the debate is over who is correct in describing it. In my estimation, the concept of moral responsibility is better understood as a more amorphous one that contains various elements, some of which might not cohere well (or at all) with others. 
    It is worth adding that if there were some single, stable concept of moral responsibility that is ours, then one would think that it would be a fairly straightforward matter to test for it by checking on how folk intuition responds to various cases. As it turns out, the tests suggest that the folk are not settled on the matter. But the data suggest a compatibilist rather than an incompatibilist (and certainly not an impossibilist) construal of the ‘ordinary’ concept. On this point, see Eddy Nahmias, Stephen Morris, Thomas Nadelhoffer, and Jason Turner, “Is Incompatibilism Intuitive?” Philosophy and Phenomenological Research 73 (2006): 28-53.

37. For a philosopher who adopts a similar treatment of Strawson’s Ultimacy Argument, see Manuel Vargas’ thoughtful paper, “Libertarianism and Skepticism about Free Will: Some Arguments against Both,” Philosophical Topics 32 (2004): 403-26.

38. I am grateful to Daniel Cohen and Nick Trakakis for inviting me to contribute to this volume. For helpful comments I would like to thank Joseph Baltimore, Daniel Cohen, Andrew Cullison, John Martin Fischer, Ishtiyaque Haji, David McNaughton, Derk Pereboom, Sharon Ryan, Saul Smilansky, Manuel Vargas, Kadri Vihvelin, and Chris Zarpentine. I have also profited by discussing these issues with Randolph Clarke and Alfred Mele. This paper is dedicated to the memory of my childhood friend, Tommy Hawes, number 52, who knew how to laugh.

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