Early in Freedom, Teleology, and Evil, Steward Goetz expresses his conviction that those who favour a compatibilist understanding of freedom do so not ‘because of its own merits’, but ‘only because they become convinced that libertarianism is too problematic and must, in the end, be abandoned for some other view’ (3). His reason for this judgment appears to be that a libertarian understanding of freedom comes closest to capturing our immediate experience of ourselves.
Seen in this light, Freedom, Teleology, and Evil can be usefully viewed as a continuation of the project Goetz pursues in the almost concurrently published Naturalism (Grand Rapids MI: Eerdmans, 2008, co-authored with Charles Taliaferro). In both works, Goetz can be seen as aiming to defend the plausibility and coherence of treating our immediate self-understanding as veridical – as something that can be explained rather than something we are forced to explain away.
In the present book his aim is to offer an original account of libertarian freedom that fits with our understanding of ourselves as agents who make free choices based on reasons. He seeks to show that his theory of ‘non-causal agency’ not only meets the main challenges to libertarian theories of free agency but has advantages over more common libertarian theories (especially the ‘agent-causation’ theory). In the course of doing so, Goetz offers a rigorous discussion that clearly explicates and critically engages with an array of alternative perspectives and opposing arguments. The result is a book that both familiarizes readers with current philosophical debates and provides a defence of an original theory with which they can profitably engage – making Freedom, Teleology, and Evil a good choice for graduate seminars as well as scholars in the field.
The early chapters explicate Goetz's theory in the light of alternatives and challenges. He begins by offering a definition of a choice, which he takes to be an ‘essentially uncaused and intrinsically active exercising’ of a ‘mental power’, one which he simply calls ‘the power to choose’ (9). Goetz distinguishes powers from capacities. A capacity is the ability to be affected in a certain way, and as such is always activated by a cause which brings about the effect. But a power is an ability to initiate an event. As such, the exercise of a power must be essentially uncaused.
Goetz stresses that to be without a cause is not to be without an explanation. And he thinks choices do need explanations. But since they cannot be causal they must be teleological. And so a choice is ‘an essentially uncaused event whose occurrence is … explained teleologically by the reason for which it is made’ (36). This need for an explanation, combined with the fact that the explanation must be teleological, leads Goetz to endorse what he calls the ‘reason–choice principle’ (RC): ‘An agent is free to choose (make a particular choice) at time t only if he has a reason R at that time for so choosing (making that choice)’ (22).
But Goetz's theory of choice requires something else as well – specifically, some possible act other than what the agent in fact did but which the agent might have done, along with a reason why the agent might have done so. Goetz wants to incorporate this as an essential feature of a choice in large measure because he accepts the Principle of Alternative Possibilities (PAP), according to which moral responsibility for our choices requires that we could have done otherwise. And Goetz seeks a theory that supports our intuition that we are morally responsible for our choices.
Hence, Goetz supplements RC with the ‘plural reasons principle of choice’ (PRC): ‘An agent is free to make a choice for a reason R1 at a time t only if he has a reason R2 to choose otherwise at t, where choosing otherwise is causally open to him.’ I should add that, although PRC does not explicitly assert this, Goetz intends for it to hold that choosing otherwise is not excluded by any sort of teleological determinism; the point is that there is a possible world identical in all other relevant respects in which the agent does otherwise. When PRC does not obtain, an agent may perform the mental act of forming an intention to do A1 for reason R1 but not the mental act of choosing. Choice, Goetz thinks, essentially involves the power to have chosen otherwise, and it is only when PRC obtains that the agent has this power.
In short, Goetz is not satisfied just to offer a theory of non-causal agency in which our actions are explained teleologically rather than causally. He wants a libertarian theory. And as Goetz notes, a determinist needn't be a causal determinist. A determinist might agree that to be an agent is to perform uncaused mental acts of choosing on the basis of reasons – and yet hold that the totality of reasons available to the agent always determine what choice the agent will make.
The first few chapter of the book are devoted to explicating this theory in the light of alternatives and in the face of objections, and Goetz is particularly concerned with objections coming from fellow libertarians who favour agent causation. In responding to these objections he repeatedly invokes a tu-quoque style of argument, the general form of which might be summarized as follows: agent causationists, who share Goetz's rejection of causal determinism, offer a reason to think that the mental act of choosing cannot itself be uncaused and so must be caused by a prior mental act of the agent (e.g. an act of will) that is uncaused. But any reasons for holding that the mental act of choosing itself could not be uncaused would also apply to this preceding mental act. To prevent an infinite regress, one must accept an uncaused mental act whose explanation ultimately lies in the agent's reasons or purposes. But if so, agent causation theories introduce explanatorily superfluous elements, since the mental act of choosing could itself be this uncaused mental act.
In chapter 4 Goetz focuses on ‘the luck objection’ to libertarian theories, according to which libertarianism implies that what we actually choose is a matter of luck. He distinguishes three versions, the first of which treats the problem as ‘intrinsic to an agent's choice’, the other two treating the problem as one that choices inherit by virtue of their relation to something else (specifically, the agent's ‘initial character, personality, or motivational structure’). There is, again, much here that is of philosophical interest. However, for reasons implicit in my critical remarks below (but which for the sake of space I will not develop), I find his response to the first version of the objection unconvincing.
In chapter 5, Goetz turns to a defence of PAP against Frankfurt-style counter-examples – a defence which he pursues in large measure because his allegiance to PAP is what justifies the inclusion of PRC in his account of non-causal agency. While interesting and philosophically important in its own right, this chapter functions more to justify the need for an account of freedom like Goetz's than to further explicate or defend Goetz's theory.
In chapter 6 Goetz brings his theory to bear on the problem of evil. In the course of developing his ideas, Goetz sides with Walls over Plantinga concerning whether a response to the problem of evil that invokes libertarian freedom must take the form of a theodicy and not simply a defence. Goetz thinks that it must – and he proceeds to sketch out his own theodicy, one which has clear implications for the doctrine of limited salvation, that is, the doctrine that not all created persons enjoy eternal blessedness. As such, Goetz finds himself taking issue with Marilyn McCord Adams, whose approach to theodicy calls for God to be acting to guarantee the salvation of all.
There is much of interest in this final chapter (and much with which I disagree). However, it is unclear to what extent Goetz's theodicy depends on the specifics of his theory of non-causal agency. It seems, rather, that it relies on the more general presumption that libertarian freedom exists. That said, it's important to note that insofar as his theodicy is premised on a libertarian freedom in which agents can be held morally responsible for their choices, an original theory of libertarian freedom which eschews the problems typically ascribed to such theories would add something of great significance.
The most important question, then, is whether Goetz's theory does this. In fact, I think that it does not. Specifically, I think there is a fundamental tension between RC and PRC which Goetz has failed to resolve. Goetz's basic error, I think, lies in confusing two things. First, there is what we might call the reason for an act (Ra). This would be the specific reason (say Ra1) one has for doing/forming the intention to do some particular act A1. Second, there is what we might call the reason for a choice (Rc): the reason one has for deciding to do/form the intention to do A1 rather than A2, given that one has reason Ra1 to do A1 and Ra2 to do A2.
Understood in this light, an Rc might be a higher-order reason such as ‘Ra2 is a prudential reason for action while Ra1 is a moral reason, and moral reasons are more worthy of acting on.’ In such a case, the Rc is extrinsic to the reasons for action (call it an ERc). Alternatively, the Rc might be intrinsic to the reasons for action if, say, Ra1 is more immediately compelling to the agent than Ra2. We might say under these conditions that Ra1 is not just a reason for doing A1, but a reason for choosing it over A2. A reason for action that itself provides a reason for choice I'll call an intrinsic reason for choice (IRc).
In cases involving IRcs, the teleological explanation for a particular action (one's reason for doing it) also explains why one chose that action over an alternative which one also had reason to choose. But if the reason (Ra2, say) for what one does (A2) is equally or less intrinsically compelling than the reason (Ra1) for an alternative (A1), then while Ra2 can serve as a teleological explanation for the doing of A2, it cannot serve as the teleological explanation for why one chose A2 over A1. In such cases, something other than the reason for the action is needed to explain the choice.
Goetz does recognize that higher-order explanations may sometimes be necessary. He notes (while engaging Nagel) that these higher-order explanations – what I am calling ERcs – might be paired against contrasting ones. In such cases there is a meta-level choice that will have to be explained either by the intrinsically more compelling character of one ERc or by an even higher-order ERc. But eventually we must reaching a stopping place – an ERc that either is not paired with an opposing ERc or is intrinsically more compelling than those it is paired against. As such, Goetz admits that there may be an ‘all-things-considered most reasonable action’ (29). What Goetz claims, however, is that libertarian free agents ‘can make irrational or akratic choices against all-things-considered beliefs’ (29).
I do not deny that this can happen – that is, I am not begging the question against libertarianism by assuming that choices are always determined by the weight of an agent's reasons. What I deny is that in such cases there can be a teleological explanation for this choice (the selection of A2 over A1) – even if, as may be the case, the action itself is done for a reason and so admits of a teleological explanation.
In other words, I'll grant that a free choice does require something like PRC, so that what one chooses is rendered indeterminate by the reasons one has for acting. Let me first consider the implications of this for cases in which the reason for choice is an IRc. Suppose Ra1 is more compelling than Ra2. I happily concede that it could still serve as the Rc for one's choice of A1 over A2 even if it does not determine that choice. But if there is an indeterminacy here, what that means is that the agent might still choose A2. What is hard to imagine is that, under these conditions, Ra2 could be the reason for that choice. While Ra2 is a reason to do A2, how can it be the reason one chooses A2 over A1 if one's reason for doing A1 (Ra1) is more compelling? It seems, rather, that the agent's choice of A2 needs in that case to be chalked up to an inexplicable arbitrariness at work in human choices (unless one assumes that one is causally determined to select A2 by some controlling affective state, in which case no libertarian freedom exists). The same problem exists if one's reasons for alternative courses of action are equally compelling.
It should be clear that moving the problem up to higher levels, by looking at ERcs, does not solve this problem. What Goetz calls an akratic or irrational choice is precisely a choice that is unexplained by any reason for choice. Thus, even if the action that is ultimately performed has a teleological explanation in terms of a reason for doing it, the mental act of choosing this over the alternatives admits of no such teleological explanation. Hence, either an akratic choice is inexpicably random (opening the door for a version of the luck objection), or it is causally determined (undermining Goetz's allegiance to noncausal agency).
I suspect that confusion over equivocal uses of ‘choice’ may explain why someone would say that a reason for an action (say Ra2) is the reason for a choice, even when it is neither intrinsically more compelling than other reasons for action nor endorsed by one's all-things-considered higher order reasons for choice. Sometimes, when we talk about the choice of A2, we mean the mental act of forming the intention to do A2 in a context in which there exist reasons for doing something else. If this is what we have in mind, then clearly Ra2 can serve as the reason for one's ‘choice’ – simply because what one means is that Ra2 is the reason one had for forming the intention to do A2. What it cannot do, however, is explain why one settled on A2 rather than A1. And so it cannot explain one's ‘choice’ in this more robust sense.
The upshot of all of this is that embracing PRC, as Goetz does, entails that many human choices (in the robust sense) will be rendered teleologically inexplicable. And insofar as Goetz's theory excludes causal explanations, many human choices are thereby rendered inexplicable simpliciter. But insofar as Goetz affirms RC because he thinks that mental actions need to be explained even if the explanations are not causal, there emerges a serious tension within Goetz's theory of non-causal agency.
