This collection of essays has its roots in a conference on free will and moral responsibility held at Monash University in November 2005, though only a few of the papers presented at the conference have made it into the current volume. We would like to thank both the participants at this conference and the contributors to this volume, as well as Cambridge Scholars Publishing for inviting us to put the collection together. Grateful acknowledgement is also made to the editor of the online journal, Sorites, where Nick Trakakis’ paper, “Whither Morality in a Hard Determinist World?
”, was originally published (in vol. 19, December 2007). Finally, thanks to Shannon Weekes for her assistance in compiling the Index. INTRODUCTION NICK TRAKAKIS AND DANIEL COHEN Much of the interest of the free will debate depends on the assumption that free will is necessary for moral responsibility. In particular, it is because responsibility seems so important for our practical lives that debates about the compatibility of free will and determinism seem so urgent.
However, much of the discussion in this volume bypasses this link. Instead, questions are raised that directly concern responsibility, such as whether it is compatible with determinism (see, for example, the essays by Fischer, Widerker, and Pereboom) and whether it is compatible with indeterminism (for example, the exchange between Levy and Kane). For the purposes of this introduction, we have not attempted to summarize the various ways in which the contributors construe the metaphysical foundations of moral responsibility.
Instead, we wish to address a more preliminary matter. In the first part of this introduction, our aim is to say something about what we mean when we say that someone is morally responsible. It is surely important to clarify this before addressing any further substantive issues because, if we don’t clarify the meaning of this key term, there remains a significant danger that different participants in the debate about the possibility of moral responsibility will simply ‘talk past each other’.
This suggests that in order to conduct a fruitful debate participants need firstly to agree on the nature of their subject-matter and, perhaps, to disambiguate different dimensions of the debate that arise if the term ‘moral responsibility’ has different connotations. In the second part of the introduction, we will discuss a neglected Wittgensteinian perspective on the notions of freedom and responsibility, a perspective that may help to clarify some of the confusion that arises when we ask what it means to say that a person is free or responsible. 1. The Meaning of Responsibility.
Before proceeding to ask whether people are, in fact, ever morally responsible, it seems that an important preliminary matter needs to be settled. That is, we need to ask what we mean when we say that a person is x Introduction morally responsible. As will quickly become clear, settling this preliminary matter is, in fact, much harder than it first seems. Many of the controversies concerning the possibility of responsibility emerge even when we try to say just what ‘responsibility’ means. Let’s start with a claim that seems relatively uncontroversial.
We will simply assume that ‘person A is responsible’ is a normative claim. That is to say, it is a claim to the effect that it is appropriate to hold A responsible in certain circumstances (circumstances, for example, where A has acted wrongly and where no mitigating, justifying, or excusing factors are present). However, this account of the meaning of ‘A is responsible’ raises at least two further issues. Firstly, we need to say more about what it means to hold someone responsible and, secondly, we need to say more about the nature and basis of the norms that govern appropriate responsibility attribution.
It is important to keep in mind that our goal, in clarifying these two questions, should not be to settle any question regarding whether people are, in fact, ever morally responsible. Rather, we want to settle the terms of this debate before it begins. To this extent, we need sufficiently neutral accounts both of the nature of responsibility attribution and of its normative basis so that we don’t beg any substantive questions before debate has even begun. As we will try to show, however, this is a rather elusive goal. The only neutral account of the nature of responsibility attribution renders the normative question deeply controversial.
And the only neutral account of the normative basis of responsibility attribution renders the nature of responsibility attribution deeply controversial. Holding Responsible There appear to be two plausible contending views regarding the nature of responsibility attribution. On cognitivist accounts, holding A responsible fundamentally involves believing something to be true of A, while on noncognitivist accounts, holding A responsible essentially involves holding some conative attitude towards A. (Cognitivists may, of course, argue that responsibility attribution is also usually associated with some conative attitude.
However, they will maintain that it is possible to hold someone responsible without holding such conative attitudes. Similarly for noncognitivist accounts, mutatis mutandis. ) Non-cognitivism appears to provide the most successful neutral basis on which to premise the debate concerning the possibility of responsibility. This is because there seems little room for debate concerning the conative attitudes that characterize normal responsibility Nick Trakakis and Daniel Cohen xi attribution. In particular, few would disagree that responsibility attribution is strongly associated with the ‘reactive attitudes’ identified by P.
F. Strawson, i. e. , resentment, indignation, anger and so on. 1 If one wishes to argue, however, that the reactive attitudes, while prevalent, are inessential to responsibility attribution, it is much harder to locate any common ground concerning the beliefs that are essential to responsibility attribution. One may suggest, for instance, that to hold A responsible is to believe that she is the source of some bad behaviour. Deep controversies quickly emerge on this view, however. One might take sourcehood to involve a psychological claim, for instance that A ‘really wanted’ to act wrongly.
2 However, others might object that any such glib psychological account fails to explain why it is fair to blame A for the wrongdoing (see Smilansky’s contribution). One might object, in this vein, that any such psychological story is unable to show that an agent really is the source of her having certain desires or values (see McKenna’s contribution), and that sourcehood thus requires some more obscure metaphysical basis (e. g. , agent-causation). Alternatively, one may suggest that sourcehood involves some impossible requirement such as that an agent was self-created.
3 On this view, holding someone responsible is essentially impossible. 4 Our goal is to account for the meaning of responsibility in neutral terms so as to provide a basis for constructive debate about the conditions (and the very possibility) of responsibility. It appears, however, that the cognitivist view of responsibility attribution quickly leads to debates that already beg these important questions before debate has even begun! This suggests that the best theory-neutral account of the meaning of responsibility must explain holding responsible in non-cognitivist terms.
The Normative Basis of Responsibility Attribution Recall that, for the purposes of this discussion, we have assumed a normative account of responsibility according to which ‘A is responsible’ means ‘it is appropriate to hold A responsible in certain conditions’. Having addressed how best to interpret what ‘holding A responsible’ might mean, without begging any important questions, we need now to turn to a second question raised by the normative account: when exactly is it appropriate to hold someone responsible? In other words, what are the norms that govern appropriate responsibility attribution?
Again, there are two plausible contending views: appropriateness may be explained either in terms of practical norms (taking ‘holding responsible’ to be analogous xii Introduction to the performance of an action) or by way of doxastic norms (taking ‘holding responsible’ to be analogous to the formation of a belief). Again, only on one of these accounts—the doxastic view—is it possible to offer an appropriately uncontroversial explanation of the norms implicit in responsibility attribution. On the doxastic view, one ought to hold A responsible if and only if it is true that A is responsible.
On this view, the normative basis of responsibility attribution straightforwardly derives from the normativity of belief. It is clear that the doxastic account presupposes the cognitivist view discussed earlier, according to which holding A responsible involves believing something about her. Given this view of the nature of responsibility attribution, the normative question— concerning when responsibility attribution is appropriate—has a straightforward answer. Unfortunately, as we saw, there is no uncontroversial way to account for the truth-conditions of ‘A is responsible’, on the cognitivist assumption that it involves belief.
So, despite the advantages of the doxastic view in providing a neutral account of the normative basis of responsibility attribution, this view at the same time precludes us from obtaining a neutral view regarding the nature of responsibility attribution (i. e. , the truth-conditions for the belief that A is responsible. ) Might we find an account of the normative basis of responsibility attribution that is consistent with the preferable non-cognitivist view outlined earlier? This would have to involve an alternative view, according to which responsibility attribution is justified in virtue of practical norms.
However, if responsibility attribution is governed by practical norms, then things are much less straightforward. One may suggest that the relevant practical norms are just moral norms, so that ‘A is responsible’ states something like: ‘It is morally obligatory (or, perhaps, permissible) to hold A responsible’. This view may appear immediately problematic because the appropriateness of responsibility attribution will now depend on further questions that are deeply controversial (for instance, questions concerning the debate between consequentialism and nonconsequentialism; see Vargas’ contribution).
A more fundamental worry arises concerning the methodological appropriateness of appealing to moral norms. One may argue, for instance, that the nature of moral obligation, itself, depends on the foundations of responsibility, which is, of course, the question at issue. Haji (in his contribution) argues that the best metaphysical basis of responsibility (i. e. , event-causal libertarianism) renders moral obligation essentially lucky. This suggests that there would be something viciously circular in explaining the meaning of responsibility Nick Trakakis and Daniel Cohen xiii in terms of some claim about our moral obligations. (See also Trakakis’ contribution. )
To avoid these worries, one may appeal to practical norms that appear to be more fundamental than any particular moral system. For instance, R. J. Wallace offers a normative account of responsibility that appeals to fairness. 5 This is meant to provide a normative basis for responsibility that remains neutral on more substantive moral issues. (See also Smilansky’s contribution. ) Clearly, however, appealing to practical norms launches us into further debates that already beg the question at issue.
Again, such an account seems ill-suited for the purpose of setting up a neutral definitional framework on which to premise further debate. A Dilemma Our aim has been to find some neutral definition of responsibility to enable further non-question-begging debates about the possibility and conditions of responsibility. It seems that this goal gives rise to a tricky dilemma. The best theory-neutral account of holding responsible is the non-cognitivist account. However, this account appears incompatible with the best theory-neutral account of the norms that govern responsibility attribution—the doxastic account.
The doxastic account, in turn, seems compatible only with the most problematic account of holding responsible—the cognitivist account. This is a puzzling result. Even though responsibility clearly gives rise to very complex issues, it is surprising that it is not possible even to define the terms of the debate without deep controversy. The worry thus remains that debates about the possibility and conditions of responsibility are essentially question-begging, insofar as different participants to the debate conceive of its key terms differently.
Must we conclude, then, that different people and different theorists are indeed talking past each other when they debate about the possibility of responsible action? This, of course, would be a depressingly deflationist conclusion. There is a possible way out, however, that is rarely canvassed. If the question concerns the meaning of ‘responsibility’, one might suggest that there are, perhaps, other ways to settle things. In particular, isn’t the meaning of a term determined by our use of the term? (Or, at least, isn’t use a good guide to meaning?)
Thus, appeal to real-world attributions of responsibility may illuminate the meaning of the term ‘responsible’. Such a methodology is sometimes rejected on the grounds that real-world attributions are rife with internal inconsistency (see, for example, Cohen xiv Introduction and Saling’s contribution). But perhaps a closer look at the Wittgensteinian ‘solution’ to the problem of freedom and responsibility will throw new light on the matter. 2. Making Sense of Free Will: A Wittgensteinian Account Wittgenstein published very little during his lifetime, and even less on the topic of free will.
He does, however, make some pertinent remarks in his Tractatus Logico-Philosophicus that mirror in some respects the contemporary debates on free will. As is well known, the early Wittgenstein claims to have definitively solved the central problems of philosophy, and by implication this would include the perennial problem of free will. Wittgenstein’s strategy proceeds by separating sense from nonsense. The realm of sense is delimited in the light of his picture theory of meaning, according to which a proposition is meaningful (or capable of expressing a fact) only if it can represent or picture a contingent state of affairs.
What cannot thus be represented is consigned to silence, or as Wittgenstein famously put it at the conclusion of his book, “What we cannot speak about we must pass over in silence” (proposition 7). 6 In the course of the book, however, it soon becomes clear that what can be meaningfully said are only the propositions of natural science, thereby leaving out of the realm of sense a daunting number of statements which are regularly made and used in language, including the propositions of logic, aesthetics, religion, and (most relevant for our purposes here) traditional metaphysics and ethics.
But unlike the positivists, Wittgenstein does not assume that what is nonsensical is of no value. As he stated in a letter to a prospective publisher, Ludwig von Ficker: “My work consists of two parts: the one presented here plus all that I have not written. And it is precisely this second part that is the important one. ”7 Further, ethical and metaphysical truths that cannot be ‘said’ or formulated in sayable (sensical) propositions can nonetheless be ‘shown’: “There are, indeed, things that cannot be put into words. They make themselves manifest. They are what is mystical” (6. 522).
In line with this austere outlook, Wittgenstein hints in the Tractatus that the very concepts at issue in the free will debate—concepts such as ‘freedom’ and ‘responsibility’—cannot be meaningfully expressed. Although in propositions 5. 1361 (“We cannot infer the events of the future from those of the present”) and 5. 1362 (“The freedom of the will consists in the impossibility of knowing actions that still lie in the future”) he seems to reject determinism as false and to accept freedom of the will as true, rather than rejecting both as nonsensical, here we arguably have a Nick Trakakis and Daniel Cohen xv minimalist conception of free will as nothing more than ignorance or uncertainty regarding the future, as opposed to a substantive commitment to the metaphysical idea of a ‘will’ that could be free or unfree.
This is confirmed in later passages where Wittgenstein takes the law of causality—the principle that every event has a cause—to be “not a law but the form of a law” (6. 32), adding a few propositions later that, “If there were a law of causality, it might be put in the following way: There are laws of nature. But of course that cannot be said: it makes itself manifest” (6. 36).
The law of causality, in other words, is not itself a law of logic nor a law of nature (or an empirical generalization), nor a synthetic a priori proposition, but rather “something purely logical” (6. 3211), a vacuous principle that tells us, not something about the world, but only something about the form our thinking about the world must take. But what is formal, according to the Tractatus, can only be shown, not said. On this view, then, the law of causality, and by extension any substantial or metaphysical doctrine of determinism, cannot be affirmed or denied, but must be placed in the category of ineffability or nonsense.
Similarly, the denial of determinism—viz. , indeterminism—is bound to result in nonsense. At one stroke, then, Wittgenstein seems to have dissolved the free will problem. Contemporary discussions of free will often take a similar turn. For example, concepts such as ‘free will’ and ‘moral responsibility’ are routinely rejected as internally incoherent or contradictory, or as incompatible with determinism or indeterminism (or both), and like the early Wittgenstein this result is achieved simply through a kind of armchair or a priori reflection on the conditions of possibility of free will and responsibility.
A glaring instance of this is Galen Strawson’s ‘pessimist’ conclusion that free will, of the sort that is necessary for genuine moral responsibility, is impossible, for in order to have that kind of free will (according to Strawson) one would per impossible have to be the ultimate cause or origin of oneself, a sort of causa sui. But what is neglected in this procedure is attention to particulars, to the variety of ways in which concepts such as free will and responsibility function in different discourses and social practices.
This, of course, is the message of the ‘later Wittgenstein’, the Wittgenstein of the Philosophical Investigations8, which effects a fundamental change of perspective: from the realm of an idealized logical language with rigorous definitions and analyses to the vagaries of everyday life and action out of which arise the multifarious ‘language games’. The earlier reduction of language to representation is now seen as incapable of doing justice to the rich fabric of human language, and so xvi Introduction
Wittgenstein famously moves from a conception of meaning as representation to a view of meaning as use: language as a kind of doing rather than a kind of picturing. We are therefore exhorted to “look and see” (PI 66) whether there is anything in common in the variety of uses to which a word is put. We cannot simply assume that words like ‘freedom’ and ‘responsibility’ must have a hidden essence, or a universally applicable meaning that can be formulated in terms of necessary and sufficient conditions.
Rather, we need to look to the complicated network of overlapping and criss-crossing similarities—what Wittgenstein calls ‘family resemblances’—between various words as these find expression in ordinary life and in various language games. Each language game, however, has its own unique ‘grammar’ (or network of rules which determine what linguistic or conceptual moves are allowed as making sense) and its own criteria of truth, rationality and intelligibility which may or may not be shared by other language games.
What counts as freedom of the will may therefore differ widely depending on which language game is being played: freedom, for some religious believers, requires the extinction of one’s will, while for purposes of legislation freedom may be conceived of as requiring a significant degree of self-determination. Underlying this view is the rejection of the mathematical ideal of the Tractatus, typified by statements that are put forward as impersonal, unambiguous and impervious to context, and whose truth is intended to be timeless and without qualification.
Language, for the later Wittgenstein, is a much more dynamic, diversified and activity-oriented phenomenon. And to be faithful to the richness and complexity of this phenomenon demands an appreciation of the bewildering and sometimes conflicting ways in which words and concepts—‘free will’ and ‘responsibility’ included—are used. It may be instructive to briefly compare (the later) Wittgenstein’s treatment of religious belief with his remarks on free will and voluntary action.
Consider, for example, how Wittgenstein, in his “Remarks on Frazer’s Golden Bough”, responds to the phenomenon of religious diversity, to the fact that different religions seem to say different and incompatible things about (e. g. ) the nature of divine or ultimate reality and the nature and destiny of the human race: Was Augustine in error, then, when he called upon God on every page of the Confessions? But—one might say—if he was not in error, surely the Buddhist holy man was—or anyone else—whose religion gives expression to completely different views.
But none of them was in error, except when he set forth a theory. 9 Nick Trakakis and Daniel Cohen xvii Against philosophers and anthropologists like James George Frazer, who construe religious doctrines as hypotheses or theories that can be confirmed or disconfirmed in light of empirical evidence, Wittgenstein views each religion as embodying a unique form of life that finds expression in language games whose ‘rules’ (relating to truth, rationality, intelligibility, and so on) may diverge quite dramatically from those of science.
On this view, the various religions of the world are not in the business of constructing hypotheses and searching for evidence, and so they are not in competition with one another, at least in the way that scientific theories may vie for the mantle of verisimilitude. The problem of religious diversity is therefore dissolved. No language game, religious or otherwise, has a monopoly on truth and on the meaning of ‘truth’.
Wittgenstein is thus opposed to both the religious exclusivist, who maintains that there is one religion which is privileged with respect to truth and soteriological effectiveness, and the scientistic philosopher who argues that the methods and techniques of science (perhaps construed broadly enough to include philosophy and logic) are our only reliable guide to truth. Instead, Wittgenstein places all language games on an equal footing, allowing a thousand flowers to bloom. Thus, as David Pears has perceptively pointed out, Wittgenstein’s later work has …an extraordinary levelling effect.
It does not assimilate one kind of discourse to another: on the contrary, it is always the differences between them that are emphasized, and particularly the difference between factual discourse and the other kinds. But it does bring all the great philosophical questions which arise within them back to the same level, ordinary human life, from which philosophy started. Philosophy is the voyage out, and the voyage back, both of which are necessary if the logical space of our ordinary linguistic practices is to be understood. 10
One of these ‘great philosophical questions’ is, of course, the question of free will and moral responsibility. On this matter, as with questions to do with religious faith, Wittgenstein refused the dominance of science on modern thinking: rather than constructing systematic theories that dictate from ‘on high’—inevitably from a position that holds up science as the model for all other discourses—how the phenomena in question are to be understood, we are brought back time and again to particular facts and examples rooted in everyday language and practices.
Taking such an approach to free will can produce startling results. For one thing, the belief in free will begins to look more like a religious commitment than a theoretical or scientific belief. Wittgenstein’s Kierkegaardian (or, more pejoratively, fideistic) account of religious belief is well known: xviii Introduction It strikes me that a religious belief could only be something like a passionate commitment to a system of reference. Hence, although it’s belief, it’s really a way of living, or a way of assessing life.
It’s passionately seizing hold of this interpretation. 11 But his somewhat similar account of free will has received less attention, despite the ever-expanding publishing industry on free will: In the sense in which asking a question and insisting on an answer is expressive of a different attitude, a different mode of life, from not asking it, the same can be said of utterances like “It is God’s will” or “We are not masters of our fate”. The work done by this sentence, or at any rate something like it, could also be done by a command! Including one which you give yourself.
12 Life is like a path along a mountain ridge; to left and right are slippery slopes down which you slide without being able to stop yourself, in one direction or the other. I keep seeing people slip like this and I say “How could a man help himself in such a situation! ” And that is what “denying free will” comes to.
That is the attitude expressed in this ‘belief’. But it is not a scientific belief and has nothing to do with scientific convictions. 13 Thus, belief in free will, much like religious belief, does not purport to express an empirical fact, but is rather expressive of an attitude, a mode of life, an imperative to live in a certain way. In the two lectures he delivered at Cambridge on freedom of the will, Wittgenstein went on to characterize belief in free will as ‘groundless’, as not supported by evidence or arguments14, again indicating parallels with religious faith.
To better appreciate this view of free will, one might introduce certain ideas from the Philosophical Investigations and, especially, On Certainty. 15 In the former work, Wittgenstein speaks of our beliefs as founded upon a ‘bedrock’ certainty: “If I have exhausted the justifications [for following a rule] I have reached bedrock, and my spade is turned.
Then I am inclined to say: ‘This is simply what I do’” (PI 217). Similarly, in OC 341 Wittgenstein states, “the questions that we raise and our doubts depend on the fact that some propositions are exempt from doubt, are as it were like hinges on which those turn” (cf. OC 88). Although Wittgenstein refers here to propositions as acting as hinges, it is commonly thought that in the more mature phase of his epistemological work Wittgenstein thought of practices rather than propositions as primary.
This, then, is no traditional foundationalism, where an inferential relationship is thought to obtain between the set of beliefs that comprise the foundations and the other beliefs we hold, with the former justifying the latter. Instead, for Wittgenstein what lies at the bottom or at the foundations of our language Nick Trakakis and Daniel Cohen xix games are not specific beliefs or propositions, but ‘ungrounded ways of acting’ (OC 110, 204), ‘communal practices’ (OC 128, 298), and ‘forms of life’ (OC 7, 358). It is in this spirit that Wittgenstein quotes from Goethe’s Faust: “In the beginning was the deed” (OC 402).
Nonetheless, our practices and forms of life can be said to show or manifest the beliefs (or quasi-beliefs) and assumptions upon which we base our lives, including such beliefs as ‘I have two hands’ and ‘The world has existed for more than 10 minutes’. However, in opposition to G. E. Moore, Wittgenstein describes these as ‘certainties’ rather than ‘knowledge-claims’, for they are not grounded in evidence or open to verification, but express an attitude and a way of acting, and so are not true or false, reasonable or unreasonable, but simply “there—like our life” (OC 559, cf. 162, 205).
It would be arguably in keeping with this epistemological account, in conjunction with the later Wittgenstein’s remarks on free will, to say that belief in free will (and moral responsibility) may also function, at least in some contexts, as one of the bedrock certainties, as one of the things that ‘stand fast for us’ in our actions and practices (cf. OC 116), or as the framework within which our ethical practices operate and are made intelligible. There are close parallels here with existentialist philosophy, where to exist as a human being and to be free (almost) come to the same thing.
Jean-Paul Sartre, for example, famously stated that we are ‘condemned’ to freedom, not free not to be free. In a similar vein, the Russian religious existentialist Nikolai Berdiaev, dubbed ‘the philosopher of freedom’, eschewed traditional accounts of freedom, which treat free will as an object that could somehow be perceived, investigated and proved or disproved from the outside, and adopted instead the Kantian position that freedom is a postulate of action: it is something we must presuppose to even think of a world in which human life and human agency are possible.
Wittgenstein would have been sympathetic to this outlook, for like the existentialists he is primarily concerned with concrete social and linguistic practices and seeks to provide a philosophical understanding of human existence that is not restricted to the explanatory framework of science (or even that of much traditional philosophy). For Wittgenstein, therefore, belief in free will, just as much as belief in God, is not threatened by scientific discoveries: “we couldn’t say now ‘If they discover so and so, then I’ll say I am free’.
”16 In line with this view, Wittgenstein spends some time in his lectures on freedom of the will in attempting to show that even if a deterministic account of the world were demanded by our best scientific theories, belief in free will need not be affected at all. 17 But he is not thereby putting forward a case for compatibilism: “All these arguments xx Introduction might look as if I wanted to argue for the freedom of the will or against it. But I don’t want to.
”18 Wittgenstein does not follow the traditional course of attempting to resolve the free will problem by proving one position or refuting another. His aim, as with other traditional philosophical problems, is to expose the problem as some kind of deep muddle or confusion arising largely from misunderstandings of the workings of language. One recurrent criticism of this view is that it appears to render the language games of science, religion, and ethics entirely self-contained and cut off from each other, if not also immune from criticism from ‘without’.
This, indeed, is a common misconception of the Wittgensteinian account of religion, and in response Wittgensteinians such as D. Z. Phillips have emphasized the many important connections between religious and nonreligious forms of life which (it is held) must be recognized if religious belief is not to degenerate into superstition. Similarly, Wittgenstein points out that scientific discoveries may have a bearing on ascriptions of free will: “A discovery might influence what you say on the freedom of the will. If only by directing your attention in a particular way.
”19 But despite these connections between the scientific and non-scientific domains, Wittgenstein insists that the distinctiveness of each language game must not be overlooked. In particular, the languages of ethics and religion must not be assimilated to the kind of fact-stating discourse and fact-finding activities that characterize the empirical sciences.