The best books on Free Will and Responsibility

What is free will?

Our interest in free will starts from our self-image. We are conscious of being agents in the world, capable of doing things and being active. We believe that we can intervene and order our own fate. We’re in control of the trajectory of our own life. That self-image immediately tracks something that is deeply important to us, which is our sense that we are also moral agents. We are accountable to one another for the quality of our actions and what flows from them.
So the problem of free will starts off at a very general level with the question ‘Are we really in control?’ In particular, is our view of ourselves as accountable, moral, ethical agents — which is intimately connected with that self-image — really accurate?

Most people feel, to some degree, in control of how they behave. There may be moments when they become irrational and other forces take over, or where outside people force them to do things, but if I want to raise my hand or say “Stop!” those things seem to be easily within my conscious control. We also feel very strongly that people, including ourselves, merit praise and blame for the actions they perform because it’s us that’s performing them. It’s not someone else doing those things. And if we do something wrong, knowingly, it’s right to blame us for that.

That’s right. The common sense view — although we may articulate it in different ways in different cultures — is that there is some relevant sense in which we are in control and we are morally accountable. What makes philosophy interesting is that sceptical arguments can be put forward that appear to undermine or discredit our confidence in this common sense position. One famous version of this difficulty has theological roots. If, as everyone once assumed, there is a God, who creates the world and has the power to decide all that happens in it, then our common sense view of ourselves as free agents seems to be threatened, since God controls and guides everything that happens – including all our actions. Similar or related problems seem to arise with modern science.

The scientific challenge is that for everything that we do, we can explain it causally. There’s some prior cause that made us do that — you can go back to childhood, to genetics, early conditioning, environmental factors. When you give the full picture, it seems there is no room for freedom.

Exactly. As in a lot of other familiar philosophical problems, critical reflection and self-consciousness about our commitments erodes our natural easy confidence, or, if you want, our complacency.

It’s a paradox isn’t it? We believe we are free but most of us also believe in a scientific world picture and naturalistic account of human beings which says that we are simply material beings and there are physical causes of our behaviour — whether those are ultimately genetic or neurophysiological or whatever. Those two beliefs seem to be inconsistent, but we cling equally to them. 

Yes, because even if, on reflection, we’re driven in the direction of scepticism, it seems incredible that we could really embrace that, because the sceptical view just seems unlivable. So there’s a deeper problem about the practical implications of scepticism itself.

We could just be in the grip of powers outside ourselves, and completely under the illusion that we’re making things happen. Wittgenstein has that example of two leaves blowing in the wind: one says to the other “Let’s go this way now,” and they get blown that way by the wind, and then they say “Let’s go that way now,” thinking they’re in control, but actually it’s the wind that’s causing them to behave the way they do…

That’s right. Examples of this kind can be traced back to Hobbes, perhaps even further back. The point of them is that we’re really just ignorant of the causes of our behaviour. Our ignorance of these causes naturally leads us to suppose we have god-like powers to bring things about. Some people think there is a problem about trying to live without any belief in free will of this kind – it would not only be depressing, it is practically impossible. What’s interesting about our belief in free will, from this point of view, is that, as you said at the beginning, we seem to experience ourselves as being controlling agents in the world and this makes it very difficult, if not impossible, to simply abandon our belief that we are free — whatever sceptical arguments may suggest to the contrary.

Are you suggesting it might be an irresolvable paradox?

Yes. There’s a very important philosopher in this area, Thomas Nagel, who wrote a very influential paper about moral luck. He describes the problem as involving not just the sceptical threat, but also the fact that the sceptical threat arises from one side of a fence, in which we have a view of ourselves as an objective part of nature. There is, however, an internal, subjective view, which is resistant to the encroachment of the objective, scientific, naturalistic, self-interpretation. That’s a very interesting problem, and I think a very deep problem when it comes to free will.

It’s usually couched in terms of determinism versus free will. What’s determinism?

Earlier you came quite close to giving at least an intuitive understanding of it, which is the view that the natural order is a causal order and everything that occurs is conditioned by antecedent events. There are laws which manifest strict regularities, such that given initial conditions, certain other events will necessarily follow, or follow in some reliably uniform, regular way. Now, there are significant problems here about how to interpret causation. In particular, the free will problem, as it’s understood in relation to determinism, operates on the assumption that whatever is caused is necessitated. That is an assumption that’s been widely held in certain kinds of empiricist or positivistic views about the nature of science and the nature of causal explanation. But it may be challenged. There’s a very famous paper by Elizabeth Anscombe that challenges it, and that’s become quite influential in some efforts to vindicate a more naturalistic conception of libertarian free will.

Let’s focus on your first book. You chose Daniel Dennett’s Elbow Room. What’s that about?

Elbow Room was written in 1984 so perhaps it’s getting a bit long in the tooth. However, I selected it because it’s a really good place for readers to start if they want to work their way into the contemporary debate. In terms of style, it’s very accessible and enjoyable. Dennett is a sharp philosopher with an engaging style. The main aim of this book is to debunk or defang the feeling that determinism presents an ominous sceptical threat to freedom and moral responsibility. Dennett sees the free will problem as a problem generated by philosophy. Philosophers abuse what he calls ‘intuition pumps’ and employ misleading philosophical analogies. By these means they generate groundless worries about the implications and consequences of determinism. An example would be, ‘If determinism is true, it would be like being in prison.’ Or ‘If determinism is true, we’re like a robot, mere mechanism.’ Dennett systematically tries to debunk claims of this kind and by doing this he wants to, as he puts it, vindicate a modestly optimistic self-image. In other words, he wants to bring us back to the original common-sense view that we talked about at the beginning. In many ways Dennett, who was a student of Gilbert Ryle’s, and in some ways belongs to the Wittgensteinian tradition, simply wants to ‘show the fly the way out of the fly bottle’ on this topic.

What does that mean?

The idea is that we’re confusing ourselves, and panicking ourselves into sceptical conundrums that make us feel worried and anxious. But when we properly and thoroughly expose these false and misleading analogies and intuitions we see there is no real problem or threat here.

So we’re like the fly buzzing around in a bottle, and Wittgenstein, Ryle or Dennett pulls the cork out and out flies the fly.

Yes. It’s good philosophical psychotherapy, that’s what Elbow Room aims to be.

The book presents the view that the kind of determinism we’ve been talking about — in terms of the scientific, causal explanation of behaviour — is completely compatible with having free will that’s worth wanting.

Yes, that’s it in a nutshell. The negative part of the book is debunking the scepticism and the anxiety generated by illusory philosophical bugbears and bogeymen. The positive agenda is that Dennett wants to show that there’s a perfectly coherent, naturalistic account of freedom and moral responsibility that’s entirely consistent with the idea that we’re part of the natural fabric of the world. Any freedom that requires something more than this isn’t worth wanting, it’s actually neither desirable nor intelligible.

So this radical freedom…I’m lost. Common sense seems to suggest you can’t have it both ways, you can’t be both determined causally and also free.

Dennett has quite an interesting and important chapter on what it is to be a self-creator or self-controller. This has influenced a whole generation of compatibilists, people who believe that determinism, or naturalism more broadly, is not a threat to our self-image as free responsible beings. The idea is that radical freedom, or what Dennett calls absolute unconditional freedom, to create ourselves ex-nihilo (out of nothing), is absurd. It’s not even intelligible for God, or at least it’s highly problematic. For human beings, however, it’s obviously absurd. We can’t have that, it’s not clear we would want it. More importantly, it’s not even clear what it would mean. What we want is to be able to reflect on our situation, to be able to consider what our interests are and how we can secure them. To give ourselves ‘elbow room,’ as Dennett understands it, is to be able to anticipate things and leave ourselves room for maneuver, where there may be variables that are unpredictable or unknown to us. This is what we actually want in order to be able to operate in the world as intelligent, rational creatures who are in control of our future. It doesn’t imply radical, 100% self-creation. It involves us developing, as we mature, and as we evolve as a species, certain capacities and dispositions of rational self-control. These can all be naturalistically accounted for and help us understand the basic distinctions between us that really matter. If, for example, I’m dealing with a small child or an animal, or a severely brain-damaged person, it’s precisely those capacities that they lack. Dennett’s common-sensical point of view is that what we’re looking for is something perfectly ordinary and intelligible that we can easily identify and recognize.

I can understand that, but it does seem as if that capacity, those dispositions, can all be completely accounted for in causal terms: there’s only an illusion of free will. It’s not as if the self-creator is actually genuinely self-creating, it’s just going through the motions of something that it’s been programmed to do.

This is an important point and Dennett’s critics would certainly press this concern. This is why, although Dennett is a good place to start, you also have to read the other books I’ve recommended. It might be a bit unfair or too severe to say that Dennett is complacent, given that he offers substantial and interesting arguments for his position. But I think his critics would say — and I have considerable sympathy with this — that his optimism comes too easily. Dennett simply wants to dismiss this sort of concern as groundless. However, while it may well be true that there is something problematic about articulating what it would be to have absolute freedom, there may still be something we care about that is missing or that we aspire to but lack. Perhaps the aspiration to immortality is impossible and doesn’t even make sense, but an awareness of our mortality may still disconcert or trouble us.

Let’s go on to the second book, Four Views on Free Will, because that’s mapping out the territory, in a way.

I suppose I’m cheating here because I’m squeezing four books into one. However, I would particularly recommend this book to readers, as the four authors of this book — Robert Kane, John Fischer, Derk Pereboom and Manuel Vargas — are all significant figures in the contemporary field and have mapped out distinct and influential positions. The reader can quickly get a sense of the general lie of the land and some of the basic available options. Just to sketch them quickly:

Kane is a libertarian, which is to say he’s an incompatibilist who thinks that for us to be free and responsible determinism cannot be true. For this we need an alternative metaphysical framework that accounts for the sort of freedom required for responsibility, which involves more than mere indeterminism. 

Fischer is a semi-compatibilist. He thinks determinism doesn’t threaten responsibility, but it does threaten free will understood in terms of open alternatives. Fischer argues that we have a rational self-control that involves what he calls ‘reasons responsiveness.’ That ability serves as a foundation for moral responsibility, but falls short of full, metaphysical freedom. Although we don’t have genuine open alternative possibilities, we don’t need them, Fischer argues, for moral responsibility. In several respects this view resembles Dennett’s general strategy, except that Fischer doesn’t buy into the freedom part of it. Rational self-control allows us to recognize reasons and respond to them in a more or less reliable and consistent way, and that’s all that responsibility requires. 

Pereboom is a hard incompatibilist, or what he sometimes calls an ‘optimistic sceptic.’ According to this view, while we are not free or morally responsible this is not that depressing. We can save most of what we actually care about, including our interpersonal relations, through some form of moral assessment that doesn’t rely on strong moral responsibility. We can, moreover, live lives that are still meaningful and personally satisfying. The supposition that scepticism does not imply pessimism is the main thrust of Pereboom’s line. 

Vargas calls his position ‘revisionism.’ Roughly what he says is that we should make a basic distinction between what he calls the ‘diagnostic’ understanding of freedom and responsibility — what we ordinarily think freedom and responsibility is — and a ‘prescriptive’ understanding. In our ordinary life, we may well have incompatibilist natural intuitions, and while we may not be able to save those, we can revise our conceptions of freedom and responsibility so that we still retain a suitably robust account that serves all the things we usually care about. Although this is revisionary it’s not radical scepticism. Vargas wants to avoid throwing the baby out with the bathwater and stops short of unqualified, extreme scepticism.

The position we should come back to for a moment is Kane’s. There’s a standard objection to classical libertarianism which is that it relies on ‘spooky metaphysics,’ postulating ghostly agents in some way intervening in the natural order. What makes Kane’s libertarianism interesting is that he argues that libertarians can avoid this and give a naturalistic account of the possibility of moral freedom. The idea here is that what’s caused doesn’t have to be necessitated. Agents can be free as long as there are crucial points in their lives when they are capable of making what Kane calls ‘self-forming’ decisions or actions. What that requires is that free and responsible agents have more than one reason that they can act on in the very same circumstances, and that in those very same circumstances, either reason could move them to act. Whichever way they act, however, they act for a reason. While it is not determined it is still not capricious. In these circumstances, Kane claims, free agents have genuine alternatives and are the ultimate sources of their conduct and character.

This makes me think of Kant…

Lurking behind Kane’s model there is something like Kant’s picture of duty struggling with desire at crucial moments of choice. Kane gives an example of a businesswoman who is running to a meeting. Somebody needs help, but she’s got to get to the meeting. Roughly, duty = help, desire = get to the meeting on time. In the exact same conditions, at that particular moment, either reason could move her to act. Kane builds up his model and makes it more complex by suggesting that there are extra things going on — mental items called ‘efforts‘ — and he suggests that we can simultaneously make an effort in both directions. You could, on this account, replay the tape several times and you’d get different outcomes.

Despite its interest, I find the model somewhat suspect. What’s crucial to Kane is that we have a plurality of reasons available to us, what he calls ‘plural voluntary control.’ But there are often cases where you may not have any alternative at all. There may be no other reason that comes up, but you still fully embrace the reason you acted on. It seems to be one that is genuinely yours, and you’re responsive to the reason that’s available to you when you’re moved by it. But for Kane, when it comes to self-forming actions that are fundamental to the possibility of moral responsibility and genuine free will, there must be occasions when we have, as he presents it, 50-50 options. This model may become very odd. What if you had a 99 versus 1 option? Suppose you have a reason that’s a very weak reason, but in the same circumstances there’s just the chance you might act on it. It seems to me that Kane and those who take a similar line are committed to the view that this would be enough to give you genuine alternative possibilities, even although the probabilities aren’t 50-50.

It seems to me that it’s a rejection of the notion that the causal explanation is there at all. He’s opting for free will at the expense of the scientific picture. It’s usually insufficient knowledge of the causes that leads us to think we’re behaving in different ways in similar circumstances. The point is that they’re not exactly similar: if they were exactly similar, we would have acted the same way.

As you know, that view can be rejected as a kind of metaphysical, necessitarian prejudice. The claim is that, in fact, when we look at what science tells us about quantum phenomena and the indeterministic order of things, especially now, in the 21st century, we shouldn’t endorse that picture. What we need is a probabilistic conception of causation rather than a necessitarian one. It’s exactly that wedge that Kane uses as the metaphysical or ontological foundation of this alternative picture. He also uses computing analogies, and suggests that you can have parallel processing systems where the outcome isn’t always the same. The system can be in a seemingly identical state, but different outcomes will issue from it. Same inputs, different outputs.

The question of whether physics — and specifically the quantum mechanics — illuminates free will or just muddies the water is a tough one.

Yes and in relation to Kane’s model, the big question is, has he really gotten rid of the problem of luck or chanciness? This is sometimes presented in terms of the contrastive question. If you have an agent who in given circumstances performs their duty, say helps, but another time, in identical circumstances, does not, the obvious question you want to ask is, why did they help in the first case but not in the other? According to this model, however, there is no further answer or explanation to be given for this variation. But then it looks like it’s just luck and the agent lacks adequate control over what they actually do in these specific circumstances. Although you may be in control in the sense you act on one reason or another, and what you do is intentional and done for a reason, the crucial problem remains that you’re not in control of why you act for one available reason rather than another. I don’t want to be unfair to Kane, because he does try to deal with this concern, but in my view this remains a serious problem for his theory. The sort of problem he faces here is just another version of the same problem that compatibilists face, that there are limits to control.

Let’s move on to the next book, RJ Wallace, Responsibility and the Moral Sentiments. Questions of moral responsibility are intimately tied into questions of free will. Part of what makes metaphysical questions about free will interesting is that they do have real life applications, because we live in a world of praise and blame and punishment and responsibility of a kind that is moral. It’s not just an arbitrary, armchair discussion. It feels like it’s part of what holds us together as human beings. This is something that is about us.

This is an important point. While I think it would be too much to say that the free will problem is just the problem of moral responsibility, they are intimately connected. Free will stretches beyond the problem of responsibility, since it touches on our conception of ourselves as creators, individuals and so on but the issue that really matters to us is the problem of our agency in relation to our moral accountability. This is why Wallace’s book is particularly interesting. As the title suggests, Wallace is drawing on a tradition that emphasizes the importance of moral sentiments for understanding this issue. The approach he takes follows that of an influential paper by Peter Strawson, an Oxford philosopher, called “Freedom and Resentment.” The approach here is to begin with an understanding of what it is to hold somebody responsible, where holding somebody responsible is understood as a matter of entertaining certain distinctive kinds of emotions towards them. Those emotions presuppose certain kinds of beliefs about them, and then we try to understand what it is for them to be responsible, by properly understanding the emotions involved in holding people responsible. So we get a better insight into the conditions of responsibility, by examining their foundation in these attitudes and practices.

What sort of emotions are we talking about here?

The philosophical jargon is ‘reactive attitudes.’ We have a certain set of expectations or standards or norms, concerning what we expect of one another in terms of our inter-personal dealings. These have ethical or normative significance for us. When people violate these norms we respond negatively to them. We hold them to those standards. What’s crucial here is that these aren’t just intellectual judgments. We must not, in Strawson’s language, over-intellectualize our responses here. It’s not like failing an exam or getting a math problem wrong – it goes deeper than that. Our responses in these circumstances are hardwired into our moral psychology. We respond emotionally and have hostile responses in those cases where these moral norms are violated.

So if somebody insults you, is that the kind of thing that produces a reactive attitude? They broke some sort of moral taboo?

Yes, Strawson doesn’t start with the impersonal moral case, but with the personal reactive attitudes, where it’s an injury or a harm done to us. If you step on my toes accidentally, it might hurt me, and I might think, “please be more careful.” However, if I think you’re jumping on my toes and aiming to hurt me then, obviously, I will get much more upset.

What does this show us about responsibility and blame?

What it shows us — and this is the Strawsonian side, that can be traced back to philosophers like David Hume and Adam Smith — is that we should start, not from some high order of moral principle or philosophical theory, but from a simple truth about human moral psychology. The relevant starting point of these investigations is that we actually care about certain kinds of attitudes and intentions that other human beings show us, just as we naturally care about our appearance, intelligence, abilities, physical prowess, and so on. Moral reactive attitudes are concerned with a particular dimension of our attributes or qualities, namely the set of attitudes and intentions and the values we manifest in dealing with each other. What’s interesting about this general approach is that it moves away from simple conceptual analysis — what concepts mean, what their logic is and how they’re related to each other — and tries to provide a better informed and more realistic moral psychology.

How does Wallace use that in relation to free will?

His model is something like this: reactive attitudes should be understood in terms of expectations. Expectations are simply standards or norms about what we may or may not do. These expectations lay down our obligations. Where those obligations are backed by moral reasons — which concern our relations to each other in social life, so that we can cooperate and trust each other — they ground our reactive attitudes when violations occur. In several respects, Wallace defends a fundamentally Kantian view of morality. When moral obligations are intentionally violated we blame people, and blame is naturally connected with our retributive disposition to punish or sanction them in some form.

This understanding of our moral psychology has to do with the interconnections between expectations, obligations and blame. It’s perhaps a weakness of this view that it leans entirely on the reactive attitudes that are all essentially negative or hostile responses to violations of moral expectations. Wallace goes on to explain that to understand our stance of holding responsible we need a theory of excuses and exemptions. As in familiar legal cases, there are certain circumstances where people appear to violate our expectations, but then we realize that they were either ignorant of the situation, or they did it accidentally, or there was some other relevant consideration indicating that although an injury occurred, the harm done was inadvertent and therefore the expectation was not strictly violated. Voluntariness and intention are essential here in triggering or occasioning our reactive attitudes. There is another important consideration that we must also take note of. Clearly, there are individuals who we see as appropriate targets of reactive attitudes or moral sentiments and some who we don’t. How do we draw that borderline? How do I decide whether an individual I’m dealing with is someone who should be included or excluded? What if, for example, it’s a crazy person or a child? This is where Wallace, again, like Dennett and Fischer, brings his own particular model of rational self-control. He tries to give a naturalistic account of rational self-control which can serve as a plausible theory of exemptions for reactive attitudes.

Is he just assuming that we have choice, that we do have genuine free will?

He certainly offers a powerful compatibilist theory. The picture he rejects is the one that runs deep and is still there, the view that responsible choice depends on genuine alternatives or open possibilities. Wallace rejects that. He thinks someone makes a responsible choice as long as she has a general disposition or capacity to recognize reasons and may be motivated by them. Although it is true that in some circumstances an individual might fail to recognize reasons or fail to be motivated by them, we should remember that someone who has the capacity to speak a language may also make a grammatical slip. Nevertheless, we still have the expectation that they will speak in an appropriate, literate manner and we hold them to these standards. When they make a slip we catch them out and respond to them and correct them. In the case of our reactive attitudes, the people we target are the people with the relevant general capacity or disposition. They might make a slip that has an explanation but if it was willful or intentional, then we can hold them to those standards and we will react to them accordingly. We will blame them.

OK so what Wallace has done is given an emotional account of our judgments of moral responsibility. He’s a compatibilist, so he’s sketched out something that will help us better understand what it is to blame someone, to hold someone responsible. He’s nevertheless, like Dennett, quite optimistic about the compatibility of determinism and free will. The next author, Bernard Williams, is far more pessimistic about free will and the human condition. I know you knew Bernard Williams because you were taught by him, can you tell us why you chose this book, Shame and Necessity?

This is a highly regarded book, though in terms of the standard free will literature, it’s very different in its approach. Unlike some of the other books I’ve mentioned, it’s not an easy book to read. There’s no simple position or model that Williams is interested in articulating. It’s a book that’s focused on ancient Greek conceptions of agency and responsibility. Not only is he interested in the contrast between the ancient Greeks and us, but he’s also interested in the difference between the Greek tragedians — Sophocles, Aeschylus and Homer — on the one hand, and Socrates, Plato and Aristotle on the other, thinkers who are really the founders of the western philosophical tradition. What’s interesting about this book is that it’s not just about free will, it’s also about our views concerning the nature of philosophy. Williams argues that methodologically, philosophers need to be historically sensitive and informed, and that typically they aren’t. That’s a problem with many discussions about free will, there is a lack of historical self-consciousness. One of the great merits of Williams’s book is that he is, among other things, a distinguished scholar of classical literature. Not only is he open to the possibility of learning from history, he also thinks we can learn from literature, and from tragedy in particular. So it’s a very interesting and challenging book.
Williams has multiple agendas – including a very ambitious agenda that involves a critique of our whole modern conception of morality. There’s a well-entrenched view that the Greeks didn’t have any adequate conception of freedom and responsibility at all. They were, it is suggested, full of primitive views about fate and the gods and the limits of agency and they didn’t really appreciate the importance of intention and voluntariness. On this view, we moderns have managed to get past all this and now, fortunately, have a sophisticated, adequate conception! Williams argues that while the ancient Greeks’ views were certainly very different from ours, and there is no question of going back to the world they lived in, they nevertheless understood and appreciated things that we can still learn a great deal from. Part of Williams’s project in the book is to suggest that there are very important respects in which the early Greeks were much more realistic about the human predicament as it concerns human agency itself, and in particular our vulnerability to fate and luck as it affects our own moral or ethical lives.

That connects with a topic Bernard Williams wrote about, which is moral luck. Certain pictures of what it is to be human seem to make all our moral choices immune from any particular circumstances in fortune. But actually, imagine these two cases: you drive your car home a bit drunk and nothing happens. Nobody stops you, you’re safe and sound. In the second scenario, you’re just as drunk, but somebody steps in front of the car. You knock them over and tragically kill them. Luck came in there, but in the second case we, perhaps irrationally, say you’ve done something worse: you’ve killed someone through drunk driving. But whether you killed them or not is ultimately not up to you, someone just happened to step out that you didn’t see. There’s so much luck involved as to whether you’re in that circumstance or not – and that determines your whole fate as a human being.

Exactly. One common-sense view is you’re only responsible for what you control. An absolutely vital vein of Williams’s book is to say that the Greeks didn’t have that view and that, in fact, the Greeks are more accurate about this. There’s a kind of dishonesty built right into our moral system now, which trickles into the free will problem. We try to evade or cover up this troubling truth about the human predicament as it relates to agency and morality.

And the truth is?

The really troubling truth is that, on one side, we are free and responsible, and it is evasion to try and deny that. We’re aware of this ourselves and, for all the reasons we have already discussed, we have the relevant abilities needed to see ourselves in these terms. On the other hand, it is also dishonest to suggest that the exercise and operation of those capacities that render us free and responsible somehow leave us immune to fate and luck. Almost all the other authors who I’ve mentioned so far are, in various ways, evasive about that and want to be more optimistic. This is what Williams rejects. He uses the Greeks and a genealogical or historical self-understanding to argue this point. It’s really a very powerful book. On the face of it it’s about free will-related problems, but it cuts deep to our whole ethical self-image and our existential predicament.

So the essence of his argument or exploration is that we could learn from the ancient Greeks and recalibrate, perhaps, our sense of what it is to be human and free and build in a greater sense of the way fate or things outside our control affect what we are.

Yes. My own inclination is to use the word ‘fate’ but it might have baggage that you want to avoid, so you could say our vulnerability to contingencies that are not of our own making in the exercise and practice of our own ethical lives. We can’t escape this. The Greeks spoke in the language of the gods, we might talk about something like ‘blind nature’ landing us in certain predicaments where the consequences of what we do could be catastrophic. But we can’t evade responsibility in these circumstances by saying “Oh that was just bad luck or fate!”

Because those very circumstances determine what we are. We can’t say, “It’s bad it happened to me!” It’s part of what I am, that it happened to me.

The Oedipus story is, for Williams, very powerful. From the point of view of western aspirations to be free agents, Oedipus might just say, “Oh well, I don’t have a real moral problem here at all. It looks like I’ve murdered my father and married my mother, but since the gods have arranged all this I can wash my hands of it – I am not really responsible.” But it’s not like that. Even now we recognize that this is not truthful about our predicament because otherwise the tragedy of Oedipus wouldn’t speak to us — and clearly it does.

Does that connect again with this notion of reactive emotions? Is that a reactive emotion to Oedipus’s own circumstances?

I think so. Williams spends a lot of time talking about this situation and other examples from ancient literature, where the agents can’t live with themselves. Even though they’ve been, in some sense, caught up in fate or bad luck, they’ve nevertheless performed deeds that they cannot live with. This may include things they did not intend or do voluntarily, which is a theme Williams draws our attention to.

It strikes me that some of the things you’re talking about resonate with themes in Thomas Hardy’s novels. The persistence of the unforeseen in how events unfold and the shaping of lives outside the participants’ control.

This is, no doubt, where philosophy meets temperament. These aren’t just theoretical problems, they resonate with our own metaphysical attitudes, our hopes, fears, and so on. I find Hardy very powerful in these respects. He was quite interested in these ancient paradigms, such as being caught up in a bad fate, in circumstances where you’ve got a cloud of doom over you. But you’re not just like a leaf that is cast along by the wind, because a crucial aspect of your fate is the way it works into the exercise of your own agency and character. That’s what makes our fate and our human predicament ethically interesting and powerful.

Let’s move from the ancient Greeks to a more recent phenomenon, the effects of neuroscience on understanding choice. Some people have suggested that various neuro-scientific experiments undermine this notion that we have free will at all. They are providing empirical evidence that free will is actually an illusion. Something as simple as moving my hand, the neurophysiological impulse that brings about that movement occurs before I have the intention of moving it, even though it feels the other way around. That’s what Benjamin Libet claims, and it’s quite a radical claim. There may be flaws in the experiment, but a lot of neuroscience is piling up to suggest that picture of ourselves as somehow sitting in the driving seat may be wrong.

In the fifth book I’ve selected, Waller’s Against Moral Responsibility, one of the sources he appeals to in support of his scepticism about moral responsibility is data coming from neuroscience. This reflects a wider trend in philosophy, which is to be better informed and more deeply integrated with advances in empirical science and to use these resources to help us understand philosophical problems. To a certain extent, this approach cuts in the opposite direction from Williams, who employs a more humanistic understanding of philosophy, as informed by history and literature. However, neuroscience is very influential in the philosophy of free will right now, as are certain kinds of psychological experiments that aim to debunk our confidence that we are agents making conscious choices and in control of ourselves. The gist of Waller’s book is that we aren’t really responsible at all, because the kind of control that we think of as essential to responsibility is illusory. His principal argument in support of this conclusion is that we are vulnerable to luck, but another layer of his scepticism relies on neuroscience data — and, as you say, Libet is really the major figure here. While we may have a conception of ourselves as conscious agents making choices, the empirical data that Libet has provided claims to show that it is not the conscious self that decides but prior events in the brain. It can be shown experimentally that the brain has already settled how we will act several hundred milliseconds prior to our awareness of making any conscious choice. This supports a seemingly sceptical position about the role of conscious choice, on which responsibility seems to rest, since how we will act is already settled prior to the occurrence of conscious choice itself. The conscious agent, it seems, is not really in charge of conduct.
A related but distinct strategy we should mention here is what’s called the “situationist objection.” This draws not so much on neuroscience as on advances in social psychology. Here again there are some seemingly embarrassing experiments. For example, in one, a coin is left in a phone booth and some people are left to find it. Shortly after this a person drops a book in front of them. Will they come and help pick up the book or not? The experimental data suggests that those individuals who just found a coin in the phone booth are much more likely to help than those who didn’t. The conclusion drawn from this is that what we do depends not so much on our particular character – whether we are kind, helpful, and so on – but on the particular situation or circumstances we happen to find ourselves in. More specifically, seemingly trivial or irrelevant background conditions, such as finding a coin, can greatly influence what we actually do. What matters to conduct is primarily a function of the specifics of our situation and not our character.

There’s another great example like that, that people are more generous outside a store from which is emanating the smell of freshly baked bread, than they are outside a hardware store. Their whole moral stance in terms of generosity has been transformed by a pleasant smell.

Exactly, and I suppose you could put people into a bad mood by having an unpleasant smell and then they will behave in less pleasant ways due to factors they may not even be aware of. Again, the conclusion here is that we lack the sort of control we generally think is necessary for moral responsibility. One response to this, which Waller does not endorse, but philosophers like John Doris do suggest, is that for the purposes of responsibility we should aim to control our situation. For example, if I know that when I go to parties I drink too much, and when I drink too much I behave in ways that I subsequently regret, I should avoid going to parties where heavy drinking is going on. This is a better strategy than trying to abstain or limit how much I drink at the party. In general, if I am more aware of these situational factors I will be in better control of my conduct.

Just to get clear about Bruce Waller’s book, he’s arguing that we don’t have, despite appearances, any moral responsibility for the choices that we make?

Right. He allows that there are various weak or feeble versions of moral responsibility that compatibilists may propose. They want to lower the bar, to accommodate their naturalism by making responsibility less robust or less demanding. This is just a kind of verbal trick as Waller sees it. What is in question is the genuine article of real, robust moral responsibility. This involves a commitment to moral desert, which sustains our retributive attitudes and practices, particularly punishment. According to Waller, the crucial question is can we vindicate this more robust understanding of moral responsibility?

Another aspect of Waller’s book, and a major concern of his, is that he thinks there’s a view that this sort of scepticism implies a kind of pessimism which supposes that it would be terrible if we’re not really, truly responsible. Against this view Waller argues that, on the contrary, these illusory beliefs about robust moral responsibility propel us into all kinds of rather nasty and unnecessary social practices, almost all of which we are better off without. These views also encourage us to remain ignorant of the relevant causes that make people commit crimes. Instead of investigating and identifying these causes, the responsibility system, simply punishes people or aims to cause them grief, which doesn’t solve anything. Scepticism offers us a way out of this mess, as Waller sees it. Abandoning the responsibility system is a basis for optimism not pessimism.

But if we’ve got no moral responsibility, doesn’t that just lead to us doing whatever we feel like doing? Because many people feel that the notion of moral responsibility carries with it limits on behaviour. That’s the point of it, in a way.

You have to believe, on this view, there’s some sense in which we can live without the reactive attitudes we are considering. One question right up front is, ‘Is this even psychologically intelligible?’ One way to appreciate the difficulties here, by analogy, is to imagine that someone suggests that, ‘Being afraid is an unpleasant state of mind, you’re far better off being purely rational in assessing a threat or a harm and not allowing yourself to get upset about it.’ We ought to stop feeling fear or ever becoming afraid. You might imagine, for example, a soldier in a threatening situation might try this exercise. At least one problem we face here is that it is extremely implausible to suggest that people can just step away from responses of these kinds, even when they have practical reasons for doing so. Much the same seems to be true of our reactive attitudes. Suppose someone harmed your child, willfully, in a very cruel way. Would you be able to live your scepticism? Waller says it’s an ideal we can aspire to, but I am psychologically sceptical about that. And even if it was psychologically possible, there’s the further question of whether or not it would be desirable, for just the reasons you’re suggesting. There may be significant costs in terms of the sorts of pressures we exert on each other to conduct ourselves in ethically desirable ways. However, on Waller’s view, these are costs we must be willing to pay since our reactive attitudes and practices are unfair and unjust, whatever social benefits we may derive from them.

Let’s take a real example. Ludwig Wittgenstein when he was briefly a school teacher – and he wasn’t a very good school teacher – hit a young kid so hard that his ear bled. He later came to seriously regret that. He blamed himself and he was very hard on himself morally. Is Waller saying Wittgenstein blaming himself for his earlier action is absurd?

The strategy being advanced here is not that we don’t care about what we do or how we act. Clearly if a wild animal came into a room and harmed a child, that would be awful. What we do in cases of this kind is to figure out why this happened and try to make sure that it doesn’t happen again. We might do the same with crime or any other sort of nasty, unpleasant action. The right approach, on this view, is to see the bad conduct as a natural event and to correct it, rather than just striking back. This is true even in first person cases, such as the one you have described involving Wittgenstein. Suppose Wittgenstein reflects on his own character and conduct in this situation. According to Waller’s sceptical view, there’s no point in saying ‘Oh my goodness I really deserve to suffer and feel horrible guilt about this conduct.’ Instead, given his troubled background and complicated personal life and relations, we might be able to come up with some sort of explanation for his bad temper and violent treatment of the child. The crucial thing is for the agent involved to learn why he acted this way and to change himself accordingly. All of this is consistent with having standards that say ‘This is undesirable conduct, this is something that is wrong.’ But being wrong and being morally responsible are two different things. We want to prevent wrong conduct in the same way that we want to prevent all kinds of other unpleasant things that occur: illness, hurricanes, fires. Ethical behaviour that fails our standards is not something we should be emotionally responding to in a retributive manner. We should aim to understand its roots, its causes, and try to improve things in the future.

I listened to a radio programme recently about a shark expert. As part of his job he pulled out sharks from the sea and tagged them. He pulled out this big shark and it bit a chunk out of his leg. He said, ‘I don’t blame the shark — it’s what sharks do.’ But if that had been a human being, who’d attacked his leg with a machete, I don’t think he would have felt the same. He wouldn’t have let him go saying it was nothing. Isn’t there a danger of this sort of view, that you rationalize all the worst behaviours? You find a causal explanation and there’s no room left for judging people and then anything goes.

It may be argued that this is a vulnerability for the sceptical position – although, as I have explained, Waller wants to resist this objection. Nevertheless, we need to be able to distinguish between our relations and responses to sharks and to people, given that people possess abilities and capacities that sharks lack. Obviously, for example, the shark doesn’t have any reactive attitudes of a kind that people are liable to, much less any capacity to see themselves in relation to relevant kinds of moral expectations or any ethical self-understanding. Compatibilist considerations along these lines certainly help to explain why we don’t respond to sharks and people in the same way. I’ve got a small puppy right now. I sometimes have strong reactive attitudes to him, like “That damn dog has eaten my book!” I know this is unreasonable but at times it’s hard to switch off. We do, at times, have unreasonable reactive attitudes. What seems extreme about the radical sceptical view is that it aims to extend this reasoning to all cases – including all human action and agents. In my view there is a kind of dishonesty or bad faith involved in this outlook because it involves ignoring very relevant and important distinctions between people and animals, or even adults and children, and why it is important and necessary to respond differently to them. In those cases where the individuals you’re dealing with understand ethical responses, and their basis in moral concerns and norms, our emotional responses not only seem appropriate, they are essential to showing that we take morality seriously and place appropriate value and emphasis upon it in our practical lives. The sceptic thinks we can make do without these responses – and that we are even better off without them – but I am unpersuaded by this.

It’s also almost impossible to do on a human level, to think of yourself as completely part of the causal order with no room for choice, and that every time you think you’re making a choice, you’re actually reporting on a choice your brain has already made and committed you to act on. I really struggle to see my own behaviour in those terms. There are elements of that behaviour, I can believe that when I am driving a car, I start to turn left and then think I’ll turn left. I can believe that, because it’s semi-automatic behaviour. But there are other things which feel so much as if I’m doing them, rather than that they’re just happening….

One interpretation of the neuroscience data is that our experience of conscious choice is all just epiphenomena – with no real causal traction in the world. Whatever is going on in our conscious states, the real, operating causal forces are brain states that we’re not even aware of. I have to confess that I don’t get as excited about these findings and interpretations as many of my colleagues do. Nevertheless, these findings do raise genuinely interesting and important questions about the nature of deliberation and choice. How we understand all this in relation to data about what’s happening in the brain when these activities are going on is both scientifically and philosophically very interesting and challenging. Having said this, too often scientists, and those who follow closely behind them, draw strong philosophical conclusions, with great confidence, which it is not obvious that they are entitled to. The extreme sceptical conclusion, for example, which is now rather fashionable, strikes me as naïve because it generally turns on a particular model of what freedom and responsibility are supposed to involve, namely a spooky metaphysical self that is making conscious choices and in doing so interrupts the course of nature. It may well be that the data provided does debunk or discredit problematic models of this kind. However, for more sophisticated, and in particular naturalistic compatibilist models, it is not clear to me that this data is so problematic at all. For example, it isn’t going to surprise compatibilists, who are committed to naturalism, that our choices and actions have causal antecedents in physical processes that can be traced back prior to those choices and actions. That’s something that compatibilists as far back as Hobbes, if not before him, have long been committed to and acknowledged.

I think there’s a tendency, in the present spirit of our times, to think that science is going to provide us with a fundamental ‘epiphany moment’ – which will suddenly clear the air and reveal our illusions on this subject. Perhaps people are looking for this because they see it as both exciting and decisive: ‘Here’s an experiment that shows – proves – that there is no free will and that we’re not responsible after all!’ To my mind this is all much too quick because it generally turns on rather superficial and implausible conceptions of what we’re actually committed to with respect to these matters. Waller’s response to these concerns is not without its own set of distinctions. He is careful to emphasize, for example, that we may have natural forms of freedom, while still insisting that these do not serve to ground or justify what he takes to be the responsibility system. Even allowing for these qualifications, he remains sceptical about our existing commitments and practices. With regard to these he wants to say, ‘Of course there’s a lot of things that we once believed that have turned out, on examination, to be wholly illusory and mistaken. We used to believe in ghosts, that the sun went around the earth, and so on. This is just part of the process of scientific advance.’ In the same way, Waller suggests, our thinking and attitudes concerning moral responsibility also need a Copernican revolution.

Is it fair to say, then, that you’re much more in the Bernard Williams camp? That we need to understand free will historically, and that we can learn from the ancient Greeks about what are choices are, and where our fate lies?

Roughly speaking, that is my, perhaps unfashionable, view. Williams is, of course, a hugely influential figure, but he doesn’t represent the main trajectory of contemporary philosophy, which is now heavily focused on the interpretation of science-based findings. Generally speaking, Williams is sceptical about the dominant effort to turn philosophy into a handmaiden of science, except in those specific areas that call for this. The overall view that I find most plausible and truthful about our human situation is that — contrary to the sceptical view that we just considered — we can defend a robust understanding of both freedom and responsibility, that can be provided in entirely naturalistic terms. However, what’s interesting about the sort of view Williams describes, which goes all the way back to the ancient Greeks, is that this does not serve to vindicate an easy optimism about the human condition. On the contrary, understanding our situation in these terms grounds a pessimism that is rooted, not in the sceptical thought that we aren’t really free and responsible, but in a more difficult and subtle truth about our predicament, which is that while we may well be free and responsible agents, we nevertheless remain vulnerable to luck, contingency, and aspects of fate in the exercise of moral life itself. It is our awareness of this that we find uncomfortable and tend to resist. Most available theories in the free will debate seek, in various ways, to evade this pessimistic conclusion. Although this understanding of our predicament as human agents may not be one we find comfortable it is, nevertheless, the most truthful interpretation.