:: Article

Time and the Philosophy of Action

Interview by Richard Marshall.

We can describe all the structures essential to a rubber ducky without taking time into account. Rubber duckies, of course, are involved in complex moving relations with other bodies, especially in water, and these relations involve time, but the temporal considerations needed to make sense of these motions are external to the ducky itself. Nothing in the metaphysical structure of a rubber ducky requires that it move in one way rather than another, and if the universe were suddenly frozen at a single moment in time, rubber duckies would still exist. Actions, on the other hand, would not.’

According to the A-theory, reality is essentially tensed: there is a past, a present, and a future. According to the B-theory, reality is tenseless, so when I say that I am going to finish this interview in the future, the fact about reality this captures is that the event of my finishing this interview occurs later than the event of my saying that I am going to finish it. So on the B-theory events are temporally related to each other by virtue of occupying different points in time rather than by virtue of happening in the past or future relative to some moving present; time does not really pass, but only seems to.’

A billionaire has offered to give you a million dollars if, at midnight tonight, you intend to drink a toxin tomorrow afternoon. The toxin will make you sick, but it won’t kill you, so you wouldn’t mind drinking it for a million dollars. But there is one catch: the money will be deposited in your account (or not) before noon. So here is the problem: you have no reason to drink the toxin (and the billionaire has told you as much), since he isn’t paying you to do so. By the time tomorrow afternoon arrives, you will either be a millionaire or you won’t be, but you will have no reason to drink the toxin and a strong reason not to. Since you know this, it seems that forming an intention to drink the toxin will be difficult or even impossible without finding some way to trick yourself into drinking.’

Roman Altshuler is interested broadly in human agency and issues surrounding it, including free will, personal identity, moral philosophy, and, a bit more distantly, issues of death and its relevance to the ways we create meaning in life. Many of our actions reflect background commitments on our part, and these commitments, in turn, shape who we are. So it would seem to follow that we can create ourselves and shape our identities simply by tweaking our commitments. At the same time, however, we didn’t get to choose where we were born, the people who raised us, or any number of other factors that have shaped our identities in countless ways. So while it might seem like we have an unlimited ability to shape our identities by tweaking our commitments, it also seems as if the commitments we select stem from influences on us that lie beyond our control. His hunch is that we can resolve the problem by understanding the role time plays in our commitments, and especially in the way those commitments are affected by our relation to the future and the narratives we use to shape our agency. Here he discusses the relation of time to action, why explanations of action require a tensed reality, basic actions, Kavka’s toxin puzzle, Heidegger and free-will, John Fischer and temporal loops, whether Heidegger is a narrativist about selves, Sam Scheffler and the afterlife, responses to the the Doomsday Scenario, whether Bernard Williams is right about whether we should desire immortality, and Kant’s views about evil. This is a long walk…

3:AM: What made you become a philosopher?

Roman Altshuler: My maternal grandfather once wanted to study philosophy; he wisely settled on math, as in the Soviet Union philosophy often meant Leninism, and doing it professionally was liable to get one arrested whenever the official orthodoxy shifted. My grandmother had a PhD in German literature, though she detested philosophy. My father is an engineer, but he has a longstanding interest in philosophy, and philosophy books have been on his shelves since my teens; whenever I visit these days he is reading a new book about memes. His favorite author is Dostoyevsky; he read Crime and Punishment and The Idiot out loud to my mother and me on camping trips. So it would be somewhat disingenuous for me to claim that my interest in philosophy was sui generis.

In high school I joined the debate team and went to debate camp, where a cranky grad student took us on a tour of Hobbes, Rousseau, Locke, Mill, Kant, Marx, and Rawls. When I got home I obsessively started reading all of these, thanks to an incredibly affordable used book store. I wasn’t personally invested in ethics or political philosophy back then, but I loved the way philosophers worked through ideas. Then I discovered Nietzsche, read Kaufmann’s Existentialism cover to cover, and spent my junior year working my way through Freud’s non-clinical works. But I still wasn’t sure if I wanted to go on to study philosophy or biology, which was my other major interest at the time. Anita Zimmerman at Brown was kind enough to hire me as an assistant in her physiology lab in the summer before I went off to college, but unfortunately the experience had the opposite of the intended effect. While I found injecting modified RNA into cells under a microscope fascinating, I got the distinct sense that I didn’t have the patience necessary for lab work. In my first year of college I took Richard Moran’s course on existentialism, where in addition to Sartre and Beauvoir, he made us read Frankfurt and Watson. I’d been reading philosophy for a few years by then, but this was my first exposure to analytic philosophy, combined with Moran’s lectures, which were almost certainly too hard for freshmen but unmistakably original. At the same time I was reading Kierkegaard and Buber in a religion class with Kimerer LaMothe. The following semester I was back in Moran’s class (aesthetics, this time), but had somehow picked up a copy of Baudrillard’s Simulations while taking a class on Wittgenstein’s Tractatus, and I was thoroughly hooked on this very weird discipline. You can probably tell that the analytic/continental divide didn’t have much of an effect on me; I simply made up for the philosophy department’s emphasis on the former by taking classes in Comparative Literature and Art History.

3:AM: One of your focus areas is the philosophy of action. One thing you’ve recently thought about is the relation of time to action and agency. You’re not so much interested in external temporal constraints on agency but internal ones – so perhaps we’d better start there by asking you to sketch out what philosophers mean when they talk about these two ways in which time can be a constraint on action. And why do you think that the important temporal structure of agency is internal not external?

RA: This contrast comes up in the introduction to Time and the Philosophy of Action, which I co-edited with Michael Sigrist. In planning the collection, we were thinking about the fact that action is an essentially temporal phenomenon. There are lots of ways in which action is temporal, but here are a few we might consider: actions necessarily take time to complete. They have to begin, develop, and end. They also involve all sorts of mental processes such as intentions or desires and agency involves reasoning and planning. To complicate things, all of these processes might involve different time-scales—for example, habitual action might be rhythmic and repetitive, but I can intend a novel action that incorporates habitual ones. These sorts of temporal considerations seem to characterize action essentially, in the sense that to be an agent is to undergo these temporal processes in a way that’s different from the way many other sorts of things—tables or salt crystals, for example—exist in time. This is more or less what we meant by talking about the temporality of action, or about action as characterized by internal constraints: something that doesn’t have an internal structure that necessarily includes time in at least some of these ways can’t be an action.

Michael suggested contrasting this with external constraints. One way to explain this is to look at something like a rubber ducky: we can describe all the structures essential to a rubber ducky without taking time into account. Rubber duckies, of course, are involved in complex moving relations with other bodies, especially in water, and these relations involve time, but the temporal considerations needed to make sense of these motions are external to the ducky itself. Nothing in the metaphysical structure of a rubber ducky requires that it move in one way rather than another, and if the universe were suddenly frozen at a single moment in time, rubber duckies would still exist. Actions, on the other hand, would not. Aside from internal constraints, of course, actions also have external ones. For example, thinking about the time available to me in life should rationally affect what I do now, and even more simply, the fact that I need to go to sleep in the next few hours means that I should not start a task that will take half a day if I need to finish it in one sitting. These kinds of constraints are fundamentally different from the sorts of internal constraints I mentioned earlier, although I suspect that in the case of actions, unlike duckies, what makes them responsive to external temporal constraints is their internal temporal structure. For example, the end of the semester places constraints on what I can do now, at the beginning of the semester, and I can respond to these constraints because my planning and acting now are already temporally structured.

3:AM: Why do you say that rational agency requires a notion of tensed reality rather than tenseless and does this make us kind of ontologically weird given that the rest of ontology isn’t full of rational agency?

RA: Philosophy of time is centered on a distinction between two ways of understanding how reality is distributed in time. According to the A-theory, reality is essentially tensed: there is a past, a present, and a future. According to the B-theory, reality is tenseless, so when I say that I am going to finish this interview in the future, the fact about reality this captures is that the event of my finishing this interview occurs later than the event of my saying that I am going to finish it. So on the B-theory events are temporally related to each other by virtue of occupying different points in time rather than by virtue of happening in the past or future relative to some moving present; time does not really pass, but only seems to. A-theorists have frequently embraced our experience as the guarantor of their view: we experience time as passing, they may claim, and the B-theory cannot account for that. For example, I may look forward to the summer, or I may be relieved to come home after a long day, and these attitudes only appear to make sense relative to the present. B-theorists have responded by arguing that while our experiences and beliefs may be tensed in this way, the underlying reality they are about is not. My relief may contain the belief that the long day is in the past, but the truth-maker for this belief is that the long day’s conclusion is prior to the occurrence of the relief.
But there are elements of our experience that can’t be reduced to truth-makers in this way. Consider a theory according to which actions are caused by desires or intentions. My intention may contain the belief that the action is in the future, and the truth-maker for that belief may be tenseless. But the intention also represents the action as future. It’s the futurity of the action that makes the intention into an intention. A desire, also, represents its object as future, and it does so independently of any propositional component it may have; an attitude that does not aim at some future object simply isn’t a desire, or at least not a desire capable of causing an action. We cannot reduce this relation of futurity to a belief that can be explained by a tenseless truth-maker. Our agency certainly makes us different from much of the universe, but whether it makes us ontologically weird depends on whether we adopt the B-theory in explaining entities unlike us. Alternatively, we might think that tenseless temporal relations are grounded in our agency, insofar as time is our way of being in the world and we utilize different temporal representations to interact with different aspects of it.

3:AM: Why do some philosophers think we need basic actions to explain action, and why do you think agents whose only actions are basic actions would not be agents of the sort humans typically are?

RA: Basic actions are typically taken to be actions that we can do directly, without doing anything else first. In other words, they are taken to be the building blocks of all actions. There are a number of different thoughts that lead to positing basic actions, but two in particular stand out for me. First, in performing a complex action, I must perform any number of subordinate actions as a means. But the subordinate actions can’t themselves all be made up of further actions, since if that were the case then performing the action would require us to first perform an infinite number of actions in a finite amount of time. So actions have to bottom-out in some set of actions that cannot be further divided, at least not without ceasing to be actions. Second, since actions take time, they can be interrupted. If I attempt to draw my gun but am shot before my hand even reaches the holster, I’ve still done something: I reached for my holster. But what if I am shot a split second before that? Or a split second before that? At any point, it still seems like I’ve done something, that is, I’ve performed some action that is more basic than my originally intended action. These two lines of thought both seem to point to the idea that there must be basic actions, but they are in tension with each other. The first suggests that basic actions must take time, since they are components of complex actions that do. But the second suggests the opposite: any action that takes time can be interrupted, and thus must have components that are more basic than the whole. An interesting way to try to resolve the problem (suggested by the authors in the first part of our collection) is to not think of actions as composed of smaller actions, but instead to think of them as processes or exercises of capacities that become actions only through the form the agent imposes on them via intention.

I think human actions are not typically one-offs. Actions are somehow picked out from ongoing activity, and how we pick them out as actions depends on our purposes in picking them out. Moreover, many of our actions are embedded in wider projects. That wouldn’t be true of agents who only perform basic actions, since their actions couldn’t aim at anything beyond themselves. So agents whose only actions are basic actions, or atomic agents, as Kieran Setiya calls them, would be very different from most of us. Their actions wouldn’t be expressions of projects, and they couldn’t, I suspect, be parts of ongoing activity.

3:AM: What is Kavka’s ‘Toxin Puzzle’ out to show, and how does it figure in helping us understand the issues of intention, will, agency, action and so forth?

RA: A billionaire has offered to give you a million dollars if, at midnight tonight, you intend to drink a toxin tomorrow afternoon. The toxin will make you sick, but it won’t kill you, so you wouldn’t mind drinking it for a million dollars. But there is one catch: the money will be deposited in your account (or not) before noon. So here is the problem: you have no reason to drink the toxin (and the billionaire has told you as much), since he isn’t paying you to do so. By the time tomorrow afternoon arrives, you will either be a millionaire or you won’t be, but you will have no reason to drink the toxin and a strong reason not to. Since you know this, it seems that forming an intention to drink the toxin will be difficult or even impossible without finding some way to trick yourself into drinking. Kavka suggests that this shows that our intentions, like our beliefs, are not fully volitional; if you could simply intend to drink the toxin at will, in the way you can simply drink the toxin at will, there would be no difficulty. It seems we cannot, or cannot easily, intend to do something we have no reason to do. Why? Some have suggested that it is simply impossible to intend for reasons different from reasons to act. Others have argued that we cannot intend to do something that, assuming ourselves to be rational, we know we won’t do. Michael Bratman has intriguingly suggested a “no-regret” condition on intending—I cannot intend something if I expect to regret carrying through on the action—and Ted Hinchman has built on this idea in interesting ways, by suggesting that agential authority requires me, in intending, not to intend in ways that my projected future self will regret having trusted my acting self to have followed through on. We might also consider the possibility that intending and acting are not two separate volitional acts: the action is a fulfillment of the intention, and intending to do something I won’t do is much like starting an action I know I can’t finish.

3:AM: You’ve argued for a Heideggerian view of agency to ensure that freedom and responsibility are irreducible features of agency. So first can you sketch the problem of free will and agency that your Heideggerian approach is out to dissolve?

RA: Action theory in general and free will in particular are both booming these days. But while there is a wide range of diversity in approaches, the range of views that are taken seriously is actually pretty narrow. One problem, I think, is the widespread acceptance of physicalism, the view that everything that exists can ultimately be explained in physical terms open to investigation by the empirical sciences. That can warp our thinking somewhat. It’s not that I think physicalism is wrong, but it can be a problematic frame for these questions. One way to be a physicalist is to start by saying, “there are all these interesting phenomena that we’ve described, and maybe one day we’ll figure out how to explain them in physical terms or reconcile them with what we will know.” But what happens frequently is something along the lines of, “we know this problem has to be explicable in physical terms, so let’s work it out in terms of elements we can potentially match up to something physical, like things going on in the brain.” Starting with that approach places unnecessary constraints on what counts as an acceptable philosophical view, especially given that we don’t know exactly what the physical is, yet we’re trying to adjust our concepts to our limited knowledge of it. It’s the main reason, for example, that action theory spends so much time on thinking of actions as events caused by something happening in the agent, which is thought to correspond to physical events in the brain. Meanwhile, views of free will are grouped into event-causal, agent-causal, and non-causal. Now, maybe everyone working on free will and action theory should have some view on how actions are caused, if at all, but I don’t see why that’s necessary—not everyone working on moral responsibility needs to have a view on the appropriateness of corporal punishment, for example—and I think it unhelpfully constrains the range of acceptable theories if one’s answer to that one question is expected to be the defining feature of their view. The same, I think, goes for the problem of whether free will is compatible with determinism; it’s virtually impossible to work on free will without having a well-defended position on that question. But we should stop to ask whether it’s all that conducive to intellectual experimentation to frame questions about free will and agency through such a limited number of issues. The other problem is that you pretty much have to be endorsed by a highly ranked program with an established leader in the field to land a job that leaves you much time to do research. That means again that, while everyone has an incentive to set their view apart, there’s pressure to work within established paradigms; the rest too often get thrown into years of contingent labor.

In any case, that isn’t to say that there isn’t plenty of exciting work. The most interesting, to me, builds a theory of freedom into a theory of agency, rather than thinking of freedom as an added extra that action theory can do without. This approach, which we see in some great recent work like Helen Steward’s, echoes much of the historical work from Sartre going all the way back to Augustine. If a libertarian adopts a theory of agency that may function just as well with or without free will, in my view something’s already gone wrong. So the problem, for me, is not one of how we can have freedom in the natural world, but rather one of what agency itself must be such that it allows us to be oriented by norms and responsible in light of them. This approach has to preserve a genuine role for choice, which I think Humean views can only bring into the picture by cheating. But at the same time the choice has to be oriented toward norms, since otherwise we might as well not have it.

[Heidegger, Axelos, Lacan, Jean Beaufret (recipient of the Letter on Humanism), Elfriede Heidegger, Sylvia Bataille (by this time married to Lacan)]

3:AM: So how does the Heideggerian view, and its view of time, help solve the problem and deliver the required goods of freedom and responsibility?

RA: There are three points from Being and Time that I find especially helpful. The first is the concept of projection into possibilities. To say that something is a possibility for me isn’t to say that it’s something I could end up doing, but to say that it’s something I am doing. What makes one a clown, for example, is one’s acting comically, making balloon animals, and occasionally taking a pie to the face. Following Bernard Williams, contemporary philosophers often discuss a similar idea under the fuzzy notion of projects. But the same is true of much simpler things, which we might not normally think of as projects, like being a slob, which may involve leaving clothes all over the floor, not cleaning up after eating, etc. A number of philosophers have usefully drawn a connection between Heidegger’s view here and Christine Korsgaard’s notion of “practical identity,” which is a self-conception guided by what one values; that is, how I see myself guides what sorts of things I find valuable, and the sorts of things I find valuable, in turn, guide how I see myself.

Second, in his discussion of guilt, Heidegger notes that we do not create our practical identities from scratch, but neither are these identities simply external constraints on our agency. Rather, we experience them as something we have always already chosen. This might seem illusory, since the possibilities we find ourselves projecting into seem to have been in place before we could even become aware of them; in what sense could we have always already chosen them? There is a constant interplay between our possibilities and our actions. Our possibilities shape our self-understanding, but at the same time are shaped by our actions, so there is at least one perfectly clear sense in which my projects are things I have always already chosen: I maintain them by continuing to press into them. So far, of course, there isn’t much room for freedom within this conception of agency: if we’re just stuck with our projects, these shape our exercise of our agency, and that exercise of agency in turn maintains our projects, it looks like we are stuck in a closed loop that isn’t really up to us, but only has the appearance of being so.

That brings me to the third point, and here I am going to continue to oversimplify. In his discussion of death Heidegger presents the powerful metaphor of a ripening fruit. He notes there that the fruit’s unripeness belongs to the being of the fruit. Its ripeness, in other words, isn’t just a possible future state to which the currently unripe fruit is “indifferent”, but instead characterizes the fruit’s being in the present, and this remains true regardless of whether or not the fruit in fact manages to ripen. The fruit’s temporality is such that its present is defined by its future. Likewise with us. Since our possibilities are never actual, in the sense that we are continually pressing forward into them, our current practical identity depends in an important way on our future identity. But unlike fruit, we don’t have a specified future, such as ripeness. In a sense, as long as you have a future, that future is open. What you are and have been depends on what you will become.

By itself, this account of agency isn’t enough to give us freedom. It doesn’t prove that we are free or show how we might be. But it does provide an opening in what otherwise seems like a closed loop. Personally, I tend to favor a Kantian move here, which I borrow from constitutivism. If there are norms constitutive of our agency, then it’s possible for us to press forward into our possibilities in a way guided by those norms. But if our current possibilities veer in a direction far from one dictated by those norms, doesn’t that mean we are not, in fact, able to pursue the norms? This is where the temporality Heidegger attributes to agency comes in handy: since my past and present depend on my future, it’s possible for me to bootstrap myself in the right direction by shaping my past in such a way that the trajectory of my projects leads to the right place.

3:AM: How do temporal loops give us greater control over our lives, greater than even John Fischer thinks?

RA: Fischer is developing a point from David Velleman: that the shape of our lives matters because our lives have a narrative structure, such that the meaning of past events depends on the meaning of future events. We redeem past misfortunes by acting in ways that give them a positive meaning in the context of our lives. In my view, these loops can do a fair bit more than simply change the meaning of past events. I think many, if not all, of our mental states are pretty indeterminate. What, for example, is my current attitude toward heath bar coffee crunch ice cream? Desire? Hope? Wishing? The only way to tell is to see what I do next. That’s not because our mental states are black boxes that we can’t get inside, but just because part of what gives them form is what follows from them. On this picture, future actions can give determinacy to the past motives and states of character they stem from. So in this sense, temporal loops don’t just give us a semblance of control over the meanings of our pasts, but over those psychological pasts themselves.

3:AM: You argue that the narrative structure of a life has the potential to open up a stronger conception of freedom than just guidance control. So first what is meant by guidance control and narrative structure here?

RA: Fischer distinguishes between guidance control and regulative control. Regulative control is the kind of control we exercise through the actual sequence leading up to action. Regulative control, on the other hand, requires us to make a “relevant difference to the world”, as Fischer puts it; it requires us to not merely exercise control via the actual sequence, but to have some ability to instigate a different sequence. On Fischer’s semi-compatibilism, determinism may rule out regulative control, but guidance control is good enough for moral responsibility.

Appeals to narrative are widespread, and yet there are relatively few discussions of just what exactly is involved in narrative structure; some of the best analysis appears in Galen Strawson, who is deeply skeptical of the role of narrative, and David Velleman, whose account is pretty far from standard. But one common way of understanding narrative structure, which I suggested earlier, is just that a narrative connects events in such a way that the meaning of earlier events is fixed by later events. How broadly we interpret “meaning” here is important. We might be thinking of meaning only in terms of significance: my putting on the shoes with wooden soles becomes significant when I slip on the ice and dislocate my pride. But events can also have properties that depend on future events: whether my action of putting on chapstick is the action of protecting my lips from shredding by cold, for example, depends on whether I go out into lip-shredding cold after applying it. As Arthur Danto notes, events can also have narrative properties that depend on whether or not later interpreters group them as parts of larger events. For example, the Stooges’ self-titled debut could only become a punk album once critics (and musicians) identified the phenomenon.

3:AM: And what is the stronger conception of freedom this gives if it’s not libertarian control?

RA: I tend to think that libertarian control requires two features. First, the agent must have genuine power over alternative possibilities. If I would have acted otherwise only if something else in the universe that isn’t up to me were different, that doesn’t count. Second, the agent must originate the action or be, in some sense, causa sui, in Galen Strawson’s term. My retroactive view of agency seems to give us both of these things. For us to make sense of actions, we have to see them as following from the agent’s motives. On one hand, this means that we use our knowledge of the motives to interpret the action, but this works both ways: motives have to make sense in light of actions, as well as in light of interpretations of the actions. Mental states are indeterminate, and motives gain their determinacy in part through interpretations we offer in light of the agent’s action. And it’s a bit trickier than that, even, since interpretations of actions can change in light of later actions and commitments on the part of the agent. So there is a sense in which we are causa sui: who we have been as agents depends on the commitments we make. And there is a sense in which we have alternative possibilities: although what I am doing at any moment follows from my motives, what my motives are in turn depends on what I will do and how they will fit into a retroactive interpretation. In articulating these two conditions of libertarian agency, however, I’ve focused on conditions that are internal to agency as marking a libertarian account. But it’s standard to take an external condition as definitive of libertarianism: whether the account of free will posited is compatible with determinism. A compatibilist could in principle adopt the account I’ve been giving, so it’s more like a semi-libertarian view. Everyone tells me the view is actually compatibilist; I’m fine with that. But my sense is that the conceptions of agency libertarians rely on are, typically, compatibilist ones, with indeterminacy introduced into the picture somewhere. I’m trying to develop a conception of agency designed for libertarian use.

3:AM: Is Heidegger a narrativist about selves?

RA: That’s a very popular view among Heideggerians, but I don’t think it’s quite right. Narrativists have drifted away from understanding the self in terms of a trajectory toward some conclusion and toward a conception of narrative on which each element takes on meaning in light of the whole. That strikes me as perfectly consistent with Heidegger’s view, and like Heidegger, the narrative view aims to provide a view of diachronic identity unified by its telos, in some sense. But selves are a lot messier than narratives. Any narrative will inevitably emphasize some features at the expense of leaving others out, because that’s what happens in any articulated account. To make the account coherent, you have to put the pieces together in a certain way, and some of the pieces have to be left out or the story won’t make sense. The bits that are left out, though, are still elements of our selves. Heidegger’s view, whether successfully or not, aims to incorporate all aspects of the self into one, and I think that’s something narrative just can’t do. That said, I think narrative gives us a much more promising picture of practical identity (Marya Schechtman, Kim Atkins, Anthony Rudd, and Patrick Stokes have done some very interesting work here). A narrative, as I said, can’t capture every aspect of the self, which is a conflicted mess. A narrative has to distill and simplify that mess, getting it to both make sense internally and cohere with external influences and situations. But that’s exactly what we need if we are going to be able to act, since action requires unity for the duration of the action, at the least, and unity among actions within projects at the most. So I think there is a nice way to combine narrative identity with a Heideggerian view without reading the former into the latter.

3:AM: Does a narrative theory of identity help preserve a neo-Lockean psychological continuity theory of the self – and if it does does that mean that such a view of the self is stronger than many think these days?

RA: Some proponents of narrative identity see it as countering neo-Lockean views, but I tend to see it as an outgrowth and elaboration of those views. It can preserve psychological continuity while maintaining that there is a unity to that continuity that is missing from other kinds of causal chains. The idea is roughly this: we tend to think that if event x causes event y which causes event z, event x must be determinate; it’s already happened, it brought about y by virtue of its particular causal powers, and thus those causal powers are fixed. A narrative view allows a bit more fluidity into the causal chain retroactively, since y and z will be involved in fixing what the causal properties of x were. This doesn’t mean that there isn’t a causal chain in play, however. It just means that when we are dealing with indeterminate elements which have interpretation-dependent identity conditions, we have a causal chain that cannot be understood without taking meaning into account. I’m not sure whether this makes neo-Lockean views stronger, but it does seem to allow a wider range of views to fall under that umbrella label. In some ways, this is a Davidsonian move: if we describe the self in purely reductionist terms, we are describing the elements the self consists in, but are doing so under a description that leaves out self-pertaining elements. The narrative description, on the other hand, lends is that selfy quality.

3:AM: You’ve engaged with Sam Scheffler’s views about the afterlife. Before explaining why you take issue with some of what he argues, could you sketch for us the salient aspects of his approach?

RA: To avoid misunderstanding, what Scheffler means by “the afterlife” isn’t what that word ordinarily means, but rather the continued existence of other human beings after our own deaths. Scheffler convincingly argues that we care a great deal about the afterlife in this sense by crafting a simple thought experiment and tracing out its implications. The thought experiment is the Doomsday Scenario: you know that you will live out a normal life span, but the Earth will be destroyed by an asteroid thirty days after your death. Scheffler asks which of our projects and activities would lose their meaning for us in this scenario, and it turns out that almost all of them would, beyond simple hedonistic ones. We couldn’t commit to multi-generational projects, for example, and continuing to maintain traditions would seem pointless. Creative work might lose much of its appeal, since there would be no future audience for it. And Scheffler even suggests, intriguingly, that our relationships might be altered in far-reaching ways, since relationships derive their value from our sense of a good life, and this sense might be radically altered in the Doomsday Scenario. His book came out with several critical essays arguing that many of our projects would remain intact, but I side with Scheffler on this. I think many of us see our identities as dependent on an ongoing cycle of generational change, and Scheffler nicely brings out this dimension, arguing that confidence in the afterlife lies in the background of much of our valuing.

3:AM: So why do you think this position is faulty? Why do you think some people care about the afterlife?

RA: Scheffler doesn’t just argue that the afterlife matters much more to us than we might normally think. He argues that its importance shows that we care more about the afterlife than we do about our own survival. I’m not convinced. First, I think his conclusions don’t follow from his premises. For example, he notes that the Doomsday Scenario would present a serious crisis for us, forcing abandonment of many of our values. Our own mortality, on the other hand, does no such thing, so it follows that we care more about the afterlife than about personal survival. But that doesn’t follow at all, since the two cases aren’t analogous. The Doomsday Scenario would be catastrophic because it would involve a radical change in our beliefs about the world. Nothing comparable happens in the case of our own mortality. We don’t have catastrophic reactions to the discovery of our mortality because it isn’t something we discover; we’re born with it. If you took a race of immortals and made them mortal, I’m willing to bet that would lead to some catastrophic reactions. You could just as well argue for the opposite conclusion: our reaction to our mortality doesn’t look catastrophic only because all of human civilization is an adaptation to that catastrophic reality. We engage in multi-generational projects, creative acts, and traditions precisely as a response to our mortality. So I’m not convinced by the claim that we value the afterlife more than our own survival; I suspect we value it because we value our own survival. People talk about achieving immortality through their work, their children, their art, and so on. I think that talk might actually be right, in a sense, but regardless, many people certainly seem to take it literally. There is a sense in which our own deaths don’t seem quite so final so long as we know others will carry on after us, and Scheffler does a great job of bringing this out. But maybe that’s why the Doomsday Scenario seems so catastrophic: it really does make our deaths final.

3:AM: Is Bernard Williams right in his approach to attitudes towards immortality? Why is your defence of his position less than fully fledged – is it to do with the role of his notion of character?

RA: I have a lot of respect for Williams’s view. The idea that immortality is undesirable is a perennial theme in fiction, but defending it on philosophical grounds is a lot trickier. Most of the responses are fairly predictable: there are lots of things that seem like they would be worth doing over and over forever, so long as we take breaks; maybe we could constantly find new ways to reinvent ourselves. Sure. And really, who wants to die if you never have to get old? But I think there is a way to at least make Williams’s view sound reasonable. He argues that if you lived long enough, your character would change so much that eventually you would be utterly different from your present self. And, from my perspective now, it’s not clear why I have any more reason to desire the existence of that complete future stranger than the continuing existence of any other person. The standard response to Williams is that he overlooks the fact that the change would be gradual. Most of us undergo changes in character over the course of our short lives, and this doesn’t make our lives less worth living. If I were changed into a completely different person overnight, I might find that prospect equivalent to death. But if I change into that person over the course of sixty years, it seems much less problematic. I’m not sure that’s entirely right. Mostly, we don’t have very good ways of predicting just how we’ll turn out. But if someone told me that over the next forty years I will gradually transition into a Nazi sympathizer who derives deep satisfaction from a daily routine of torturing children, I wouldn’t find the fact that this transition will be gradual very reassuring. Heck, even if someone told me that I’ll get sick of philosophy and spend my days reading tabloid magazines, that seems less objectionable but also not especially appealing, from my current perspective. Of course it’s true that, after I have finished transitioning into that person, I might look back and think all is good. But from my current perspective, the prospect of such drastic change in character seems unappealing. So to some extent the desirability of immortality depends on whether I take a forward-looking or a backward-looking perspective.

Second, I’m not sure it makes sense for me to have a desire for immortality, as opposed to simply longevity. Some people, inspired by Williams, think that immortal beings couldn’t have projects and so would find their lives empty and pointless; they couldn’t have a character at all, since having a character depends on having projects. I don’t think there is a conclusive case for that. Immortal beings, particularly ones who were born immortal, might come up with all sorts of projects they would find fulfilling, like documenting every change in the solar system around them. If we were to become immortal, on the other hand, we might not be able to find such projects, because we are used to pursuing projects that somehow fit into a finite life; our patience is pretty limited. But we’re adaptable, and perhaps we could adjust. What we would be adjusting to, however, would be a fundamentally different kind of life with fundamentally different kinds of values and fulfilling activities, and we just can’t imagine what that would be like. When we desire things, normally, we have at least some sense of what the desired object is like, but that wouldn’t be the case here. So if we can have a desire to be immortal, it’s fundamentally unlike any other desire we have. This isn’t a fully fledged defense of Williams because ultimately I think one could want to become a completely different kind of being and just hope that, looking back in a few millennia, my future self will think that all was good. But that’s mostly just guess-work, and if the best response to Williams rests on guess-work, then I think he has a pretty decent argument.

3:AM: Finally switching to another area of interest in your work, you’ve written about the problem with evil in Kant. Why is the view of evil problematic for Kant and why do you argue that diabolic evil does play a role in human behaviour?

RA: In Kant’s system every free action must have its ground in either nature or freedom. That means that if I am responsible for my actions, they cannot be caused by nature, but must issue from my freedom, and freedom is defined in terms of the ability to act from respect for the moral law. But Kant also takes it that human beings are by nature evil. We cannot be evil by nature in the ordinary sense, since that would exclude responsibility, so Kant argues that evil is freely chosen by all of us. This creates a further problem: every free action has to have a maxim, and every maxim must be grounded in something. Good maxims are grounded in the moral law, which is just the law given to us by Reason. But evil actions are grounded in an evil maxim, which for Kant is a maxim of occasionally allowing deviations from the moral law. What Kant rejects is the possibility that human evil could be diabolical, i.e., that it could be grounded in a maxim opposed to the moral law, rather than simply one that allows for exceptions. I didn’t actually disagree with him on that; I was only interested in trying to explain why it might appear as if diabolical evil is possible. The big problem for Kant, though, is that the evil maxim itself can’t be grounded in anything, which means Kant has to make its ground inscrutable. He even proposes that scripture has a “spirit” tempt Eve into eating the apple in order to convey just this point: the ground of evil has a nature that necessarily evades human comprehension. That means that Kant’s moral system is a restricted economy: in order to define freedom and morality reciprocally, he has to take evil out of circulation. The late Laszlo Tengelyi, who generously spent months working through the Religion with me in Germany, insisted that I was completely wrong here; it was the only time I had ever seen him angry. He was probably right.

3:AM: And finally for the curious readers here at 3:AM, can you recommend five books that will help us understanding your philosophical world better?

Korsgaard, Self-Constitution
Together with R. Jay Wallace’s Normativity and the Will and Richard Moran’s Authority and Estrangement, this book represents to me a way to do philosophy of action that captures some of what is central to existentialism. Watson’s Agency and Responsibility and Velleman’s The Possibility of Practical Reason are less existentially tinged, but to my mind belong in a similar category. But then you have to read Nomi Arpaly’s Unprincipled Virtue to counter the overrationalization of our agency.

Kant, Religion Within the Limits of Reason Alone
Everyone reads the Groundwork at some point in their philosophical career and usually walks away with a pretty truncated understanding of Kant’s view of freedom. Once you get through the second Critique, that view is somewhat improved, but then the Religion throws it all into chaos again. Kant is usually seen as grounding freedom in an atemporal noumenal world, but something else is happening here, since all of us have already made the choice of an evil maxim, and yet the possibility of acquiring a holy will remains open to us. Dieter Henrich’s The Unity of Reason is a nice but overlooked contribution to understanding the relation between freedom and morality. As an added bonus, the Religion presents an account of religion so rationalized it might be acceptable to atheists; a feat reminiscent of Maimonides’s Guide.

Heidegger, Being and Time

This, of course, has to be on the list. But it’s helpful to read together with Sartre’s Being and Nothingness, which to my mind overturns Heidegger as much as it draws on him, and
Beauvoir’s Ethics of Ambiguity, which tries to make good on an existentialist ethic. Bergson’s Time and Free Will is an important and sadly overlooked precursor to the views on the relation of freedom and temporality (Mead’s Philosophy of the Present tries something similar). And as long as you’re reading existential philosophy, Jonathan Webber’s forthcoming Rethinking Existentialism is a valuable guide. Mishima’s The Temple of the Golden Pavilion is my favorite existentialist novel. It doesn’t get as much attention from philosophers as Sartre, Beauvoir, Camus, and Murdoch, since Mishima isn’t laying out philosophical theses via fiction, but it’s got all the elements of good existentialist literature and, frankly, who hasn’t wanted to overcome their mortality by destroying a thing of eternal beauty?

Donald Davidson, Essays on Actions and Events
This is a classic, of course. In many ways, his later Problems of Rationality works out some of the problems in more interesting ways. Unfortunately, in my view, the part of Davidson that action theory latched on to was the causal theory of action rather than the antireductionist psychology. It’s helpful to counter that tendency with a healthy dose of the papers in Constantine Sandis’s edited collection, New Essays on the Explanation of Action. And I’m still hoping a very different consensus on action theory might begin to emerge from Thompson’s Life and Action and McDowell’s recent work on action.

Georges Bataille, Visions of Excess
In this vein, I’m also fond of Baudrillard’s Simulations and Artaud’s The Theater and its Double. Are these books good philosophy? Are they even philosophy? Should we take them seriously? I’m not sure I can answer any of these questions, but these books add a zany weirdness far beyond the strangest thought experiments, and I hope philosophers can make room for them as an exercise in imagination on the margins of the discipline.

Richard Marshall is still biding his time.

First published in 3:AM Magazine: Tuesday, March 14th, 2017.