In his first Kant Lecture, “Against Temporal Neutrality,” last Wednesday, Sam Scheffler argued that a bias toward the future is not in fact irrational, and that the wider framework of bias versus neutrality yields a distorted view of the requirements of rationality. (I was able to ask Scheffler about this later, so I can promise that my exegesis is at least good enough not to be worth objecting to in Q&A. I’ll discuss his response to my questions at the end of this post.)

Part of the story he told is that adopting a temporally neutral perspective would make our lives worse, because it would make it harder to develop the strong attachments that make our lives valuable. (Scheffler offers this as part of his larger “attachment-sensitive conception” of what it is to lead a life.) A good life is one that contains strong, particular attachments to people and to sustained pursuits (“projects”). Scheffler claims that these attachments would be more difficult to establish or sustain without an essentially forward-looking orientation in time.

Intuitively it seems plausible that a temporally neutral or “timeless” perspective is more “detached” and makes us less prone to attachments. But it is worth asking, by what mechanism does it do so? The answer Scheffler proposed in his talk seems to me to prove a bit too much. It seems to show either that a timeless agent will fail to do what is best even from a timeless perspective, or that a dreamlike life of dwelling on past pleasures is, from a timeless perspective, better than an active life. Neither conclusion seems right to me.

Future Bias

Parfit’s My Past or Future Operations neatly exemplifies future bias (my paraphrase):

I wake up in the hospital. One of the following is the case: (1) I am about to be subjected to a long, necessary surgery without anesthesia, after which I will be given a drug that makes me forget the experience; (2) I have just been subjected to the surgery and taken the forgetting drug.

Intuitively, (1) is a lot worse than (2). I would hate to learn I was in (1), and be relieved to learn I was in (2). Indeed, I would rather have a 10-hour surgery in my past than a 1-hour surgery in my future.

Parfit criticizes this reaction: a pain is equally bad whenever it occurs. Since, whether it is past or future, there is no way to avoid the pain, the only reasons the future pain could be worse would be either its absolute position in time, or its relation to the present moment. But (says Parfit) neither of those things matter in themselves.

Timeless’s attachments

Parfit imagines a character called Timeless who lacks this future bias.

When he is told that he will later have some period of great enjoyment, he is pleased to learn this. When he is reminded that he once had just such a period, he is equally pleased. (Reasons and Persons, 174)

Timeless would have certain advantages over us future-biased beings: he could selectively look back on good episodes from the past and enjoy them as though in anticipation.

However, Scheffler argues, his life would be worse than ours, because he would be less able to form valuable attachments. Scheffler illustrated this with an example which went roughly like this:

Last weekend, Timeless had a really good time hanging out with an acquaintance. Next weekend, Timeless could (by putting in some effort) arrange to have a really good time hanging out with that same acquaintance again.

The problem is this: since Timeless has already hung out with his acquaintance last weekend, and can mentally revisit that pleasure at will, why should he put in the effort to hang out with the acquaintance again next weekend? A future-biased person has an additional reason to put in the effort: to such a person, past pleasures are less good, so putting in the effort to hang out next weekend is the only way to achieve a pleasure of that intensity. But valuable attachments are made by repeated efforts. The future-biased person will be more motivated to undertake such efforts, and so (says Scheffler) will end up with a better life, one richer with valuable attachments.

One aside here: there is nothing particular to attachments about Scheffler’s argument. It’s a general argument about the role of pleasures and pains in our instrumental motivations. Future bias makes future pleasures and pains relatively more motivating for us, the argument says, which makes us take more future-oriented actions than we would without it. The future-biased take more actions to build attachments because they are more motivated towards future-oriented actions in general (not because of anything particular to attachments). So if the argument is right, it suggests that Timeless would not only have trouble forming valuable attachments, but would deviate widely from intuitively correct patterns of instrumental reasoning. For example, Timeless would judge there was vastly diminishing reason to ever do the same thing more than once; it would almost always be better to reminisce or to do something new.

A timeless dilemma

I am not ready to accept Scheffler’s argument. It forces us to one of two interpretations of Timeless’s behavior, both of which I regard as implausible. They are:

  1. A life of hanging out once and reminiscing is timelessly better than a life of hanging out twice; or
  2. Timeless is not motivated to choose the life that is timelessly better.

One life is “timelessly better” than another if it is better from a temporally neutral perspective.1 Timeless, in making his weekend plans, is choosing between two lives: (1) the life containing a single episode of hanging out followed by reminiscing about it; (2) the life containing two episodes of hanging out, together with the efforts needed to make that happen. Prima facie, (2) is better than (1): it is a fuller and richer life. And, I claim, (2) is timelessly better than (1). Judging that (2) is better than (1) does not require a future bias. (2) would be better than (1) even from a purely backward-looking perspective. Someone on their deathbed could look back on their life and be glad that their life was (2) rather than (1), or regret the opposite. More generally, a life with close attachments to others is (it seems to me) timelessly better than one without them.

If we judge that hanging out is timelessly better than just reminiscing, and Scheffler is right about Timeless, then we are forced to conclude that Timeless is not motivated to choose the life that is timelessly better. If this is correct, Timeless is seemingly irrational by his own lights.2 There is an akrasia built into Timeless’s agency: a systematic failure to be motivated towards what he regards as better. This is a very interesting prospect; it would be good to know if timeless agency was in some way oxymoronic. But I don’t think we should be convinced yet. In particular, if Timeless suffers from some akrasia, we would like to know whether that akrasia is essential to being a timeless agent, or instead emerges from some quirk of how we are thinking about Timeless.

Is it better to hang out than to reminisce?

I want to return to the question of whether it is timelessly better to hang out twice, or to hang out once and reminisce. Although I suggested that hanging out was timelessly better than reminiscing, perhaps I was not sufficiently inhabiting the timeless perspective when I made that judgment.

Here is a way that reminiscing could be better than hanging out. Perhaps (from a timeless perspective) reminiscing about an episode of hanging out is just as good as actually hanging out. We might be forced into this position by the following thought: the only value-relevant difference between hanging out and reminiscing about a past episode of hanging out is the reminiscer’s temporal relation to the episode of hanging out. If this is the case, then the two lives are equally good from a timeless perspective, and if we judge that they are not actually equally good, we must concede the timeless perspective is inappropriate.

However, I think this is wrong, because it undercounts the timeless value of the second episode of hanging out: it considers only the contemporaneous pleasure of it. If the first episode of hanging out is valuable not just as a contemporaneous pleasure but also as an anticipated and reminisced pleasure, then the second episode of hanging out would be too. While, in life (1), Timeless would be able to enjoy reminiscing about having hung out once, in life (2), Timeless would be able to enjoy reminiscing about having hung out twice. This would be more enjoyable, for example because with two episodes of hanging out, the reminiscing could include thinking about the beginnings of a closer relationship. The conclusion that reminiscing is timelessly better than hanging out entails that reminiscing about a single episode of hanging out is timelessly just as good as reminiscing about two episodes of hanging out, but we should reject that implication.3

Can you inhabit a moment forever?

Scheffler seems to suggest that Timeless, by virtue of his temporal neutrality, can use his relation to past moments of pleasure as a substitute for ever doing something else. Parfit seems to encourage this interpretation when he explains how our lives would be better without future bias.

When we look backward, we could afford to be selective. We ought to remember some of the bad events in our lives, when this would help us to avoid repetitions. But we could allow ourselves to forget most of the bad things that have happened, while preserving by rehearsing all of our memories of the good things. It would be bad for us if we were so selective when we are looking forward. Unless we think of all the bad things that are at all likely to happen, we lose our chance of preventing them. Since we ought not to be selective when looking forward, but could afford to be when looking backward, the latter would be, on the whole, more enjoyable. (174-5)

We could be selective in our reminiscence, and preferentially remember good episodes. This would make our lives better because a life with more pleasant and less unpleasant reminiscences in it is pro tanto better. Scheffler seizes upon this contribution timelessness would make to our experiential quality of life and goes a step further. Because selectively enjoying the past would be much easier than acting towards the future, suggests Scheffler, Timeless would focus his energies on enjoying the past, to the detriment of acting towards the future. He would choose not to hang out because he has access to reminiscences that are just as good as the real thing.

There is something tricky about this. On this interpretation of timeless agency, all the events of our life potentially contribute to its value twice: first as actual occurrences, and second as mental re- (or pre-)enactments. Scheffler seems to further think that, for Timeless, reminiscing about a past pleasure is comparable or equal in value to experiencing a contemporaneous pleasure. I argued above that, even granting this, hanging out could be better than reminiscing because reminiscing about two episodes of hanging out could be better than reminiscing about one episode of hanging out. But now I want to cast doubt on the whole idea of evaluating a life by adding up the values both of the events in it and the reminiscences about those events. This seems to me a suspect form of double counting.

We seem to be forced into this double-counting by denying that contemporaneous pleasures are special. Perhaps this is a feature of timelessness, but I think that conclusion should be resisted. Contemporaneous pleasures are special not (only) because they are contemporaneous but also because they are direct experiences of an actual event. By contrast, a reminiscence is like a copy of a copy, and an anticipation is an experience of an imagined event. Living in reminiscences is a bit like living in the experience machine. A life in the experience machine is filled with pleasant experiences, but empty of actual events in the world. A life in reminiscences is filled with pleasant experiences, but almost empty of actual events in the world. But, whatever might be wrong with a life in the experience machine, it doesn’t seem to have anything to do with temporal bias. If that’s right, and we have reason to reject a life in the experience machine, then we have reason (not rooted in temporal bias) to regard the life of reminiscence as worse than the life of actual events.

Instead, I think we should conclude that hanging out makes Timeless’s life better, and more hanging out makes it even better – all from the timeless perspective, all without regard to his temporal position relative to the hanging out. It may still be the case that temporal neutrality would undermine our valuable attachments, but a further explanation is needed.

Coda: Scheffler’s reply

After I wrote most of this, we had a discussion seminar with Scheffler and I got to ask him about it. He agreed with my characterization of the argument (although agreeing in a Q&A doesn’t guarantee complete accuracy), and he also agreed with both of my two suggestions:

  1. Timeless could judge that a life with close relationships is better than one without.
  2. Timeless could judge that hanging out is better than reminiscing, because when he hangs out he then has more to reminisce about.

Both of these could motivate Timeless to hang out. The second suggestion is a bit of a strange motivation and we might think it’s not completely compatible with a normal friendship. But the motivation in the first suggestion seems pretty normal. So it seems like Timeless should be able to form close attachments.

I was puzzled by this, but I think the right way to understand what Scheffler was up to is to put him more in context of Parfit. Scheffler pointed out that Parfit’s diagnosis of future bias is actually a bit narrow. Parfit thinks that we don’t suffer from future bias in most of our concerns – for the most part, we exhibit a (laudable) temporal neutrality. Future bias comes out only in particular kinds of cases: most clearly in our attitudes towards experiences that are pleasant or unpleasant in themselves (hence My Past and Future Operations). Let’s call these pure pleasures or pains. When Parfit creates Timeless, the character without future bias, he is careful to change just those attitudes that are affected by the future bias. Timeless is just like us, except in place of our attitudes which exhibit future bias, he has non-future-biased counterpart attitudes.

However, as Scheffler points out, many of our pleasures and pains have an intentional character: they are good or bad partly because of their relation to events in the world, rather than purely in themselves. Among these “event-dependent” pleasures and pains are the ones relating to friendship: hanging out is both pleasurable per se and because of its role in constituting a relationship with another person.

What Scheffler aims to show is that Timeless’s attitudes towards pure pleasures and pains extend awkwardly to the event-dependent pleasures involved in a relationship. Even if that is right, though, I don’t think advocates of temporal neutrality are necessarily committed to such an extension. Instead, with Parfit, they can argue that we are already appropriately temporally neutral in regard to the goods and bads in our life that are not pure pleasures or pains. Because of this, I don’t think this part of Scheffler’s argument against temporal neutrality cuts much ice. It is an argument against a particular conception of temporal neutrality that advocates need not adopt.4

However, Scheffler has convinced me that this material is quite tricky, and even Parfit’s treatment contains unexplored conceptual difficulties. I would not be surprised if a more convincing version of this basic argument could be (or has been!) worked out.

  1. It is tricky to say what exactly it means to be better from a temporally neutral perspective. I believe the following is a sufficient (but not necessary) condition: $A$ is timelessly better than $B$ if it is better without regard to the temporal position of the events of the life relative to the temporal context from which the life is evaluated. 

  2. Perhaps there is room to deny that Timeless will judge better the life that is timelessly better. But if Timeless is not an agent who regards as better the life that is timelessly better, then opponents of temporal neutrality also owe us a critique of the latter agent. 

  3. When Scheffler argues that Timeless is less motivated towards future hanging out, because of the mental accessibility of a past episode of hanging out, is that because Timeless finds “reminiscing” on the past episode just as good as anticipating it? That is how I will present the argument, and I think it is close enough for these purposes. However, Parfit explicitly warns against this interpretation. The backward-looking counterpart to anticipating a pleasure or pain is not memory, because we don’t have memories of our future pleasures and pains. Accordingly, Timeless should be able to have the backward-looking counterpart attitude even if we don’t have any memories of the targeted pleasure or pain. The counterpart attitude can occur purely as one “to past pains and pleasures about which we know, but of which we do not have painful or pleasant memories” (Reasons and Persons, 172). If anything, substituting the counterpart attitude for “reminiscing” in the argument I make here would strengthen the argument. If it is plausible that reminiscing about two episodes of hanging out can be better than reminiscing about one, it is all the more plausible that knowing you have hung out twice can be better than knowing you have hung out once. 

  4. Of course, this is not the only argument Scheffler presented against temporal neutrality. He thinks that, even if being temporally neutral wouldn’t make our lives worse, we are wrong to see temporal neutrality as a requirement of rationality.