top of page

Preferring cake tomorrow, to cake yesterday

Recent research from the Centre for Time at the University of Sydney undermines a widely held assumption that our preferences about the temporal location of our own experiences will be biased in a certain kind of way, but that our preferences about the temporal location of other people’s experiences will not be biased in that way. It suggests that preferring to have your tasty cake tomorrow, rather than yesterday, may not be irrational after all.

Suppose you are asked whether you prefer that you had very tasty cake yesterday, or that you will have very tasty cake tomorrow. You get the same amount of cake no matter when in time it is located. Holding everything else fixed—for instance, holding fixed that the cake tomorrow will definitely arrive, and that you won’t, for instance, be incredibly ill and unable to eat it and so on—which do you prefer? Or are you indifferent between these two options (you have no preference). If you have a preference, do you think that preference is rational, or do you think you really ought not care when you get the tasty cake? Now ask yourself the same question again, but this time about a painful dental procedure.

If you prefer to have your tasty cake tomorrow rather than yesterday, and to have had your painful dental procedure yesterday rather than tomorrow, then your preferences are what are known as future biased.

Some researchers argue that you ought to be indifferent in these cases. After all, you get the same amount of tasty cake, (or the same mount of pain via the dental procedure), whether you had it yesterday or will have it tomorrow. They argue that it would be be irrational to prefer tomorrow’s cake to yesterday’s cake, (or the other way around). By contrast, others argue it is at least rationally permissible to prefer to have your cake tomorrow rather than yesterday, and to prefer to have your painful dental procedure yesterday rather than tomorrow.

Until relatively recently future bias had not been experimentally studied. So we didn’t know what people’s preferences look like, and we had no real basis on which to evaluate their rationality.

The first study to investigate these kinds of preferences was in 2008. Caruso, Gilbert, and Wilson asked participants to determine fair compensation for boring data entry work that either occurred in the past or will occur in the future. The amount of boring work was the same regardless of where in time it was located (past or future).

Participants assigned themselves 60% more compensation for the work if it was located in the future, compared to in the past. However, when asked to recommend compensation for the work of others they recommended the same compensation regardless of when the work occurred. This suggested that people prefer to have negative experiences (like boring data entry) located in their past rather than their future, and hence they required more compensation for the same work located in the future. This study however, did not look at people’s preferences regarding positive events (like eating cake), and it didn’t explicitly ask people what they prefer: it only allows us to infer their preferences based on the amount of compensation they would award themselves, or others, for the work. Still, it suggested that there is an asymmetry between our first-person preferences (our preferences about the temporal location of events that we will experience) and our third-person preferences (our preferences about the temporal location of events that someone else will experience).

The idea that there would be such an asymmetry can be traced back at least to Parfit [1984: 181] who presents the following case:

I am an exile from some country, where I have left my widowed mother. Though I am deeply concerned about her, I very seldom get news. I have known for some time that she is fatally ill, and cannot live long. I am now told something new. My mother’s illness has become very painful, in a way that drugs cannot relieve. For the next few months, before she dies, she faces a terrible ordeal. That she will soon die I already knew. But I am deeply distressed to learn of the suffering that she must endure.

A day later I am told that I had been partly misinformed. The facts were right, but not the timing. My mother did have many months of suffering, but she is now dead.

Parfit reports that in this case he would have no preference regarding the temporal location of his mother’s suffering. In contrast, he would care a great deal about whether his pain is in the past or future.

On the basis of this predicted asymmetry, some researchers such as Brink (2011), Greene and Sullivan (2015), and Dougherty (2015) argued that future bias is irrational. Something goes awry in the first-person case—perhaps the fact that we can anticipate the various experiences—and this leads us to make an error.

Recent research by Greene, Latham, Miller and Norton (2020) at the Centre for Time at the University of Sydney set out to test the idea that there is an asymmetry in future bias between first-personal and third-personal cases. The study took 931 participants and presented them with a vignette that describes a space voyage during which an astronaut (who they are told is themselves) wakes up and, briefly, cannot remember whether he/she had a rare tasty meal yesterday, or will have the rare tasty meal tomorrow, or wakes up and cannot remember whether he/she had a rare horrible meal yesterday, or will have the rare horrible meal tomorrow. The former participants are asked whether they (as the astronaut) would prefer to learn that they had the tasty meal tomorrow rather than yesterday, and the latter are asked whether they would prefer to learn that they had the horrible meal yesterday rather than tomorrow.

Other participants saw the same vignette, but rather than being told that they are the astronaut in the vignette, they are told that a third party, Freddie, is the astronaut. They are then asked whether they prefer that Freddie learns that he received the tasty/horrible meal tomorrow, or yesterday.

The study replicated the earlier results of Caruso et al, in showing that people prefer to learn that the horrible meal was had yesterday, rather than tomorrow. It also showed that people prefer to learn that the tasty meal will be had tomorrow, rather than yesterday. So in the first person condition in which participants were told that they were the astronaut, they showed future biased preferences.

But contrary to expectation, this study found the same pattern of preferences in the third-person condition. People preferred that Freddie have his tasty meal tomorrow and his horrible meal yesterday. This is in stark contrast to Parfit’s own predictions about his suffering mother, and appear to contradict the earlier findings of Caruso et al.

Greene et al hypothesized that the crucial difference between the two experiments is that in their experiment the vignettes described events surrounding some particular individual—Freddie—an astronaut on a special mission. The results suggested that participants were imaginatively putting themselves in Freddie’s position, and then answering the question about what they would prefer to learn, based on what they think things are like for Freddie. By contrast, in Caruso et al.’s study, participants are told to provide a dollar figure of compensation to work that has been, or will be, completed by a ‘randomly selected person from the local area’. No details about the person are provided. So, there is little prompting to take on the perspective of a particular person.

Greene et al note that these results undermine the widely accepted contention that there is a deep asymmetry between our first-person and third-person preferences. Many, and indeed perhaps most, ordinary situations are ones in which we can take the perspective of the relevant third party. It is only in special situations—for example, when a preference is elicited for a randomly selected stranger or someone who has already died—that perspective-taking is blocked.

This suggests that arguments that aim to show that future bias is irrational because we only show this bias when we are considering our own experiences, and not those of others, is mistaken. Of course, Greene et al are quick to point out that this doesn’t show that future bias is rational. But it does eliminate one reason why we might have doubted its rationality.

So if you prefer your tasty cake to be located tomorrow, rather than yesterday, it might still be that this preference is perfectly rationally permissible.

Brink, D. O. 2011. Prospects for Temporal Neutrality, in The Oxford Handbook of Philosophy of Time, ed. C. Callender, Oxford University Press: 353–81. 

Caruso, E., D.T. Gilbert, and T.D. Wilson 2008, A Wrinkle in Time: Asymmetric Valuation of Past and Future Events, Psychological Science 19/8): 796–801.

Dougherty, T. 2015. Future-Bias and Practical Reason, Philosophers’ Imprint 15/30: 1–16.

Greene, P., Latham, A. J., Miller, K., and Norton, J (2020) “Hedonic and non-hedonic bias towards the future”. The Australasian Journal of Philosophy. 

Greene, P. and M. Sullivan 2015. Against Time Bias, Ethics 125/5: 947–70.

Parfit, D. 1984. Reasons and Persons, Oxford University Press. 

26 views0 comments


bottom of page