Introduction
If you are like most people, you prefer negative experiences to be in the past and positive experiences to be in the future. Furthermore, you may prefer more painful experiences in the past to fewer painful experiences in the future, and fewer pleasurable experiences in the future to more in the past. If this describes you, you most likely have bias towards the future, or future bias.
I open this article with a closer look at the manifestation of future bias in a thought experiment. It is followed by a consistency argument, which, while not definitive, lends support to the notion of time neutrality. The core of the article questions future bias through the lens of a reductionist personal identity. The traditional view sees a person as one unchanging entity moving through time; in contrast, I propose that distinct consciousness-moments (or time-slices) better capture the essence of a person. This reductionist view, as I will argue, is incompatible with differential treatment of past versus future events. Finally, I discuss recent empirical findings on future bias, including the conditions under which it is more pronounced and potential explanations for it.
Motivating Example
A well-known thought experiment by Derek Parfit illustrates future bias (Parfit, 1984). Imagine a patient lying in a hospital bed. A nurse approaches and explains that she has mixed up reports of two patients and does not know which one describes you. The reports read:
(i) You had a painful, ten-hour-long operation yesterday. You were conscious throughout because anaesthetics were not an option, but the operation was a success. A drug administered afterwards erased all your memories of it.
(ii) You will have a painful, one-hour-long operation tomorrow. You will be conscious throughout because anaesthetics are not an option, but the operation will succeed. A drug administered afterwards will erase all your memories of it.
Which patient do you prefer to be?
It turns out most people would prefer the painful experience to be in the past, even if it lasted much longer. Some justify this by saying, “What’s done is done,” or “The past doesn’t matter because it’s not 'me' anymore.” I think that in most circumstances, this is very likely true: you cannot change bygone events, but you can shape your future experiences to some degree, so that is where you should concentrate all your efforts. Therefore, the reasoning would lead you to opt for the option with the least future suffering, which is option (i).
Before challenging the standard response, I want to address a few pitfalls related to this particular thought experiment.
Beware of Thought Experiment Pitfalls
One of the issues in the original scenario is that if I have completely impartial preferences, I'm indifferent between being one patient or the other. I know one of us experienced the ten-hour-long surgery yesterday, while the other is due to have the one-hour-long surgery tomorrow; it does not matter who is who, all else being equal. This could be easily fixed by constructing a scenario in which only one patient must decide whether they prefer the operation to be in the past or in the future.
One confounder that may explain people's future bias is that they already know the operation was a success, whereas they cannot be sure the future operation will be equally successful. There is less perceived certainty in the outcome, even though the scenario guarantees that the effects of the operation will be identical (Lee et al., 2020).
Another pitfall is that when we usually think of an operation, we associate some kind of urgency with it. The sooner the beneficial outcome is delivered, the better for us. This would again favour the operation in the past. There are, however, plenty of non-life-threatening operations for which a delay of one or two days makes little difference.
Yet another pitfall people may fall into is considering the affective state of the patient: past pain makes one feel relieved, whereas future pain makes one apprehensive (Lee et al., 2020). Ideally, the scenario should not account for these experiences if we are trying to isolate temporal preferences, as it's the case with future bias. (The drug-induced amnesia serves a similar purpose – to avoid possible memory-related confounders influencing one's decision – although it complicates the scenario.)
The original future scenario could be modified by stipulating that immediately after the patient's preference is stated, they are put to sleep until the scheduled time of the operation tomorrow. This ensures, at least in theory, that affective experiences caused by anticipation are eliminated
Inbuilt Inconsistency in Future Biased Decisions
Consider a similar scenario: do you prefer to have a ten-hour-long surgery tomorrow or a one-hour-long surgery in three days? (There are no other patients, you will be given post-operative amnesia just as before, and other mentioned pitfalls are implicitly assumed to be overcome.) There can be no doubt that you would choose the shorter operation, even if you were not near-biased. That is, even if we abstract away from people's tendency to prefer negative experiences further in the future, the opportunity to reduce future suffering tenfold is too compelling to ignore.
Alex thinks the same. Alex is not a typical patient, though. They have been abducted and are at the mercy of Drs. Predictors, who can predict the future with certainty. Nevertheless, Alex is asked the question with which this section began and contends that they prefer the shorter operation. Moreover, if given the hypothetical choice, Alex would be willing to pay $1.00 to be rescheduled from the longer operation to the shorter one, but they are unaware of the Predictors' decision procedure for operation allocation.
Now, Alex wakes up two days later, having no memories of enduring, or not enduring, any operation. Alex is asked the same question, except that now the longer operation would have already happened yesterday, whereas the shorter one is still due tomorrow. Alex, being biased towards the future, prefers negative experiences to be in the past – so much so that they would be willing to pay $1.00 to already have had the operation. The Predictors subject patients to the longer operation if and only if they predict that the patient would be future biased. Since the Predictors knew in advance that Alex was future biased, Alex suffered through the ten-hour-long operation yesterday.
What this means, in sum, is that Alex's preferences are inconsistent: Alex is willing to pay $1.00 to switch from the longer operation to the shorter one before any operation takes place, and is equally willing to pay $1.00 to switch from the shorter operation in the future to the longer operation in the past, even though they gained no new information in between.
Consciousness-Moments at the Center of Personal Identity
There is a different line of reasoning that supports time neutrality and it is related to the ethics of personal identity. The standard conception of personal identity is that a person moves through time, where the past "me" no longer exists but where the (near) future "me" is very real. (This thinking is not absolute, as we will see in the next section, because people's tendency to prefer to have good lives overall is also present.)
If a person is subjected to a painful event, we usually say it's bad for that person. It's not very clear what it's meant by this, however. On the view of continuous personal identity, why would some past or future event be good or bad for me now, except for instrumental reasons? I lack access to the phenomenal experience, and hence can't derive any value or disvalue from it.
In contrast, according to the reductionist view, the intrinsic value exists only at the moment when it's directly experienced. The moral badness of a painful event stems from the fact that at each particular moment in time, the person-instance would prefer the pain to stop over its continuation. It is these consciousness-moments – infinitesimally brief instances of conscious experience – that intrinsically matter. A single person, as usually conceived of, is comprised of countless of these conscious entities.
This view also better encapsulates the relativity of time. After all, what we take to be the past or future was, or will be, at one time experienced as the very presence for a particular consciousness-moment. It's then, strictly speaking, incoherent to talk about "past" and "future" experiences. Consciousness-moments may lie in our relative past but they were just as real and "present" as the one reading this sentence.
According to many ethical views, the broad ethical principle is to reduce unnecessary suffering. Combined with the alternative view of personal identity, the aim stated in normative terms would be something like this: At each temporal location, there should be the fewest number of suffering consciousness-moments possible, because it is each valenced experience that intrinsically matters to the unique conscious entity having it at that time. Although memories and anticipation can cause negative experiences, the original phenomenal experience will never be reenacted in its purest form, as it belongs to a different consciousness-moment.
Returning back to the bias towards the future, notice that when we prefer more painful experiences to be in the past rather than fewer in the future, it just means that we prefer the existence of more suffering consciousness-moments overall rather than fewer of them.
Therefore, future bias, in a sense, condemns more consciousness-moments in our relative past to needless suffering than it spares consciousness-moments in our relative future. Future bias weights two equally disvaluable experiences differently based on their location in time, but this is not right because the moral badness of suffering always exists in some present moment and as such is independent of its relative temporal location.
Clarification
To clarify, I am not claiming there is no practical reason to focus on the future, but that there is a priori no reason to prefer "your" future experiences over "your" past experiences. Naturally, in most circumstances, the suffering of past consciousness-moments cannot be alleviated – the disvaluable experiences have already occurred – so there is a good reason to focus on preventing suffering consciousness-moments that are yet to arise and over which we have some control.
In the realm of normative ethics, however, where preferences and judgements are detached from practical intuitions and considerations, I see no rational basis why decisions should be influenced by future bias. It's unlikely that such philosophical theorizing has any immediate implications but the observable coherence of the consciousness-moments view in such hypotheticals still lends general support for this view as opposed to others.
Empirical Findings
Factors Influencing Future Bias
It has long been tacitly assumed that some kind of future bias is rational, though serious attempts at providing plausible explanations have been emerging only in recent years. The body of literature is still growing, but there are early hints as to what the explanations of future bias are and which situations make people more prone to biased evaluations. That being said, various experiments have been conducted, but the results do not always match, which might be caused by slightly different experimental designs and a lack of data.
First, it has been tested if future bias emerges only in tiebreaker situations, but is absent in unequal situations; that is, if people prefer to have some equal amount of pain in the past rather than in the future, but prefer to have less pain in the future rather than more in the past. Lee et al. (2020) found that people's preferences for past pains and future pleasures are far from absolute as participants were quick to switch once the amount of value became unequal. However, there is still substantial residual future bias even in unequal conditions (also Latham et al., 2021).
Moreover, future bias seems more pronounced in negative than in positive experiences (e. g. Greene et al., 2021). Greene et al. (2020) found the ratio to be at least ten units of past pain for one unit of future pain, whereas the ratio for positive experiences hovers around two to one (Greene et al., 2022a).
The reason why participants in Lee et al. (2020) switched sooner might be that they were made to think in terms of "who they would prefer to be", which encourages broader life evaluation and time neutrality. On the other hand, participants in Greene et al. (2021) were asked whether they prefer some event to be past or future, which encourages them "to take a more embedded temporal perspective". (Greene et al., 2022a)
In any case, future bias does not seem to be just a matter of preference in tiebreaker conditions but is persistent even in unequal situations. On the other hand, once the ratio of past vs future experiences reaches a certain threshold, people's future bias tendency diminishes.
Second, it was investigated whether people's preferences differ between first- and third-person conditions. It was assumed that third-person decisions are more distanced and rational. It appears that there is only a slight reduction in future bias when deciding about others rather than ourselves, especially when "one is prompted to take on the perspective of a particular person" (Greene et al., 2020; Lee et al., 2020). Time neutrality was found, however, when deciding about a random person's past or future work compensation (Caruso et al., 2008). A recent study by Caruso et al. (2024) found that people are equally future biased in self, friend and stranger conditions.
Third, both hedonic events and non-hedonic elicit future biased responses, though non-hedonic experiences seem to push people a bit toward time neutrality. Greene et al.(2020). It, however, depends on the valence of the event and the person's perspective. Out of these three, negative valence is the strongest determinant of future bias.
Causes of Future Bias
Latham et al. (2021), instead of testing out which conditions make people express future bias, set out to investigate what explains future bias in the first place. They came up with the "practical irrelevance" hypothesis, which stipulates that people are biased towards the future because the past is usually causally inaccessible, whereas the future is amenable to change. The first version of the practical irrelevance hypothesis – the cognitively mediated hypothesis – predicts that if the past is made causally accessible, people will exercise their choice and exhibit less future bias. The second version predicts narrow evolutionary adaptation without any bias reduction in such situations: the adaptation is taken to be hard-wired and inflexible.
In Latham et al. (2021), future bias was attenuated in situations in which participants' choices had retro-causal consequences compared to preference statements of causally inaccessible past events; this lends support to the cognitively mediated hypothesis. It does not, however, tell the whole story because there is still residual future bias left unexplained.
It is consistent with the findings (Greene et al., 2022b) that suggest participants are less future biased when "they think of themselves as ‘placing’ events in the past/future" because it encourages "agentive perspective" (as quoted in Greene et al. 2022a).
Aside from the reliance on the practical heuristic, why do people fail to show impartiality in these scenarios? My hunch is that people consider their future consciousness-moments to be closely tied to their present self, whereas those in the past are perceived as belonging to a stranger. Given that egoism is the prevailing norm in nature and society, it could partly explain why most of us are future biased: we want to help those consciousness-moments with whom we identify the most. The asymmetric identification is, however, just an illusion, because both past and future consciousness-moments are equally inaccessible to us at this very instance.
Conclusion
I hope to have demonstrated that the seemingly intuitive judgements supporting the rationality of future bias fall apart when examined through a reductionist lens of personal identity. This view—a compelling candidate for a metaphysically coherent understanding of personhood—shifts the focus from a continuous, unchanging self to the fleeting yet intrinsic significance of individual consciousness-moments. By granting equal moral weight to each of these moments, irrespective of their placement in time, the preference for time neutrality emerges not merely as an abstract ideal but as a more ethically consistent and impartial perspective.
While I have highlighted evidence that future bias is pervasive among people, I acknowledge that its practical implications in everyday life remain uncertain. However, the true significance of this discussion extends beyond the critique of future bias—it lies in advancing an ethically coherent view of personal identity, one that can inform broader philosophical and ethical considerations moving forward.
Acknowledgements
I'd like to thank Pavel Cahyna and Martin Vlach for valuable comments on this draft, which have helped me refine my thoughts on this matter.
References
Caruso, E. M., Gilbert, D. T., & Wilson, T. D. (2008). A Wrinkle in Time: Asymmetric Valuation of Past and Future Events. Psychological Science, 19(8), 796-801.
Caruso, E. M., Latham, A. J., & Miller, K. (2024). Is future bias just a manifestation of the temporal value asymmetry? Philosophical Psychology, 1–40.
Greene, P., Latham, A. J., Miller, K., & Norton, J. (2020). Hedonic and non-hedonic bias towards the future. Australasian Journal of Philosophy. , 99(1), 148–163
Greene P., Latham A. J., Miller K., & Norton J. (2021). On Preferring that Overall, Things are Worse: Future-Bias and Unequal Payoffs. Philosophy and Phenomenological Research, 105: 181–194.
Greene, P., Latham, A. J., Miller, K., Norton, J., Tarsney, C., & Tierney, H. (2022a). Bias towards the future. Philosophy Compass, e12859.
Greene, P., Latham, A. J., Miller, K., & Norton, J. (2022b). Why are people so darn past biased? In C. Hoerl, T. McCormack, & A. Fernandes (Eds.), Temporal Asymmetries in Philosophy and Psychology. Oxford University Press.
Herrán M, (2018). Empty, Open and Closed Individualism. Manu Herrán (blog)
Kolak, D. (2004). I am you: The metaphysical foundations for global ethics. Dordrecht: Springer.
Latham, A. J., Miller, K. & Norton, J. (2021). Future bias in action: does the past matter more when you can affect it?. Synthese 198, 11327–11349
Lee, R., Hoerl, C., Burns, P., Fernandes, A.S., O'Connor, P.A. and McCormack, T. (2020). Pain in the Past and Pleasure in the Future: The Development of Past–Future Preferences for Hedonic Goods. Cogn Sci, 44: e12887.
Parfit, D. (1984). Reasons and Persons. Oxford, GB: Oxford University Press.
Roberts, M. A. (2024). "The Nonidentity Problem", The Stanford Encyclopedia of Philosophy (Fall 2024 Edition), Edward N. Zalta & Uri Nodelman (eds.).
Vinding, M. (2017). You are them. Ratio Ethica
