![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
It has widely been assumed, by philosophers, that our first-person preferences regarding pleasurable and painful experiences exhibit a bias toward the future (positive and negative hedonic future-bias), and that our preferences regarding non-hedonic events (both positive and negative) exhibit no such bias (non-hedonic time-neutrality). Further, it has been assumed that our third-person preferences are always time-neutral. Some have attempted to use these (presumed) differential patterns of future-bias—different across kinds of events and perspectives—to argue for the irrationality of hedonic future-bias. This paper experimentally tests these descriptive hypotheses. While, as predicted, we found first-person hedonic future-bias, we did not find that participants were time-neutral in all other conditions. Hence, the presumed asymmetry of hedonic/non-hedonic and first/third-person preferences cannot be used to argue for the irrationality of future-bias, since no such asymmetries exist. Instead, we develop a more fine-grained approach, according to which three factors—positive/negative valence, first/third-person, and hedonic/non-hedonic—each independently influence, but do not determine, whether an event is treated in a future-biased or a time-neutral way. We discuss the upshots of these results for the debate over the rationality of future-bias.
Disclosure Statement
No potential conflict of interest was reported by the authors.
Notes
1 For a more formal characterisation of hedonic time biases, see Greene and Sullivan [Citation2015: 948–9].
2 Explicit supporters of the rationality of hedonic future bias include Prior [Citation1959], Hare [Citation2007, Citation2008], and Heathwood [Citation2008]. Recent critics of hedonic future bias include Brink [Citation2011], Greene and Sullivan [Citation2015], and Dougherty [Citation2015]. Hedden [Citation2015] argues that hedonic future bias is merely rationally permissible, while Parfit [Citation1984] remains neutral about its rationality.
3 Indeed, ‘past-bias’ is not a term that appears in the literature on time bias. However, the concept is clear: it is simply the inverse of future-bias.
4 Some empirical evidence for this asymmetry can be found in research by Caruso, Gilbert, and Wilson [Citation2008: Study 4].
5 Hare [Citation2008] argues that the predicted asymmetry between first- and third-person preferences over the temporal location of hedonic events requires the other person to be spatially distant. When the other person is spatially close, Hare predicts, hedonic future-bias returns. He takes this to be consistent with the view that hedonic future-bias is rational. See note 15 for more discussion of this hypothesis.
6 Evolutionary accounts of future-bias focusing on the role of emotions include those by Horwich [Citation1987: 196–8], Maclaurin and Dyke [Citation2002], Suhler and Callender [Citation2012], and Greene and Sullivan [Citation2015: sec. V]. Historical precedent for these views can be seen in the work of Hume [Citation1738: sec. 2.3.7.6] and Adam Smith [Citation1759: pt. 6].
7 There were no significant main effects of age and gender; nor were there any significant interaction effects involving age and gender.
8 Not only do our third-person vignettes allow for a direct comparison to the first-person vignettes; they also avoid a potential problem with Parfit’s ill-mother thought experiment (introduced in section 1)—namely, that his mother suffering in the past entails that she had less time alive overall, which provides a time-neutral reason to prefer her suffering to be in the future.
9 There was no significant main effect of statement type, nor were there any significant interaction effects with statement type. Statement type has no influence on our reported results. As such, it is permissible to amalgamate the two statement types in the manner that we have done.
10 Participants were quite confident in their judgments regarding their preferences. Results of a 2×2×2 between-subjects ANOVA found significant main effects of perspective F(1, 805) = 18.188, p < .001 and event F(1, 805) = 15.054, p < .001. No other significant main effects or interaction effects were observed. The main effect of perspective showed that levels of confidence were significantly higher for first-person conditions (M = 5.61, SD = 1.43) than for third-person conditions (M = 5.16, SD = 1.60). The main effect of event showed that levels of confidence were significantly higher for hedonic conditions (M = 5.86, SD = 1.59) than for non-hedonic conditions (M = 5.18, SD = 1.43).
11 The ANOVA assumes (1) that residuals are normally distributed and (2) a homogeneity of variance. We first tested these assumptions by using a Shapiro-Wilk test (W = .973, p < .001) and Levene’s test (F = 2.741, p = .008). According to these tests, both assumptions have been violated. However, on large samples (n = 813) the ANOVA is robust to non-normality, and visual inspection of the normal Q-Q plot shows that the residuals in this study do approximate a normal distribution. Further, an investigation into the homogeneity of variance violation revealed that it was being driven solely by the lower variance in condition (1). The ANOVA is robust to non-homogeneity when the variance ratio between the largest variance and lowest variance conditions is less than 1.5 [Blanca et al. Citation2018], and in this study the ratio is 1.46 (between condition 1 and condition 3).
12 Whenever you perform multiple statistical tests on the same data, there is an increased chance of encountering Type I errors (false positives). The Bonferroni correction helps to correct this problem.
13 There is one limitation to the study that we should note. We have assumed that time-neutral participants will respond with 4, and that those who are time-biased in some manner (past or future) will respond with 1–3 or 5–7. It is, however, possible that a small number of strongly time-neutral participants will respond in a ‘deviant’ way, by responding to statements of the form ‘I would prefer to learn that [the event will occur] tomorrow, and [did not occur] yesterday’ with 1–3, in order to indicate strongly that they have no time-biased preferences. Our methodology will classify these participants as either past-biased or future-biased. Importantly, even if there are such participants, this will not affect the mean levels of agreement that we see, since we would expect that they will be evenly allocated to each statement type in a given condition, and that they will respond in this manner regardless of the question that they see. For each participant misclassified as future-biased, another will be misclassified as past-biased, and this effect will cancel itself out. The results of the ANOVA, then, would be unaffected by any such effect. Nevertheless, if there is a small population of participants who respond in this ‘deviant’ manner, it will lead us to overestimate slightly the percentage of people who are past-biased and future-biased. While we have no reason to suppose that there are participants who responded in this manner, the reader should nonetheless keep this in mind in what follows. A useful follow-up study would replicate this study by using a single sliding scale from ‘I would prefer to learn that [the event will occur] tomorrow, and [did not occur] yesterday’ to ‘I would prefer to learn that [the event occurred] yesterday, and [will not occur] tomorrow.’ Then this type of deviant response would not be possible. Thanks to an anonymous referee for pointing this out.
14 With thanks to an anonymous referee for this suggestion.
15 Recall that Hare [Citation2008] maintains that people will be time-neutral if the third person is spatially isolated from the chooser (see note 5). However, the case of Freddie the astronaut is extremely spatially isolated, and so Hare’s view would predict that participants will be time-neutral in their preferences regarding Freddie.
16 There is a difference between the cases that we discuss by Parfit [Citation1984] and Brink [Citation2011] and those by Hare [Citation2008] and Greene and Sullivan [Citation2015] that might be relevant to the debate over future-bias. Parfit’s and Brink’s cases compare situations in which the same negative event is either past or future, while Hare’s and Greene and Sullivan’s cases compare events involving more past pain against ones involving less future pain. Hare and Greene and Sullivan predict that people will prefer more past pain in the first-person condition and less future pain in the third-person condition. Our experiments are like Parfit’s and Brink’s cases: they compare the same negative events being past or future. It is possible for our results and Hare and Greene and Sullivan’s predictions to be mutually consistent: it could be that, when the negative events are the same, people use the fact that one of them is past as a ‘tiebreaker.’ It would be interesting for future work to compare unequal trade-offs between past and future, like those discussed by Hare and Greene and Sullivan. Notably, Parfit’s [Citation1984: 165] influential My Past or Future Operations case in support of negative hedonic future-bias also features an unequal trade-off.
17 More carefully, perspective only seems to matter when the events are hedonic.
18 We thank David Braddon-Mitchell, Mark Colyvan, Michael Duncan, Brian Hedden, Alex Holcombe, and Arden Koehler for helpful comments on an earlier draft.