Von Hippel & Trivers (VH&T) argue that self-deception evolved not as a defense mechanism but as an offensive weapon in an evolutionary arms race of deception and deception-detection. Their proposal explains how the deceiver can also be the deceived and why evolution may have possibly selected for mechanisms that represent the world inaccurately. This is thus a powerful perspective that, as illustrated by the latter part of the target article, sheds new light on a range of phenomena. Here I will suggest that this approach might also offer a way out of the vexing problem of systematic biases in human foresight. First, however, I note that the authors fail to discuss when such an arms race may have got off the ground. They thus sidestep an important aspect of the evolution of self-deception.

Humans are clearly not the only creatures to deceive. Many animal signals are not honest but were selected to deceive predators, prey, or competitors. This deception is not limited to simple mimicry but includes various curious behaviors and even counterdeception. Some primates, especially our closest living relatives, appear to engage in quite flexible “tactical” forms of deception (Whiten & Byrne Reference Whiten and Byrne1988). Do they hence have the prerequisites for the purported arms race between deception and deception detection?

One important aspect of deception that nonhuman animals may not be capable of is lying. Parts of the target article seem to use the terms deception and lying interchangeably, perhaps reserving the latter to describe verbal deception. Yet to lie, one really must do more than declare something that is in fact not true. One must also know that it is not true; otherwise mistakes would be called lies. Furthermore, one must want the other to believe what one knows not to be true, to be true. Thus, lying implies intentionally implanting a false belief. In spite of persistent efforts, there is as yet no convincing evidence that non-human animals can represent false beliefs (Krachun et al. Reference Krachun, Carpenter, Call and Tomasello2009; Penn et al. Reference Penn, Holyoak and Povinelli2008). If this is correct, then they lack the very opportunity to deliberately manipulate such representations. Lies, and the self-deceptions they may have spawned, appear to have evolved only over the past five million years, after the split from the last common ancestor with chimpanzees.

A second purportedly unique skill that may have played an important role in the evolution of human deception and self-deception is mental time travel (Suddendorf & Corballis Reference Suddendorf and Corballis1997). Humans can flexibly imagine a range of potential future episodes (and hence can plan complex deceptive ploys) and can also mentally reconstruct past events (and hence can uncover past deceptive ploys). These travels in both temporal directions are closely linked in mind and brain (e.g., Addis et al. Reference Addis, Wong and Schacter2007; Suddendorf & Corballis Reference Suddendorf and Corballis2007). From Bartlett (Reference Bartlett1932) we know that recollecting past events is an active construction that may be biased. Selection for memory must be based on what fitness benefits it brings, not on how accurate it is per se. The same must be true for thinking about future events. Foresight is implied in various contexts in the target article (e.g., optimism, plans, and goal achievement), but was not addressed directly. Yet, I think (possibly because I deceive myself about the importance of something I have been working on for too long), that VH&T's theory can throw new light on the evolution of biases in foresight.

Our ability to imagine future scenarios has obvious adaptive benefits, allowing us to prepare in the present to secure future rewards or thwart future disaster. Why, then, is it that humans display systematic errors in anticipation? For example, various lines of research (see Gilbert Reference Gilbert2006) have demonstrated that we tend to exaggerate the positive or negative emotional consequences of future events (e.g., of handing in one's PhD thesis; or of losing a leg). When the event occurs, we tend to feel not quite as happy as we had imagined, and we tend to cope much better with a negative event than we anticipated. We also systematically misjudge the likelihood of events. VH&T allude, for example, to the optimism bias where we generally tend to judge the likelihood of good things happening to our future self above that what is rational.

On the face of it, there are some clear benefits to these biases. One apparent benefit of exaggerating the hedonic value of future consequences, for instance, is that it may help motivate future directed action. One can only fully reap the benefits of anticipating future events if such thought can appropriately guide present action. The system that governs motivation, however, has long been based on present rewards. Evolution had to modulate this system rather than build a new one from scratch. An important problem foresight poses, then, is the need to motivate prudent action in the present when this is costly, or when it prevents more immediate hedonic rewards. To compete with current rewards and motivate future-directed action, it may thus make sense to exaggerate the future reward or punishment.

The logical problem, though, is that one would expect the system to learn with experience (and modulate decision making). People should learn to adjust their predictions and create a more accurate representation of future hedonic values. On some level, perhaps, we do appreciate the truth (the German vernacular, e.g., tells us that “Vorfreude ist die schönste Freude” [anticipated joy is the greatest joy]). Yet, we continue to display the same forecasting biases.

VH&T's theory explains how a system might have evolved that mis-represents the facts by the clever proposal of a social arms race between deception and deception detection. This approach could also offer a solution here. With language, a third purportedly uniquely human capacity, humans exchange their plans and coordinate them. Indeed, language may have evolved initially for the sharing of mental time travels (Suddendorf et al. Reference Suddendorf, Addis and Corballis2009). In order to elicit cooperation on a project, one may benefit from exaggerating the likelihood and positive consequences of success (or negative consequences of failure). As in the argument of VH&T, I propose that one may be much better at doing this if one actually believes this exaggeration. Again, such belief may also reduce potential punishment if the future fails to bring what was promised. Thus, our biases in judging future hedonic values and likelihoods might be self-deceptions that have their origin in an evolutionary arms race between deception and deception detection mechanisms of a social other. VH&T's theory may potentially go a long way further in explaining the evolution of the complexities of the human mind than even they have anticipated.