17 Apr 2019
There are two ways to be fooled. One is to believe what isn’t true; the other is to refuse to believe what is true.
– Soren Kierkegaard
The visionary lies to himself, the liar only to others.
– Friedrich Nietzsche
It’s a well-known dictum that everybody lies. However, what is less known is the other side of the coin, the one that the quotes above expose – lying to ourselves. The ubiquity of self-enhancing biases, such as the Lake Wobegon effect or the self-serving bias, raises the question of why would we be so consistent and eager to deceive ourselves. To explore that question William von Hippel and Robert Trivers (hereafter referred to as VH&T) (2011a) put forward a valuable proposal to explain self-deception through an evolutionary lens. The purpose of this critical review is to put VH&T’s theory to the test: Does this theory make presuppositions that cannot withstand a philosophical and/or logical rebuttal?; Have the authors omitted critical aspects of self-deception that destabilise their theory and does their evolutionary approach meet the relevant evolutionary criteria?
The theory
VH&T argue that self-deception is an evolved form of interpersonal deception. It results from information-processing biases that favour welcome over unwelcome information, as long as the biases facilitate the individual’s goals.
The main evolutionary advantage of self-deception is two-fold: 1) the cognitive load of conscious deception is circumvented and 2) given that interpersonal deception is highly taxed in the evolutionary struggle (e.g. cheater-detection modules, Cosmides, 1989), self-deception allows to reduce the taxing should the deception of others be detected. Above and beyond the benefit of self-deception of convincing others of specific lies, there is a secondary advantage of self-deception: the ability to exude more confidence than warranted. Confidence is boosted through self-deceptive self-enhancement and self-deceptive other-derogation. The increased confidence, in turn, provides an array of psychological benefits that enable the individual to climb the social hierarchy, which is a target for natural selection.
However, at first glance, self-deception raises a paradox – how can a person deceive themselves and be deceived; in other words, how can the mind be both the deceived and the deceiver? VH&T offer several broad types of mental dissociations that restrict the access of the mental processes, that are subject to self-deception, to the mental processes, that enable the self-deception. For instance, we can hold the deceptive information in our explicit memory while the accurate information is stored in our implicit memory.
The next logical question is how incoming information is processed to be later sorted into the different types of mental processes. VH&T argue that a series of mental gymnastics facilitate the biased encoding of self-deceptive information. Such mental gymnastics start from the search of information and progress temporally to interpreting the found information in a biased way or even if we found and interpreted the information correctly, we might not be able to recall it. Even if we recall it, we can construct a different version of that information by rationalization. It can even go as far as convincing the self that the lie is true.
The paradox of dual representations
Ever since Trivers (1976/2006) brought the topic of self-deception under evolutionary lens, a philosophical paradox has scourged empirical advancements: how can our mind be the deceiver and the deceived simultaneously? VH&T resolve this paradox by suggesting that inserting an audience into the deceiver-deceived loop yields social advantages which can subsequently be selected for. However, using self-deception to explain overconfidence or optimism might be unwarranted as in those phenomena there is no third party to deceive (Pinker, 2011). Even more, self-deception requires two representations of reality, a biased one and an accurate one, but the advantages of overconfidence or optimism can be gained even if only one representation, the biased one, is present: thus, the theory of self-deception becomes superfluous.
Such criticism, if justified, renders VH&T’s proposal meaningless due to Occam’s razor: the explanatory value of the theory is zeroed out. Fortunately for VH&T, this does not seem to be the case. Dual representations are not necessary for self-deception to take place. For instance, in one study the participants stopped searching for information when the results were favourable and never had to encode a second representation to self-deceive (Ditto & Lopez, 1992).
Another similar criticism surrounding the dual representation paradox can be levelled at a more fundamental level. Mapping self-deception onto interpersonal deception assumes that one has the intention to self-deceive. Therefore, not only does the self have two competing representations of reality, but such mapping also requires that self-deceivers are aware of their self-deceptive intentions (Smith, 2011). To resolve this problem VH&T take a deflationist approach, which is to say that they argue that self-deception is the unintended product of biased information-processing. However, the issue is that they simultaneously advocate for an intentionalist approach, which is to say that self-deception is intentional because humans who self-deceive have been selected for as they are better interpersonal deceivers (Smith, 2011; for an overview of philosophical approaches to self-deception see Lynch, 2011).
However, the problem with such a line of criticism is that it mixes different levels of analysis. The purpose of self-deception (for interpersonal gains) is different from the mechanism of self-deception (via biased information-processing). Therefore, given that the purpose is intentional, it does not follow that the mechanism is intentional too. As VH&T (2011b) point out, people can intentionally self-deceive as it allows them to climb the social hierarchy, but this intention can be entirely unconscious.
Self-deception has costs
Any exhaustive, mature theory should account for all findings and questions within its scope. However, VH&T’s proposal is, at best, limited in its account of a critical aspect of self-deception: the costs. Such a blatant omission of half of the story can prove to be devastating to any theory, should the information and/or predictions prove to be inconsistent in light of new evidence. To their credit, VH&T do briefly touch on only two obvious costs: 1) the loss of quality of information when self-deceiving and 2) the possibility that ubiquitous self-deception might make the mind tolerant of holding contradictory information and hence vulnerable to exploitation by others. Still, there are many more costs and corner-cases to consider.
For example, although the original proposal argues that self-deception should be practised in moderation, it does not specify to what degree exactly. When a ceiling effect is not specified, it is hard to ascertain how much self-deception is too much. Answering this question is important as excessive self-deception can quickly undermine one’s own interpersonal trustworthiness in an environment where social bonds are of paramount importance (Bandura, 2011).
Another cost of self-deception is that we cannot apologize if we do not know we have lied (Mercier, 2011). Given the importance of apologizing in maintaining interpersonal trust (e.g. Kim, Ferrin, Cooper & Dirks, 2004), depriving ourselves of this opportunity can be costly. However, for a complete analysis, this cost must be weighed against the benefit of looking more sincere when lying in order to determine which has a greater evolutionary advantage.
Yet another cost is highlighted by experiments on choice blindness, where participants’ preferred choices are swapped after they have made their choice but they go on to passionately defend the non-preferred choice (Johansson, Hall, Sikström, & Olsson, 2005). These experiments demonstrate that instead of self-deception facilitating interpersonal deception only, it can end up changing the self. For instance, if I try to make someone else believe that I have a belief p, and the only reliable, advantageous way for me to do it is to trick myself that I have a belief p, it is possible that I end up changing myself to actually have the belief p at all levels of processing (Johansson, Hall, & Gardnefors, 2011). This cost, however, highlights the non-unitary nature of the mind, discussed in the previous section.
Last but not least, in deliberate, rational thinking we sometimes construct lines of argumentation, where, along the road, we assume a proposition to be true for the purposes of the argument (Frankish, 2009). However, we might end up believing the proposition without evidence and accept it as an implicit premise in future argumentations or actions (Frankish, 2004, 2011). However, this acceptance (especially of favourable premises) can also be interpreted as a way to self-deceive in our efforts to deceive others. This assertion can be verified if we find that people are more likely to accept implicit, but unproven, premises as true for the sake of an argument in which they have something to gain; then this acceptance might be another bias/route through which we achieve our goals.
Fortunately for VH&T, although the suggested costs raise interesting questions, such as whether a ceiling effect can be specified, or even serve to confirm and/or expand the theory as shown by choice blindness experiments and the literature on reasoning, they do not destabilise it. Albeit more work remains to be done, VH&T’s theory lives to see another day.
Evolutionary stability and confidence
Given the above-mentioned array of costs associated with self-deception, the next challenging question to answer is how it is possible that Mother Nature has not figured out a better way to facilitate advancement in the social hierarchy but through self-deception. That is to say, has self-deception always been an optimal strategy in the evolutionary struggle that enables us to reap the benefits of overconfidence? This line of thinking calls into question the theory and whether it fulfils a crucial evolutionary criterion – evolutionary stability (Cowden, 2012).
Marshall and colleagues (2013) raise precisely this issue. The main premise of VH&T’s theory is that the information-processing biases evolved to enable self-deception which gives us higher confidence than warranted. However, this does not necessarily mean that overconfidence is always advantageous. On the contrary, an ubiquitous application of overconfidence might be useful for attracting a partner as there are little costs for being taken at your word but can have serious costs in animal conflict. Hence, an explanatory mechanism is needed to address this cost-benefit analysis.
Although the original theory does not address this issue, some prospects for resolutions are available. One such example is a recent approach which attempts to explain overconfidence using game theory (Wolpert, Jamison, Newth, & Harre, 2011). Specifically, rational agents in non-cooperative games typically reach a Nash equilibrium, whereby none of the rational agents’ outcomes improves if they change only their own strategy. However, deviations from Nash equilibria, or nonrational behaviour, can be explained if one of the agents signals their irrevocable commitment to a ‘persona’ in the upcoming game. This hypothesized persona, for example, enables individuals to signal more confidence than warranted. The overconfident persona, in turn, might deter the other rational and stronger parties from entering in a contest for low-value resources as it will inevitably result in a costly challenge relative to the reward.
Conclusion
That other people lie is not surprising to anyone. What is less acknowledged is the fact that all too often we also lie to ourselves. Therefore, a better understanding of why self-deception occurs is pressing. Unfortunately, philosophical debates, such as that surrounding the dual representation paradox, have stumbled scientific advancements for too long and it’s high time to break the vicious cycle. VH&T break the shackles of endless, abstract arguments and lay out an empirically-driven approach to self-deception in an attempt to stimulate research from biologists, biological mathematicians, evolutionary psychologist and others. Although some important aspects of this new proposal are being called into question, such as the costs and evolutionary stability of self-deception, it remains a valuable theory that accounts for a wide range of findings and makes testable predictions.
Call it self-deception but I believe that VH&T’s theory lays out the foundation for a complete understanding of self-deception.
References
Bandura, A. (2011). Self-deception: A paradox revisited. Behavioral And Brain Sciences, 34(01), 16-17. doi: 10.1017/s0140525x10002499
Cosmides, L. (1989). The logic of social exchange: Has natural selection shaped how humans reason? Studies with the Wason selection task. Cognition, 31(3), 187-276. doi: 10.1016/0010-0277(89)90023-1
Cowden, C. C. (2012). Game theory, evolutionary stable strategies and the evolution of biological interactions. Nature Education Knowledge, 3(6)
Ditto, P., & Lopez, D. (1992). Motivated skepticism: Use of differential decision criteria for preferred and nonpreferred conclusions. Journal Of Personality And Social Psychology, 63(4), 568-584. doi: 10.1037//0022-3514.63.4.568
Frankish, K. (2004). Mind and supermind. Cambridge: Cambridge University Press.
Frankish, K. (2009). Systems and levels: Dual-system theories and the personalsubpersonal distinction. In J. Evans & K. Frankish, In two minds: Dual processes and beyond (pp. 89-107). Oxford University Press.
Frankish, K. (2011). Conscious thinking, acceptance, and self-deception. Behavioral And Brain Sciences, 34(01), 20-21. doi: 10.1017/s0140525x10002554
Johansson, P., Hall, L., & Gärdenfors, P. (2011). Choice blindness and the non-unitary nature of the human mind. Behavioral And Brain Sciences, 34(01), 28-29. doi: 10.1017/s0140525x10002591
Johansson, P., Hall, L., Sikström, S., & Olsson, A. (2005). Failure to Detect Mismatches Between Intention and Outcome in a Simple Decision Task. Science, 310(5745), 116-119. doi: 10.1126/science.1111709
Kim, P., Ferrin, D., Cooper, C., & Dirks, K. (2004). Removing the Shadow of Suspicion: The Effects of Apology Versus Denial for Repairing Competence- Versus Integrity-Based Trust Violations. Journal Of Applied Psychology, 89(1), 104-118. doi: 10.1037/0021-9010.89.1.104
Lynch, K. (2011). A Defense of a Deflationary Theory of SelfDeception (Ph.D). University of Warwick, Department of Philosophy.
Marshall, J., Trimmer, P., Houston, A., & McNamara, J. (2013). On evolutionary explanations of cognitive biases. Trends In Ecology & Evolution, 28(8), 469-473. doi: 10.1016/j.tree.2013.05.013
Mercier, H. (2011). Self-deception: Adaptation or by-product?. Behavioral And Brain Sciences, 34(01), 35. doi: 10.1017/s0140525x10002281
Pinker, S. (2011). Representations and decision rules in the theory of self-deception. Behavioral And Brain Sciences, 34(01), 35-37. doi: 10.1017/s0140525x1000261x
Smith, D. (2011). Aiming at self-deception: Deflationism, intentionalism, and biological purpose. Behavioral And Brain Sciences, 34(01), 37-38. doi: 10.1017/s0140525x10002657
Trivers, R. (1976/2006) Foreword. In: The selfish gene, R. Dawkins, pp. 19–20. Oxford University Press. (Original work published in 1976.)
von Hippel, W., & Trivers, R. (2011a). The evolution and psychology of self-deception. Behavioral And Brain Sciences, 34(01), 1-16. doi: 10.1017/s0140525x10001354
von Hippel, W., & Trivers, R. (2011b). Reflections on self-deception. Behavioral And Brain Sciences, 34(01), 41-56. doi: 10.1017/s0140525x10003018
Wolpert, D., Jamison, J., Newth, D., & Harre, M. (2011). Strategic Choice of Preferences: the Persona Model. The B.E. Journal Of Theoretical Economics, 11(1). doi: 10.2202/1935-1704.1593
