Before I begin this post, I just want to note that I will be using many complex terms from evolutionary biology throughout. I will do my best to stop and explain them, but should I let one pass without explaining it then Google is your friend!
Firstly, a brief outline of the post. Several theorists working on the psychology of art have, of late, presented (or assumed) a theory of a universal aesthetic psychology, grounded in evolutionary psychological reasoning about human nature. In this post I will criticise this theory by arguing that we should expect to see mixed strategies in creative displays (of which the most obvious is art) which serve as costly hard-to-fake displays of protean cognition. If these mixed strategies are maintained as balanced polymorphisms in protean cognition both across and within individuals, then there is not a single universal aesthetic psychology.
So, what is a universal human aesthetic psychology? Some philosophers and psychologists working on an evolutionarily informed account of aesthetic psychology (such as Denis Dutton 2002 & 2009, Ellen Dissanayake, 1992, Ramachandran & Hirstein, 1999) have, in recent times, focussed on, and attempted to defend, the concept of aesthetic universalism. This is not a simple concept to unpack, but there are two broad trends in this research: claims about psychological universals and claims about cultural universals. Dutton argues that there are universal features of art, and that similar kinds and contents of art are appreciated cross-culturally. Dissanayake argues that artistic practices can be situated in a wider cultural behavioural context, that of “making special”, along with various other cultural practices that humans display the world over. These are claims about culture rather than psychology. Dutton and Dissanayake are also, however, constructing an evolutionary psychological account of the human aesthetic response. The basic idea behind their claims is that art occurs in all human societies and make us feel good. Things that are universal and intrinsically pleasurable tend to be the products of evolutionary forces. For evolutionary psychologists, then, underlying these cultural universals (and generating these cultural universals) are human psychological universals. Ramachandran takes a slightly different route. He argues that, rather than having a specific adaptive function per se, art is pleasing because (most) works adhere to one or many of his “eight laws of aesthetic experience” rules, each of which has its own adaptive function. Again, however, he is assuming an underlying universal human nature. These eight laws apparently explain all human aesthetic experience. He just doesn’t see art as being an adaptive complex like Dutton and Dissanayake.
So, I need to unpack the concept of a universal human nature. Tooby and Cosmides (1990) argue that all humans share the same complex cognitive adaptations, which they refer to as “universal human nature”. Tooby and Cosmides are interested in what aspects of the human mind can be seen as species-typical design, in an approach analogous to theories of species-typical adaptations in evolutionary biology. On this view, all humans share essentially the same kinds of genetic “developmental programs” for cognitive structures and hence will all, after various critical periods of learning, develop into adults with the same kind of cognitive structure – a species-typical universal cognitive architecture. (Note it is commonplace to refer to these cognitive structures as “mental modules”, but I want to avoid this term as it is liable to cause a lot of philosophical disagreement). Moving on, this does not mean that all human minds are the same; there is obviously a great degree of variance between the cognitive abilities of different people. Tooby and Cosmides argue that phenotypic diversity in cognition does not arise because of fundamental differences in human genotypes, but rather through the process of different environmental circumstances affecting the outcomes of universal developmental programs. This process is called adaptive plasticity, and – for Tooby and Cosmides – is responsible for all phenotypic psychological diversity in human populations. Evolutionary biologists, however, do not consider adaptive plasticity to be the only mechanism via which phenotypic diversity can occur. Another mechanism that can lead to phenotypic variation is genetic polymorphism (Ford, 1940). My argument hinges on showing that there are likely to be polymorphisms underlying an essential component of aesthetic psychology, namely creativity. Essentially, if there are polymorphisms underlying our creative faculties, then these polymorphisms are complex enough to mean that Tooby and Cosmides’ claim that adaptive plasticity is the only mechanism underlying human psychological variation is false. I will return to this mechanism later.
What this all means, according to evolutionary psychology, is that underlying artistic production and appreciation is a universal human aesthetic nature, i.e. the cognitive adaptations responsible for our appreciation of art and our desire to produce art are universal in the species. Any surface variation is caused by environmental variation affecting the development of these species-wide cognitive adaptations. This is the thesis of aesthetic universalism that I will be focussing on in this post and I will attempt to show it to be problematic by arguing that there are polymorphisms in our creative abilities.
I need to begin my argument by defining sexual selection, sometimes called Darwin’s “forgotten” theory. Natural selection is the well known mechanism via which much genetic change in a species occurs. Genetically unfit individuals (and the genes which contribute to their lack of fitness) are more likely to be removed from the gene pool because they are less likely to survive than individuals with beneficial genes, who tend to survive, have more offspring, and pass on these beneficial genes to their offspring. The other major mechanism via which genetic change in a species occurs is sexual selection. Sexual selection is the mechanism which explains how individuals are able to pass on their genes in the first place; it argues that certain traits or signals an individual possesses can be seen as attractive by individuals of the opposite sex, and that those who possess the most attractive traits or signals are likely to pass the genes for these traits/signals on, thus increasing their incidence in the population. Natural selection is about survival and genes beneficial to survival. Sexual selection is about sex and genes beneficial to being able to have sex.
In general, sexually selected traits are reliable indicators of fitness. By reliable, I mean they are hard to fake. For if any old person can fake a desirable trait, then that trait no longer serves to reflect something beneficial about that individual. For example, a person can signal their monetary wealth by buying a very expensive car; only those with a lot of money can buy these cars, so possessing an expensive car is an honest signal of monetary wealth (unless, of course, the car has been stolen). Selection seems to usually achieve this in accordance with the handicap principle, developed by Amotz Zahavi (1975). For an individual to signal to other individuals that he or she is genetically fit, what is needed is an elaborate and costly signal which seems very wasteful. If an individual is able to maintain this handicap effectively, then receivers know that the signal indicates quality because inferior quality signallers cannot afford to produce such wastefully extravagant signals. This is what we find in many other species (for example, the mating displays of birds of paradise, and the large antlers of moose, or the peacock’s tail). Geoffrey Miller (1997) argues that displays of human creativity (in humour, in story-telling, and, of course, art) count as costly reliable signals of imaginative abilities, being dependent, as they are, on the healthy functioning and distinctive size of the human brain. For our brains take many years to grow to maturity, and are extremely costly to maintain in terms of their energy requirements (accounting for 40% of the body’s glucose consumption, for example). In addition, Miller thinks, there was probably sexual selection in hominids for honest advertisements of protean cognition, of which the most obvious are partly-random displays of creative thought. Protean cognition is genuinely random cognition, i.e. cognition which is not consciously random but rather occurs via some mechanism unknown to the individual. For example, many animals show protean behaviour. Rabbits escape from predators in a random zigzag pattern, a behaviour which is instinctual. Certain creatures randomly change colours when faced with a predator in order to confuse the predator’s perceptual expectations (sometimes called search images). These escape strategies are protean, and their genuine randomness is what makes them adaptive – the predators cannot predict the prey’s behaviour.
I now need to introduce the concept, borrowed from evolutionary game theory, of frequency-dependent selection. This is perhaps best described through examples. The Hawk-Dove game is a game where contestants play either a Hawk of Dove in competing for a resource worth 40 points (points here are “fitness points” – a way of quantifying fitness in a game theoretical model). Hawks are always aggressive until seriously injured. Doves produce a threatening display but never attack and retreat if attacked by an opponent. Doves, then, always lose to Hawks. But let’s assume Hawks have a 50% chance of defeating another Hawk and Doves have a 50% chance of defeating another Dove. Also, assume that a serious injury costs -60 points and wasting time and energy in a very long contest costs -10 points. I won’t go into the maths, but given the fitness costs and benefits in this simple model, selection favours an evolutionary stable mix of 75% Hawks and 25% Doves in a population. If this number shifts, then the model predicts that the numbers of Hawks and Doves in a population would rapidly fluctuate until it, again, reached this evolutionary stable mix. Too many Hawks or too many Doves is an unstable strategy – what works best, in this case, is what is known as a mixed strategy.
If behaviours have frequency-dependent fitness, like in the Hawk-Dove game, then “selection often maintains a particular proportion of alternative variants” (Buller, 2005, p. 40). When this happens, it is referred to as frequency-dependent selection. Stable mixes of genotypes in a population are referred to as stable polymorphisms. The garter snake has highly variable colour patterns, ranging from stripes to mottled blotches (Brodie, 1990, 1992). The snake’s colour seems to be coupled with differences in escape strategy – striped snakes flee in a straight line, creating a visual illusion which makes it difficult to pinpoint the exact position of the snake, and mottled snakes remain motionless, using the mottles as camouflage. Paracerceis sculpta is a marine isopod crustacean which has males of three different sizes (Shuster, 1987). The large males guard harems of females, and thus secure their copulations in this fashion. Medium sized males resemble females and are able to mimic the female courtship display and fool the large males into allowing them into the harem, where they are able to copulate with the females. Small males are small enough to sneak past unobservant large males and enter the harem, securing copulations. These various strategies all have equal success and are a clear case of a stable complex cognitive polymorphism being maintained in a population. Furthermore, the genes underlying them have been identified.
The Machiavellian Intelligence hypothesis presents us with a good potential set of balanced cognitive polymorphisms in creative abilities; more specifically in protean cognition (Miller, 1997). The hypothesis itself suggests that apes and humans have evolved special cognitive adaptations for predicting and manipulating the behaviour of other individuals (Humphrey, 1976; Byrne & Whiten, 1988; Whiten & Byrne, 1988) which are all grouped together under the umbrella term “Machiavellian Intelligence”. These adaptations are postulated to include a “Theory of Mind” module for attributing beliefs and desires to others, to better predict their behaviour (Leslie, 1994; Baron-Cohen, 1995; Dennett, 1988). Suppose the Machiavellian Intelligence hypothesis is right. It seems unlikely that evolution would stop there, allowing every individual to manipulate every other individual. This would be an unstable strategy. What seems more plausible is that certain defences to exploitation would evolve. Miller suggests that a potential defence against Machiavellian exploitation is protean cognition, or genuinely unpredictable cognition.
Let’s talk about animals again for a second. In terms of animal fleeing behaviour, unpredictability can be useful at many levels. When threatened, octopi, cuttlefish, and sea pansies use “colour convulsions” across the fast-response chromataphores on their skin, quickly going through different colour patterns to defeat the search images (perceptual expectations) used by their predators (Driver & Humphries, 1988). Would unpredictability still prove adaptive if we shift attention from trajectories through physical space to more abstract trajectories through the space of possible social behaviours? When would “social proteanism” be selected?
Well, what if an individual could behave in a way similar to a threatened octopus when faced with the prospect of Machiavellian exploitation? The Machiavellian succeeds when the individual’s behaviour and cognition is predictable and ordered. If genuinely random cognition is possible, then the Machiavellian would not be able to predict and exploit. This cognition would have to be genuinely random, and the best way of achieving this would be by making the individual themselves unaware of the processes generating the random cognition. It’s almost like a double-blind. The individual, then, cannot give away clues as to what their behaviour will be when this behaviour is caused by protean cognitive processes.
I have been talking as though protean cognition would mean that an individual displays a certain amount of random and unpredictable behaviour. Clearly, individual unpredictability would make it difficult for a Machiavellian individual to exploit someone. People are not entirely unpredictable, however. In fact, people are, in many ways, quite predictable. If individuals were totally unpredictable then this would be disastrous for social cohesion. You need to be able to trust that the person you’ve paid to paint your house will actually paint it and not set it on fire, and that your girlfriend or boyfriend will not one day just suddenly decide to kill you. Clearly, if protean cognition exists, then, it is unlikely to be displayed solely as unpredictable variation in the individual over time. Individual unpredictability over time may exist, and may be an important way of resisting exploitation, but it is unlikely to be the whole story.
So how is protean cognition to be fully understood? Miller has argued in a previous paper (Miller & Cliff, 1994) that animal pursuit and evasion contests are likely to result in mixed strategies between pursuer and evader. This work is based on complex game theoretical models that I won’t go into here, but the conclusion appears to be that there is no one single best strategy in many different kinds of contests between pursuer and evader. The kinds of strategies used by either party tend to be dependent on the behaviour of the other party, i.e. the strategies show frequency-dependence. It seems plausible to see evasion of Machiavellian exploitation as an example of a pursuit and evasion contest. So we should expect to see mixed strategies in individual’s responses to Machiavellian exploitation. As Maynard Smith (1982) has pointed out, any frequency-dependent balance can be implemented either between individuals, as a genetic polymorphism, or within an individual, as unpredictable strategic variation over time.
Social strategies which show wide degrees of intra-individual variation are often subject to frequency-dependent selection between individuals. Examples include the balance between aggressiveness and bluffing in the Hawk-Dove game (Maynard Smith, 1982), between deceptive and honest signals in animal signal theory (Zahavi, 1975; Dawkins & Krebs, 1976), and between extroversion and introversion in evolutionary personality psychology (Buss, 1991; Wilson, 1994). As I have just argued, unpredictable variation in an individual’s response to Machiavellian exploitation is unlikely to be the whole story. Thus, we are also likely to see unpredictable variation across individuals as well as within an individual. So this means that proteanism is sometimes likely to be displayed as different personalities and behaviours distributed across a population. The randomness aspect comes from not being able to predict whether a particular individual one meets will have a certain personality or display a certain behaviour. Furthermore, in order to retain the genuinely random aspect of this distribution, the individual themselves would need to be unaware of the source of their behaviour so as not to give away any clues to the Machiavellian. This seems to be what we see in actual human behaviour: people are notoriously bad at introspection and understanding the sources of their behaviours (Nisbett & Wilson, 1977).
Furthermore, Clarke (1962) postulated that “apostatic selection” (in which predators overlook rare prey types while consuming an excess of abundant ones) in favour of low-frequency body types and behaviours could maintain substantial morphological and behavioural polymorphism in a species. For example, birds have more trouble finding and eating snails in species with high levels of polymorphism in shell colour and pattern, because the birds’ perceptual expectations don’t work as well. The same argument may apply to polymorphism at other levels: the predictive power of Machiavellian intelligence may favour greater diversity in personality (i.e. stable social strategies). Buss (1991) and Wilson (1994) recognize that frequency-dependent selection can maintain genetic variation in personality traits, but they neglect this possible apostatic effect. Consider how much easier social interactions would be if every individual one encountered had exactly the same-personality — i.e. the same social-strategic repertoire, the same goals, the same tastes in mates and friends, the same thresholds for anger or gratitude, etc. Social complexity is not just a function of the number and quality of social relationships in a group (Dunbar, 1993), but also of the inter-individual variations in personality that must be perceived, remembering, and strategized about. Given Machiavellian intelligence, apostatic selection may favour a rapid diversification of personality types, and may maintain behavioural polymorphisms at higher levels than they would otherwise attain.
Suppose, then, that protean distribution of personalities and individual protean cognition have become an aspect of human social intelligence. Insofar as divergence in personalities, along with individual protean cognition, contributed to competitive success, along with size, strength, health, and social skills, we might expect that mate choice mechanisms would evolve to favour exaggerated displays of proteanism. Zahavi’s (1975) handicap principle might predict that only proteanism-displays that are costly would be reliable, so proteanism-displays would probably become elaborated and specialized, leading to special “protean courtship displays” that might bear little resemblance to the original social protean tactics. Proteanism-displays in courtship, then, may evolve to be quite distinct from social proteanism in competition, and their informational features may become elaborated (e.g. into unpredictability at higher and higher levels of cognitive or strategic performance). We should expect to see some elaborate way for individuals to display 1) their individual protean abilities and 2) their differences to others in the population.
It seems to me that art is a prime candidate for this role. Artists who toil and strive for hours and months and years over works seemingly irrelevant to their immediate survival could be displaying an exaggerated form of their individual protean cognitive abilities and their personality to prospective sexual partners in a highly costly hard-to-fake fashion. As with most displays, the link may not seem obvious. All that would have to be the case is that their artworks are highly individual and strikingly different to many other kinds of artworks in existence, because, as explained earlier, unpredictable variation is likely to exist between individuals as well as within individuals, if not to an even greater extent. Art, then, serves as an exaggerated display of one’s polymorphic protean abilities and differences to other individuals. (Note that artistic output is clearly influenced by ones contemporaries, historical influences and the overall social and historical context in general in which the art is situated. I don’t wish to deny this. What I am saying, however, is that quite a large part of what makes art art (creativity and individuality) could stem from its functioning as a display of protean personality distribution and cognition). As mentioned before, the link between art and personality is not obvious, and, more importantly, should not be obvious (for it to be a costly, hard-to-fake signal). Based on Zahavi’s logic, the display is likely to be exaggerated and highly idiosyncratic. This is what we tend to see in art.
It seems likely, then, given the individuality of these displays (stemming from underlying polymorphisms) that there will be multiple kinds of aesthetic psychology, or, perhaps, multiple art instincts, with various artists (and art appreciators) preferring various different kinds of art, for it is within this variation that we see the protean distribution of personalities and their associated displays wrought large.
Polymorphism and adaptive plasticity
It’s time now to return to Tooby and Cosmides’ argument that the only mechanism that can lead to human psychological variation is adaptive plasticity. Adaptive plasticity, remember, is the process of different environmental circumstances affecting the outcomes of universal developmental programs. As I have argued, however, there are likely to be polymorphisms in protean abilities, and, therefore, in protean courtship displays, e.g. art. Biologically, polymorphisms are encoded in an individual as a genetic switch, set very early on in development. This is the crucial difference between polymorphisms and adaptive plasticity; the ultimate phenotypic differences caused by adaptive plasticity are caused by the same genotype acting under different environmental conditions, whereas polymorphic phenotypic differences are due to actual differences in the genotype, caused by genetic switches that shunt development irreversibly down one of several pathways. Say for the purposes of this discussion that there are ten distinct protean polymorphisms. This means there are ten distinct ways a genetic switch can be set, shunting development irreversibly down one of these ten pathways. The important thing to note here is that the individual, once the switch is set, can never change to a different polymorphism – the environment cannot have an effect on the kind of polymorphic ability they eventually show (of course, the environment can have an effect on how well this ability is eventually displayed, for example given both inadequate nourishment and human interaction an individual would display very stunted psychological growth, and thus their protean abilities could seem indistinct from the abilities of any other individual). What this all means is that – given that polymorphisms exist for protean distribution of personalities – there cannot be a single set of universal developmental programs for cognitive architecture. Differences are encoded in the genotype. Given that protean courtship displays such as art are products of this protean distribution, which likely to be polymorphic, there is not likely to be a universal aesthetic psychology.
An objection and a conclusion
Suggesting that art is a protean courtship display captures the originality aspect of art. It explains why we like very original and creative art – it shows the artist to have a highly individual personality and highly protean cognition, meaning that they are less likely to be susceptible to Machiavellian exploitation. Art, however, is also respected for many other reasons: things like technical prowess and moral and political messages. Isn’t analysing art as a protean courtship display missing many things that are very important about art?
Well yes, it is. But I do not wish to claim that a protean courtship display is all that art is. Much more can be said about art as a cultural or psychological phenomenon that this analysis just doesn’t touch. My aim, however, was only to present one possible way of criticising the theory of a single universal aesthetic psychology. Perhaps it will be possible to show that other aspects of our aesthetic psychology are universal. I have my doubts, however. I can’t discuss exactly why in this post, but come back to me in a few years time when I’ve finished my thesis and I’ll have some answers.
Baron-Cohen, S. 1995. Mindblindness. Cambridge, MA: MIT Press.
Brodie, E. D. I. 1990. Genetics of the Garter’s getaway. Natural History, 45-51.
Brodie, E. D. I. 1992. Correlational selection for colour pattern and antipredator behaviour in the Garter Snake Thamnophis ordinoides. Evolution, 46, 1284-129.
Buller, D. J., 2005. Adapting Minds: Evolutionary Psychology and the Persistent Quest for Human Nature. Cambridge, MA: MIT Press.
Buss, D. M. 1991. Evolutionary Personality Psychology. Annual Review of Psychology, 42, 459-491.
Byrne, R. & Whiten, A. (eds). 1988. Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes and Humans. Oxford: Clarendon Press.
Clarke, B. C. 1962. The evidence for apostatic selection. Heredity, 24, 347-352.
Dawkins, R. & Krebs, J. R. 1978. Animal Signals: Information or manipulation? Behavioral Ecology: An Evolutionary Approach. 2nd edn, ed. J. R. Krebs & N. B. Davis, pp. 282-309. Oxford: Blackwell Scientific Publications.
Dennett, D. 1988. The intentional stance in theory and practice. Machiavellian Intelligence: Social Expertise and the Evolution of Intellect in Monkeys, Apes and Humans, ed. R. W. Byrne & A. Whiten, pp. 180-202. Oxford: Clarendon Press.
Dissanayake, E. 1992. Homo Aestheticus: Where Art Comes From and Why. New York: Free Press.
Dunbar, R. 1993. Coevolution of neocortical size, group size and language. Behavioral and Brain Sciences, 16, 681-735.
Dutton, D., 2009. The Art Instinct. Oxford: Oxford University Press
Ford, E. B. 1940. Polymorphism and taxonomy. J. Huxley, The new systematics. Oxford: Oxford University Press.
Humphrey, N. 1976. The social function of intellect. Growing Points in Ethology, ed. P. P. G. Bateson & R. A. Hinde, pp. 303-17. Cambridge: Cambridge University Press.
Leslie, A. 1994. ToMM, ToBY, and Agency: Core architecture and domain specificity. Mapping the Mind: Domain Specificity in Cognition and Culture, ed. L. A. Hirschfield & S. A. Gelman, pp. 119-148. Cambridge: Cambridge University Press.
Maynard Smith, J. 1982. Evolution and the Theory of Games, Cambridge University Press.
Miller, G. 1997. Protean Primates. Machiavellian Intelligence II: Extensions and Evaluations, ed. A. Whiten, & R. W. Byrne. Cambridge: Cambridge University Press.
Miller, G. F. & Cliff, D. 1994. Protean Behaviour in Dynamic Games: Arguments for the co-evolution of pursuit-evasion tactics. From Animals to Animats 3, Dave Cliff, Philip Husbands, Jean-Arcady Meyer & Stewart W. Wilson, eds. MIT Press.
Mithen, S. 1998. The Prehistory of the Mind: A Search for the Origins of Art, Religion and Science, Pheonix.
Nisbett, R. E., & Wilson, T. D. 1977. Telling more than we can know: Verbal reports on mental processes. Psychological Review, Vol. 84, Issue 3, p. 231-259.
Ramachandran, V.S. & Hirstein, W. 1999. The Science of Art. Journal of Consciousness Studies 6, No. 6-7, pp. 15-51.
Shuster, S. M. 1987. Alternative Reproductive Behaviors: Three Discrete Male Morphs in Paracerceis sculpta, an Intertidal Isopod from the Northern Gulf of California, Journal of Crustacean Biology, Vol 7, No. 2, pp. 318-327.
Tooby, J. and Cosmides, L. 1990. On the universality of human nature and the uniqueness of the individual: the role of genetics and adaptation. Journal of Personality 58, 1.
Whiten, A. & Byrne, R. 1988. Tactical deception in primates. Behavioral and Brain Sciences, 11, 233-273.
Wilson, D. S. 1994. Adaptive Genetic Variation and Human Evolutionary Psychology”, Ethology and Sociobiology 15, p. 219-235.
Zahavi, A. 1975. Mate selection – a selection of handicap. Journal of Theoretical Biology, 53, 205-214.