Is art an adaptation?
Prospects for an evolutionary perspective on aesthetic emotions
late uncorrected draft of a paper published in Jerrold
Levinson & Jenefer Robinson, ed., Art, Mind, and Cognitive Science: special issue of Journal of
Aesthetics and Art Criticism, 62/2 (2004): 109-118.

Ronald de Sousa
University of Toronto
sousa@chass.utoronto.ca
http://www.chass.utoronto.ca/~sousa

One day, however, a guest noticed it and said: "Oh, I see you got one of Nick's rejects.  Did you find it on the beach? Nick is that hippy sculptor who carves souvenirs for tourists.  This looks like one of the failures he sometimes dumps on the beach....."

Now the pebble was nothing but the charmless reject of a mediocre craftsman. All the strange and wonderful beauty the man had so much admired was gone.

Whether you believe, with Hume, that the actual world shows evident signs of being a botched effort, or whether you take, with Leibniz, the more gloomy view that this is the best of all possible worlds, the thought that it has been intentionally contrived seems, as in the case of the wrought pebble, a sad disappointment next to the wonder of a universe unfolding according to the laws of chance and necessity. Yet paradoxically when looking at the gifts of nature even the most skeptical too often find themselves looking for purpose. We will see this in at least some of the speculations the question of the "function" of art has prompted.

But the parable suggests a number of other morals:

·It illustrates the fact that our aesthetic attitude to an object may be radically different depending on whether the object is thought to be a product of "nature" or an artifact.

·In doing so, it reminds us of the two faces of the aesthetic: on the one hand, the aesthetic question concerns the origin and nature of our sense of beauty.  On the other hand, it raises the question of the origin and nature of the impulse to create art.  The two may seem inseparable, just as speaking seems inseparable from understanding. But the appearance of inseparability can be misleading, as the variety of types of aphasia shows. Neurologically, it turns out that capacities that we think of as merely aspects of the same thing can be controlled by entirely different mechanisms. (FIG 1:)

     So it might be with art and beauty: this is an eventuality that needs to be kept in mind even though, at the level of generality appropriate to the present discussion, I will ignore it. I shall here speak metonymically of "art" to refer indiscriminately to the sense of beauty, the emotional response to art, and the urge to create it. 

·The parable draws attention to the question of the universality of aesthetic taste.  Specifically,  my own (tendentious) interpretation of it rests on a commonplace about the aesthetic sense: namely that it tends to privilege surprise along with order: so that order is valued more highly when it arises where it is least expected.  Later I shall consider some more detailed and sophisticated hypotheses that have been offered in the literature about the "laws of beauty". 

2. Adaptations, exaptations, or spandrels? 

The claim that art has a biological function might be meant in at least four senses, corresponding to four distinct meanings of "function":

(i)"Cummins functions":  The most intuitively straightforward sense of ‘the function of i in system S is F' is that by doing F,  i contributes to the overall goal of S. (Cummins 1975) Thus the heart's pumping action contributes to the overall goal of the circulatory system, i.e. roughly to distribute nutrients and remove waste from all parts of an organism.  But the simplicity of this characterization is only apparent.  For it is unhelpful in practice unless we can specify just what the system S and its "overall goal" are.  In the present case, the question would be: if art has a function, what is the system to which it contributes?  Is it individual health? individual happiness? group harmony? social and technological progress? A case could be made for each of these and more; but none is self-evidently biological.  No doubt we shall be tempted to bring in "survival of the fittest".  But this expression is a misleading one. Actually no organism ever survives: the best I can hope for is to pass on half of my genes. (But my genes aren't me, and their purposes may be deeply alien to mine.)  We therefore need to specify the relevant "system" as a type rather than a token organism: for types do get reproduced; tokens at best only get copied. The Cummins-function of art, then, would be consist in whatever contribution it makes to the fitness of a certain type of organism.  But this interpretation of the notion of function is still relative to a stipulation of the "system” in which we are interested.  Cummins-functions are interest-relative.

(ii)The  aetiological concept: "proper function." A genuinely objective specification of function requires that we define (though not necessarily discover) intrinsic functions, functions that are not interest-relative.  This is precisely what Ruth Millikan's notion of proper function is designed to do.  (Millikan 1984;  1989).  This requires an objective way of distinguishing the  proper function of an organ or part from its mere side-effects. The heart doesn't only pump the blood: it also produces rhythmic sounds. But intuitively only the former is its function: the sounds it makes are merely side-effects.  Proper functions as those effects that have aided the reproduction of ancestral organisms of which effects of that type were present. A slightly more precise formulation is due to Paul Griffiths:

Where i is a trait of systems of type S, a proper function of i in Ss is F iff a selective explanation of the current non-zero proportion of Ss with i must cite F as a component in the fitness conferred by i. (Griffiths 1998, 442)

In terms of this concept, the claim that art has a function amounts to saying that our ancestors happened on a mutation, endowing them with what for simplicity we may call the "art gene", which caused them to leave more offspring than those who were stuck with the old artless alleles. 

For later reference, it is important to note that if Fis a proper function of i, it by no means follows that it is universal.  On the contrary, as Millikan stresses with reference to the example of spermatozoa, F (in this case the fertilization of an ovum) can be the proper function of an item i,despite the fact that only one time in millions doesi accomplish F.

(iii)Exaptations, a term coined by (Gould and Vrba 1982), are features that originally have one function but which are shaped by new selection pressures to serve novel functions. Examples are plentiful in evolution.  A particularly striking one concerns the small bones in our middle ear:  "Now used for hearing, two of these bones (the malleus and incus) were originally part of the lower jaw of our reptilian ancestors, who used them for chewing."  (Ramachandran and Blakeslee 1998, 210)  (FIG2) 

(iv) "Spandrels".  Exaptations focused attention on the haphazard nature of evolution—its capacity opportunistically to change direction and hijack an organ selected for one function and turn it to novel use. But there are also cases where side-effects of an existing functions never were subjected to selective pressure. Strictly speaking these are not adaptations at all. They may nevertheless come to be valued, but they had no part in promoting the reproduction of their ancestors.  To single out such cases, (Gould and Lewontin 1979) adapted the architectural term for the roughly triangular areas resulting from the design of fan-vaulted ceilings.  The spandrels of San Marco frame beautiful mosaics, but (Gould assumes) they were not originally meant to frame designs. (FIG 3)  [from  Glossary of Medieval Art and Architecture]  They were just necessary consequence of other architectural decisions, and architects were forced to make the best of them. 

In evolution, spandrels are characteristics that find a use for which they were never selected.  A toy devised by Elliott Sober illustrates the distinction on which this point rests, between  selection for some trait, and that trait's merely being selected. (Sober 1984, 99)  (FIG 4)

In the toy, the green balls are selected, i.e. got to the bottom. But their colour plays no part in the causal explanation of that fact. They got there because they are the smallest. Small size is what the apparatus has selected for.  Their colour just happens to be associated with their size, so green ones were indeed selected, but not selected for. 

So it might be with various human capacities.  Two examples that come to mind are higher mathematics and female orgasm.  In both cases, the argument that these are "spandrels"  rests on an important truism and a methodological principle about natural selection.  The truism is that a gene cannot be selected for unless it has phenotypic consequences that are "visible" to natural selection.  The methodological principle is one of parsimony: we shouldn't assume that something is a result of adaptation if it can be satisfactorily accounted for in terms of non-selective aspects of its causal history, e.g. as a side-effect.  In both the cases in question, the probability is that they could not have been reinforced by selection because they never surfaced as such in the phenotype. In the case of the female orgasm, the anthropologist Frances Burton found that the female macaques on Gibraltar were capable of orgasm, but that it took twenty minutes of vigorous stimulation to produce it. Copulations in the wild, however, never last more than a minute and would probably endanger the lives of the participants considerably if they did.^[1] So the truism argues that selection for female orgasm couldn't have taken place.  Furthermore, given the homology of penis and clitoris, there is no more mystery about the capacity for orgasm to be produced by stimulation of the latter as of the former: thus parsimony dictates that we don't need an adaptive explanation any more than we need one for the presence of male nipples. Such capacities are spandrels: Cummins-functions that lack the appropriate aetiology to be proper functions.

3.  Does it matter if art is a spandrel? 

In the light of these distinctions we can refineour question.  One could ask a number of questions about the "function" of art in the welfare of individuals, education, societies, and so forth.^[2] But questions about the biological function of art seem to be aimed at the difference that art made to evolution, not to its present uses. Does art have a proper function? is it an exaptation? Or is it merely a spandrel?

Recall one of the morals of my opening parable: the feeling induced by the idea that something is an artifact is quite different from the one likely to be inspired by a "natural" object. But in the cases in question the objects are all "natural objects"; at the same time they are also all crucially like "products of design" in that special way that applies, if only by analogy, to the products of natural selection. Much of the discussion that follows will consist in attempting to reconcile the intuitions prompted by these conflicting considerations. 

Let us begin with Helen Dissayanake, who has argued in a number of very stimulating books and articles (Dissanayake 1992; 1999; 2000) that art is an adaptation in the strongest sense of having (multiple) proper functions. She contrasts this view with the more common thesis that art rides on the back of a number of other adaptations. Here is one list she offers of adaptations of which art has been held to be a spandrel: 

"communication, play, display,and, exploration and curiosity, amusement and pleasure, creativity and innovation, transformation , the joy of recognition and discovery, the satisfaction of a need for order and unity, the resolution of tension, the emotion of wonder, the urge to explain, and the instinct for workmanship"  (Dissanayake 1992, 27); cf. also  (Dissanayake 2000). 

All are plausible enough, but she finds them dissatisfying because she is keen to show that the arts are themselves an adaptation, not a "mere" spandrel.  She seems to think that if art is to be vindicated as really valuable, it must be selected for in its own right. She wants to show how "art could have arisen and persisted, not as an epiphenomenon of other behavior but as a positive and primary motivation in its own right".  (Dissanayake 1999, 33).  And again:

Western modernity might view the arts as "useless," but biologists, using the evolutionarily salient criteria of universality, energy investment, and pleasure, would have to concede that engaging with the arts—like eating, sleeping, sex, socializing, and parenting—is a fundamental and essential part of human nature."  (Dissanayake 1999, 29)

In short, on her view, if the arts are taken seriously, we must think of them as 

a) having genuine proper functions, not just being spandrels or exaptations. 

b) universal; and 

c) being "an essential part of human nature", apparently by virtue of being genetically innate. 

All three of these ideas are questionable. 

a) Actually it is likely that the majority of adaptations have their roots in original exaptations, serendipitous hijackings of features developed for other reasons. That is one of the beauties of evolution, which, as François Jacob memorably remarked, consist essentially in bricolage. [Sp. ¿"chafallada"?] Bricolage suits exaptations particularly well, but spandrels may be just as valuable for the life of an organism.  And if "having a function" is taken in the Cummins sense, then nothing follows about whether it was selected for. What we value may well be the result of pure chance; indeed, as my parable illustrates, a product of pure chance may be more valuable than one of design.

b)  As we saw a moment ago with the example of spermatozoa, universality is irrelevant to the genuineness of functions.  It could be that art does indeed have a proper function, but that it seldom fulfills it.

c)  What innateness means isn't that clear. It is a concept often used for divergent and  often sometimes politically charged effect in psychology and biology, and some have recently argued that it is "irretrievably confused" (Griffiths 2002; cf. Godfrey-Smith 2002).  In particular, one can't really insist on universality as a mark of the innate, since some traits—such as eye colour—are innate if anything is, but far from universal. (In fact, no non-trivial characteristic is ever universal in biology.)  As for whether universality entails innateness, consider this thought experiment.  According to British Council figures, English language is spoken by about a quarter of the people alive today. Imagine that some day it becomes universal. Obviously that will not mean it's been selected for in the biological sense.  Nor would it mean it's "essential" to human nature, if that means that anyone who didn't speak English wouldn't really be human.

The example returns us to the normative character of the notion of function. When Dissayanake says "universal", "essential",  "innate", she doesn't really mean any of those things literally.  Rather, she means something that is rather well captured by the notion of "proper function": namely that art (1) is a good thing,  and (2) that some of the ways in which it is a good thing have contributed to the dissemination of heritable dispositions to artistic appreciation and production.  But remembering that f can be the proper function of i even though i very seldom performs f, the claim that art has a proper function should not by itself be taken to imply anything much about its frequency.

4.  But what is art anyway?

Before proceeding to look at some specific explanatory hypotheses, we should take a moment to sketch what sort of thing is in question when we speak of art.  Here we could do worse than to follow the guidance of (Abrams 1953), who distinguishes four kinds of theories of art, classified in terms of their emphasis on one of the four elements that are typically related in a work of art: the universe, the author, the audience, and the work itself. Although Abrams thinks of this as a classification of theories, we can as well think of it as a classification of types of art, in so far as each type of art is partly conditioned by the prevalent assumptions about the point of art.

·Mimetic theories stress the idea of representation, and thus focus on the universe outside the author; mimetic art is what we think of as conventional, representational art.

·Pragmatic theories stress the effect of the work on its audience.  "Pragmatic" art, then, is the kind Plato approved of: providing it was successful in improving its audience.^[3]

·Expressive theories shift the focus to the artist. The paramount value becomes sincerity. Expressive art could be thought of as a variant of the mimetic, with a change in the object imitated. For the expressive theory of art asks simply that one be true to oneself, as opposed to the outside world:. 

·Finally Objective theories focus on the work of art itself.  Typical objective art is modernist art, focused, like Aristotle's god, on the contemplation of its own contemplation.

As we look at some of the ideas advanced about the function of art, we shall see that some are more appropriate to some types of art and theories of art.

5.  Some proposed functions of art

Proposed biological functions of art are legion. In what follows, I will limit myself to a sampling of recent suggestion.  Like Dissanayake, Frederick Turner defends the idea that the sense of beauty is "an adaptation". He gives this idea two interpretations.  One is the plausible one that art has served our ancestors in navigating the world in practice.  The more ambitious thesis is that art actually reveals to us the deep structure of the laws of the universe. Let us start with the more moderate thesis:

"Antiphonal birdsong, the brilliant shapes and colors of flowers (what more precise record could there be of the aesthetic preferences of bees?) and the gorgeous ritual mating garments of tropical fishes and birds of paradise, all attest to a more-than-utilitarian attraction in certain forms of organization.....Beauty is an adaptation" (Turner 1999, 119).

Is this scientific insight or romantic rubbish? In trying to answer this, I leave aside the question of whether it is reasonable to attribute to bees the sort of consciousness that would seem to be entailed by the possession of "aesthetic preferences". 

First, it would be better to say: beauty is a co-adaptation rather than an adaptation. It's not as if the bees came with a pre-existing set of aesthetic preferences, and the flowers had just had to bring their colour schemes into line. Nor, presumably, is there any reason to think it was the other way around.  So this leaves open the question:  Why is it just these colours and shapes that the flowers flaunt and the bees prefer?

This question can be given two senses: a strictly causal sense, and an information-theoretic sense. To see this, consider the question: Why aren't  flowers green?  The answer is that they wouldn't be as visible against the background of green leaves: that is a requirement of the signalling or informational function of flowers.  But why not make leaves a different colour? Because, for strictly causal reasons you can't do this: green is already taken, pre-empted by the chlorophyll essential to photosynthesis.  That is the causal constraint on the signalling function.  So when we take the existing causal constraint together with the informational function, we can expect that  flowers will be any colour except green. 

Just which colour each flower will be may depend entirely on chance.  But the ecology of which the flower is a part is constrained by a requirement of variety.  As Darwin first observed in his famous island finches, niches are formed among differentiating populations by specialization or "adaptive radiation": if I can feed on what you ignore, there will be that much less to fight about.  From the informational point of view, it doesn't matter whether you take the roses and I take the lilacs, or vice-versa.  Chance may well rule.  Or some pre-existing slight difference in perceptual or behavioral bias may make it slightly easier for you to take the roses. 

So far, there does not seem to be any justification for Turner's claim that we are dealing with a "more-than-utilitarian attraction" on the part of flowers and birds or bees. Colours are entirely functional. That doesn't mean that every aspect of it is an adaptation: but that's just what we expect from the logic of evolution, where chance proposes and natural selection (sometimes) disposes.  Once the differentiation gets started, however, there will be selective pressure for it to proliferate.

6.  Non-standard mechanisms of selection

The cases considered so far illustrate an important point: that some of the most important parts of the environment to which organisms must adapt is constituted by other organisms that must reciprocally adapt to them. 

There are two variant mechanisms of natural selection that share that characteristic, and  provide promising models for the evolution of the sense of beauty.  The first is sexual selection. The other is (one interpretation of) "Baldwinian selection" which, as I will shortly explain, can be thought of as including sexual selection as a special case.  Both these mechanisms complicate in interesting ways the simplistic idea that beauty is an "adaptation". 

The basic mechanism of sexual selection is easy to grasp. The background is this:  (Consider only mammals for simplicity.)  Females are limited by the number of gestations they have time and resources for. Males can mate indefinitely often at little cost.  Since males and females must, by the laws of arithmetic, have the same average number of  offspring, it follows that the variance in offspring among males is much higher than among females.  Females must therefore "choose" mates, rather than the other way around. Now suppose females prefer, say, males with larger antlers. Providing large antlers are passed on to offspring, they will proliferate along with the females' preference.  This sets up a positive feedback loop, launching a process which can result in an increasingly burdensome handicap. The Irish Elk may have met with extinction in precisely in that way. (FIG 5) 

So much for the mechanism.  But what about the original female preference that set the mechanism in motion?  There are two hypotheses in the literature (Cronin 1991): the first is that the original preference is due entirely to chance: Cronin calls it the "good taste" hypothesis. The other is that the preference is triggered by a genuine advantage that was originally signalled by the large antlers:  the "good sense" hypothesis.  Notice that it makes no difference to the efficacy of the actual mechanism, but at first sight, at least, it makes the claim that "beauty is an adaptation" feel very different.  If it is indeed just an accident that sets off the sexual selection, the trait in question—the peacock's tail, the elk's antlers, the giraffe's neck—may look beautiful to us, but unlike the flowers' colour it can't claim to serve any purpose beyond itself.  In some cases where it comes into direct conflict with other selective pressures, beauty may be faced with the options of becoming lethal or inverting its own standards. 

Here is an example. Elephant tusks are a product of sexual selection. But the animals with the largest tusks have been at greatest risk from ivory hunters.  As a consequence, Steven Jones relates that a third of adults are now tuskless (Jones 1999, 78). Tuskless elephants have come up from behind, as it were, in the race for sex. One can imagine that in time tusks will have been bred out of the population altogether. In consequence of the selective pressure introduced by the ivory hunters, tusked elephants will seldom survive to leave offspring, and the most successfully breeding females will be those (if the genetic variance has not bred such genotypes out altogether) with a positive aversion to tusks. In this way, one will have found an evolutionary equivalent to fashion.

Is that what we mean by beauty?  The fact that standards of fashion are both changeable and arbitrary doesn't disqualify it, but it does locate the biological aspect of beauty not in any specific standards, but rather in the fact that there are standards, relating to positional values. These are values that depend not on any intrinsic qualitative characteristics, but on the current distribution of traits in the population.  "Evolutionists and art historians agree that fashion is all about status."  (Ridley 1993, 301). On the other hand, there is evidence that insofar as human beauty is concerned there is a good deal of intersubjective agreement about what is beautiful, and that standards change slowly. (Etcoff 1999).  And of course that is just what we would expect, even if beauty is purely an effect of coordination on intrinsically arbitrary features, like the colours of flowers: if it happens at the glacial pace of evolution, it will seem entirely stable on the scale of social science.

Nevertheless Turner is not alone in speculating on the specific nature of beauty, and in claiming that this can be explained in terms of "good sense" rather than "good taste."  If he is right, we should also expect him to be right in predicting that there will be found a "natural classicism" of taste, together with a general rejection of "modernist and postmodernist" taste.  And indeed there is an intriguing project which appears to support this view: the Komar & Melamid Dia project has canvassed viewers in many countries including the USA, Portugal (not Spain, unfortunately) France and China, and elected on the basis of this research the best and least liked pictures. We'll see in a moment some reasons for skepticism on this point; meanwhile, you can get the flavour of the Komar & Melamid site, which can be visited at http://www.diacenter.org/km/index.html, from this selection of illustrations showing most and least liked pictures in Turkey (FIG 6), Portugal (FIG 7);   China (FIG 9):  Kenya (FIG 9)  and the USA (FIG 10)

Baldwinian Selection.

Turner and others have also suggested that art and the sense of beauty have evolved under the joint influence of culture and natural selection.  Collaboration between genes and culture, however, is not the unproblematic thing it seems.  A cultural pressure for a certain result may take the pressure off natural selection rather than act in synergy with it. Consider the relation between the incest taboo and biologically conditioned incest avoidance.  Grant, for the sake of argument, that the bad consequences of incestuous matings in the past have selected for a gene making for strong incest avoidance. But imagine that this is complemented by a socially conditioned incest taboo so strict that incest simply never occurs.  Unless incest avoidance has been definitively fixated in the population (which would make the incest taboo otiose), we must surely expect that after a few generations the gene for incest avoidance, in the absence of continued maintenance by selection, will have become less rather than more common. Far from reinforcing one another, the genetically conditioned avoidance and socially conditioned taboo will actually undermine one another:  A perfectly effective genetic disposition would leave no room for the taboo; and conversely, once a powerful taboo has existed long enough, any easing of the taboo would then produce a sudden rash of incest, showing that incest avoidance has actually been weakened. 

This example shows that one can't take for granted that social and evolutionary forces will necessarily make for the same result. Baldwinian selection is a mechanism that has attracted attention because it appears to mimic Lamarckian inheritance of acquired characteristics. Just how it is supposed to work is somewhat controversial; but here is the most plausible reconstruction, derived from (Deacon 1997, 322-334).   (FIG 11 ) 

It starts with some innovative form of non-instinctive behaviour, which causes a change in the environment. That change, in turn, generates new selective pressures that favour genetic dispositions for certain types of behaviour. When the behaviour at the end of this loop is of the same type as the behaviour at the beginning, we may have something that looks rather like an episode of Lamarckian evolution:  a particular (group of) organisms’ choosing a certain mode of behaviour leads to a predisposition for that type of behaviour to be coded in the genes.  It is easy to see why this process is sometimes referred to as "niche construction" (Griffiths 2002): it involves a feedback loop that begins with some behavioral change, inaugurated, again, either by chance or because of some existing selective advantage and results in a modification of the environment which in turn favours selection for the behaviour in question. 

It is just this "niche creation" aspect that is of interest for the development of the sense of beauty. Sexual selection also involves a positive feedback loop, in which the hypertrophy of a certain trait (call it B for beauty) results from the mate's preference for B. Thus sexual selection is a special case of Baldwinian selection. It favours the survival value of B, by favouring the predisposition to choose it on the part of potential mates.  This propagates both B and the disposition to choose it.  As we saw, sexual selection need not result in any trait that is in the ordinary sense adaptive, and that is generally true of other forms of Baldwinian evolution. But they are of interest in this connection because they could be said to create beauty, in a way that accounts both for the temptation to call it "adaptive" (since once the mechanism has got going, beauty will actually promote fitness), and also for the intuition that the whole point of beauty is that it is gratuitous (since there was not in the first place any reason to prefer long over short necks, large over dainty antlers, and so forth.)

Baldwinian selection, then, including sexual selection as a special case, is well suited to the explanation of how some capacities will be favoured both by genetic selection and by cultural change, mutually reinforcing each other.  It's clear that art modifies our environment. Or at least artifacts do.  Did those, in turn, create an environment sufficiently different from the previous niche occupied by the population to result in novel selective pressures?  If so, one might surmise that the capacity for art will have begun with some random change, but that this event set a directional course for a while. In the generally non-directional pattern of evolution, this is one of the few patterns that actually constitutes an exception.  Needless to say, no directionality carries a guarantee of continuation. At any time, the path could be interrupted in any number of ways, including crashing headlong, like the Irish Elk, into a fatally maladaptive extreme.  While it lasts, however, something like "beauty" is being created, and the production of things experienced as beautiful constitutes art. 

At this level, beauty is essentially the phenomenal trace of a solution to a coordination game. But Turner offers a far more daring second thesis: Besides the sort of self-made beauty that is set up by Baldwinian or sexual selection, he thinks "the experience of beauty is a recognition of the deepest tendency or theme of the universe as a  whole" (Turner 1999, 124). 

Among the "themes" in question is "self-similarity" or fractal structure, a feature  found often in nature where the shape of large shapes recurs indefinitely many times in smaller parts. 

Turner might find encouragement from a striking study of Jackson Pollock's drip paintings.(FIG  12) ,  which found them to exhibit a chaotic fractal structure, constituting "a direct expression of the generic imagery of nature's scenery".  (Taylor, Micolich and Jonas 2000, 137).  Pollock's drip paintings, it turns out, share the property of "statistical self-similarity" characteristic of natural scenery (p. 140)  This means that patterns at different degrees of magnification, while not identical, "are described by identical statistics".(ibid.).  One can get a feel for what's involved here from this illustration, showing, above a 2.5m section of a Pollock painting, a 10cm section of snow on the ground and a 50m section of forest .(FIG 13) .

Notice that there is no question here of Baldwinian selection or of arbitrary solutions to coordination problems.  Instead the claim is that the very workings of our aesthetic sense are favoured by natural selection because they mirror and thus contribute to our knowledge of the deep structure of the world, which Turner argues exhibits, besides a preponderance of fractal structures or "self-similarity", a number of other wholly general features.  The complete list is the following: (Turner 1999, 126-127):

·Unity in multiplicity

·Complexity within simplicity

·Generativeness and creativity

·Rhythmicity

·Symmetry

·Hierarchical organization, 

·Self-similarity. 

This is, to be sure, an impressive list. But it's not clear that Turner is in a position to claim that these seven features are in the world, as opposed being features that our brains happen to be good at using to make sense of the world.  In the sixth Meditation, Descartes warned that the representations of the senses might be fulfilling their natural purpose without necessarily  giving us any information about the world as it is: "the proper purpose of the sensory perceptions given me by nature is simply to inform the mind of what is beneficial or harmful.... But I misuse them by treating them as reliable touchstones ... about the essential nature of the bodies located outside us.....''. (Descartes 1641/1986, 83)  It's not clear that the apparent constancy of the world's forms reflects more than precisely the methods favoured of our brains. 

7.  Art's functions in the brain: Ramachandran

A more modest but still powerful hypothesis is—in the words of (Myin 2000, 45) that "given that both the brain and the artist are in the same business of representation, perhaps the overt representing of the artist is highly constrained  by how the brain represents the visual world internally." 

This is the avenue pursued by (Ramachandran and Hirstein 1999).  Their claim is not to show that art reveals the ontological structure of the world, but rather to explore that other "deep structure" of beauty that reflects the way the brain constructs a representation of the world. 

Prominent among their "laws of beauty", are the peak shift principle and the principle of perceptual grouping.

I begin with the latter, because while it is reminiscent of Turner's principle of "unity in multiplicity", its significance stands in interesting contradiction to Turner's claim of ontological representation. Instead of resting on some claim about nature intrinsic "unity in diversity", it focuses on how we perceive unity in diversity, or what neuropsychologists call the binding problem.  It also stresses the emotional concomitants of the exercise of very specific mechanisms responsible for constructing an object out of the data presented to the senses. Here are two examples of images that at first sight look like a meaningless jumble but can be seen as coherent pictures. The integration and recognition, most people find, are intrinsically pleasurable.

(FIG14:) 

Another whole "genre" which supports his argument is to be found in the computer-generated elaborations on Bela Julesz's discovery that random dots appropriately displaced generate three-dimensional images without any further cues^[4] (Julesz 1971).  Here are two example that don't employ random dots, but generate a nice effect: (FIG 15);  (FIG 16).

The striking fact about these examples is that they generate distinct aesthetic pleasure, despite the utter triviality of their subject matter. The pleasure is taken merely in the fact that a coherent three-dimensional form appears.  Watching someone coming to see the three-dimensional form in these pictures is quite instructive: they commonly  give out very strong literal expressions of the  "Aha! experience." 

This sort of process, as opposed to the others I have surveyed, no longer puts any direct premium on representation. Instead it deals with the tools and mechanisms of representation.  Hence, pace Turner's commitment to "natural classicism", it would lead us to expect that people can take pleasure in all kinds of "abstract" art, insofar as these play on the capacities of the brain that are involved in the immensely complex task of constructing representations (Hoffman 2000). 

This applies to all the other "laws of beauty" that R&H propose: the isolation of visual outlines; contrast extraction; symmetry; metaphor; and the "generic principle" that tends to reject interpretations that rely on rare coincidence.  I have time to comment only on more. 

The peak-shift effect is summed up in the slogan 'All art is caricature' (Ramachandran and Hirstein 1999, 18).  It refers to the tendency of the visual system to learn a tendency rather than a characteristic, and thereby to react more strongly to an exaggeration of the original characteristic than to the characteristic itself. Once a rat has been trained to respond to a rectangle of ratio 3:2, "the rat's response to a rectangle that is even longer and skinnier (say, of aspect ratio 4:1) is even greater than it was to the original prototype on which it was trained". (ibid.)   It's intriguing to speculate that this trait lies at the basis of the directionality of sexual selection: just as the rat reacts to the "more rectangular" shape rather than to the prototype, so the chooses not the mate whose neck is the precise length that has proved "successful" in the past, but the longest neck. In matters of beauty in representations of the human body, R&H suggest, "idealization" amounts to exaggerations of characteristics regarded as "essential", as in this illustration of the Goddess Parvathi (FIG 17). 

These cognitive mechanisms of chunking and "binding" information, as well as others typically exercised in recognizing objects, generate limbic activation which is in turn responsible for the emotional impact that they are able to have even in the absence of otherwise emotionally interesting cues. (The Parvathi picture is hardly a fair example, but the Pollock will do.)  And the "mystery" of art is surely due in part to the fact that these mechanisms work without our having the slightest consciousness of their existence, let alone the manner of their functioning.  The Pollock example is particularly striking: for I venture no one before Taylor et al. had any idea that fractal structure might have anything to do with the impact of Pollock's drip paintings:  and although viewers would surely distinguish and have a preference between merely the merely random lines of a  and the fractal ones of b and c (whether produced by machine or by Pollock), it is equally sure that virtually no one could have diagnosed the source of the aesthetic preference. (FIG 18)

All these images exemplify a principle that is distinctive in Ramachandran and Hirstein's hypothesis: namely that aesthetic pleasure appears to have at least the Cummins-function of reinforcing the exercise of specific exercises of the most important mechanisms the brain employs to construct our visual world. "Every man," said Aristotle at the very beginning of the Metaphysics, "desires to know.  An indication of this is the delight we take in our senses ...; for even apart from their usefulness they are loved for themselves; and above all others the sense of sight.." (Met. I.1) The prescient phrase here is "loved for themselves".  The senses, it appears, generate pleasure even by their gratuitous exercise out of context. In the same vein, though usually with the much higher degree of self-consciousness, painters in modern times have abandoned realistic interpretation not so much because they have given up on mimesis, but because they have turned to exploring for their own sakes the shapes, colours, textures, rhythms, and other effects that enter into representational art.  From the present vantage point , though modern art  appears to constitute a spectacular counter-example to Turner's principle of "natural classicism", the exception is more apparent than real. 

8.  Conclusion: Four Functions of Art.

In this brief survey, it has emerged that the biological functions attributed to art fall into four kinds.  We can think of these as delineating four types of beauty:

a)beauty as a solution to a coordination problem: colours are arbitrary signals, but the pleasure we take in them is related to the fact that they work to guide us. As is always the case in evolution, an informational problem is solved, subject to immensely complex causal constraints. 

b)beauty as the phenomenal correlate of non standard mechanisms of selection, namely Baldwinian selection and its more specific (and more familiar) species, sexual selection. This might in turn be one of two types:

-  purely positional, where the feature selected for acquires value only in relation to others in its range as they occur in the population; or 

-  as an index of some other desirable quality, such as fertility or strength.  While this makes no difference to the actual mechanism of selection, we saw that many people thirst to find something that will count as "real" biological functions of art, as if that alone could justify the importance we attribute to art.  But I have been skeptical of that demand:  there is, I maintain, no connection between the type of origin of a trait and the importance it can legitimately claim in our lives. 

c)beauty as a reflection of the innate structure of the universe.  This is Turner's grand scheme.  The metaphysical flavour of that proposal does not inspire my sympathy, but it is plain that it corresponds to something very real in the phenomenology of art and beauty. And while that does not warrant Turner's bold leap into the ontological plane, it does seem ripe for a Kantian reinterpretation as a hypothesis about the deep structure of the mind.  In terms of contemporary cognitive science, this means we should attend to the mechanisms employed by the brain to structure the world of our experience. That is precisely what the fourth proposal aims to do:

d)beauty as the pleasure taken in the exercise of the cognitive mechanisms of the brain. 

This last proposal stresses a type of "beauty" the "function" of which is likely similar to that other close kin of art, play; but it is play located in the very heart of the brain's cognitive modules.  Each of the mechanisms detailed by R&H (and many more that are rapidly emerging into the scientific literature) is a good bet for a module with a proper function—whether by adaptation or by exaptation. That each is tied, as R&H argue, to the limbic system and therefore to the emotional system, strongly suggests that the influence of a further selection pressure might have been at work to cement those mechanism into our cognitive toolbox.  Just exactly why some have become conscious while most can only be discovered by painstaking experiment, is an intriguing question. But I shall resist the temptation to indulge in further speculation. 

References

Abrams, M. H. 1953. The mirror and the lamp. Oxford: Oxford University Press.

Bowlby, J. 1973. Attachment and loss. Vol. 2, Separation: Anxiety and anger. New York: Basic.

------. 1980. Attachment and loss. Vol. 3, Sadness and depression. New York: Basic.

------. 1982. Attachment and loss. Vol. 1, Attachment. New York: Basic.

Cronin, H. 1991. The ant and the peacock: Altruism and sexual selection from Darwin to today. Cambridge: Cambridge University Press.

Cummins, R. 1975. Functional Analysis. Journal of Philosophy 72:741-64.

Deacon, T. W. 1997. The symbolic species: The coevolution of language and the brain. New York: Norton.

Descartes, R. 1986. Meditations on First Philosophy with selections from the objections and replies. Trans. J. Cottingham, R. Stoothoff, and D. Murdoch. Cambridge: Cambridge University Press.

Dipert, R. R. 1993. Artifacts, art works, and agency. Philadelphia: Temple University Press.

Dissanayake Ellen. 1992. Homo Aestheticus: Where art comes from and why. New York: Free Press.

------. 2000. Art and intimacy: How the arts began. Seattle and London: University of Washington Press.

Dissanayake, E. 1999. “Making Special”: An undescribed human universal and the core of a behavior of art. In Biopoetics: Evolutionary explorations in the arts, ed. B. Cooke and F. Turner, 27-45. Lexington, KY: ICUS.

Etcoff, N. L. 1999. Nature 405:139The survival of the prettiest: The science of beauty. New York: Doubleday.

Godfrey-Smith, P. 2002. On the evolution of representational and interpretive capacities. Monist 85(1):50-69.

Gould, S. J., and E. Vrba. 1982. Exaptation: A missing term in the science of form. Paleobiology 8:4-15.

Gould, S. J., and R. L. Lewontin. 1979. The spandrels of San Marco and the Panglossion paradigm: A critique of the adaptationist programme. Proceedings of the Royal Society of London B 205:581-98.

Griffiths, P. 2002. Beyond the Baldwin effect: James Mark Baldwin’s ‘social heredity,’ ‘epigenetic inheritance’ and ‘niche construction.’, in Learning and evolution: the Baldwin effect reconsidered.  eds B. Weber and D. Depew. Cambridge, MA: MIT Press.

------. 2002. What is innateness? Monist 85(1), January.

Griffiths, P. E. 1998. Functional analysis and proper functions. Reprinted from: British Journal for the Philosophy of Science 44:435-52 . References are to the reprint in Nature’s purposes: Analysise of function and design in biology, vol. 44, eds C. Allen, M. Bekoff, and G. Lauder. Cambridge, MA: MIT, A Bradford Book.

Hoffman, D. D. 2000. Visual Intelligence: How we create what we see. New York: London: Norton. Http://aris.ss.uci.edu/cogsci/personnel/hoffman/.

Jones, S. 1999. Darwin’s ghost: The Origin of Species  updated. New York: Doubleday.

Julesz, B. 1971. Foundations of Cyclopean perception. Chicago: University of Chicago Press.

Millikan, R. 1984. Language, thought, and other biological categories. Cambridge, MA: MIT Press, A Bradford Book.

Millikan, R. G. 1989. In defense of proper functions. Philosophy of Science 56:288-302.

Murray, L., and C. Trevarthen. 1985. Emotional regulation of interactions between two-month-olds and their mothers. In Social perception in infants, ed. T. Field and N. Fox, 177-98. Norwood, N.J.: Ablex.

Myin, E. 2000. Two sciences of perception and visual art: Editorial introduction to the Brussels papers. Journal of Consciousness Studies 7(8/9):43-55.

Ramachandran, V., and S. Blakeslee. 1998. Phantoms in the brain:  Human nature and the architecture of the mind. New York: William Morrow.

Ramachandran, V., and W. Hirstein. 1999. The science of art. Journal of Consciousness Studies 6(6-7):15-51.

Ridley, M. 1993. The Red Queen: Sex and the evolution of human nature. New York: Macmillan Viking.

Sober, E. 1984. The nature of selection. Cambridge:  MIT Press.

Stern, D. 1985. Intepersonal world of the infant. New York: Basic books.

Taylor, R. P., A. P. Micolich, and D. Jonas. 2000. Using science to investigate Jackson Pollock’s drip paintings. Journal of Consciousness Studies 7(8/9):137-50Metaphysics. Garden City: Prentice-Hall.

Turner, F. 1999. An ecopoetics of beauty and meaning. In Biopoetics: Evolutionary explorationsw in the arts, ed. B. Cooke and F. Turner, 119-37. Lexington, KY: ICUS.