Philosophy of art assignment

V.S. Ramachandran

Sharpening Up ‘The Science of Art’ An Interview with Anthony Freeman

Introduction

Freeman: The article ‘The Science of Art’ by William Hirstein and yourself

(Ramachandran and Hirstein, 1999) has proved one of the most stimulating ever

published in JCS. Criticisms of it abound, so to focus your response I will put to

you questions that summarize the chief points made by our formal commentators

[see Journal of Consciousness Studies, 6 (6–7), 1999, and 7 (8–9), 2000] and also

pick up other questions that are frequently asked about it. But first I should like to

know what motivated you to write the paper — and did you anticipate the reaction

it would provoke?

Ramachandran: We mainly did it for fun. Also we hoped the essay would serve

to generate a useful dialogue between artists, neuroscientists, perceptual psychol-

ogists and art historians — to bridge C.P. Snow’s two cultures. The article was

intended to be whimsical, provocative and slightly tongue-in-cheek, and to serve

as a starting point — it certainly wasn’t intended to be a complete theory of art

(even assuming there is such a thing!). Judging from the enormous number of

comments received — both in the journal and at subsequent meetings on the topic

— it seems to me that our article has at the very least accomplished our minimum

objective: to make some specific experimental predictions and so to stimulate a

lively debate.

Freeman: What is your overall reaction to the tone and content of the comments

and questions that have been thrown up?

Ramachandran: As one would expect, with such a wide range of commentators,

some are favourable and others frown on the whole enterprise — while at the

same time disagreeing among themselves or even contradicting each other. My

general impression has been that scientists (e.g. psychologists and neurologists)

find the essay stimulating whereas some art historians don’t. But I was pleasantly

Journal of Consciousness Studies, 8, No. 1, 2001, pp. 9–29

Correspondence: V.S. Ramachandran, MD, Center for Brain and Cognition, UCSD, La Jolla, CA 92093-0109, USA. Email: [email protected]

surprised to find that many practising artists also find the article to be both inter-

esting and provocative. (‘It makes what we do seem less “crazy” or ad hoc —

indeed it legitimizes what we do,’ was the way one artist put it while commenting

on a lecture I recently gave at the Getty [Ramachandran, 2000b].)

I especially enjoyed reading the clear, cogent, and insightful arguments of

Richard Gregory, Colin Martindale, Jaron Lanier, Julia Kindy, Ron Mangun,

Jennifer McMahon and Amy Ione and would like to thank them for taking the

time to respond.

When reading some of the critical commentaries I was reminded of a story

about the great German physicist Hermann Von Helmholtz, the first scientist to

propose the First Law of Thermodynamics — that energy can neither be created

nor destroyed. Even though the ‘law of conservation of energy’ was starting to be

widely accepted it was violently opposed by many biologists who argued that the

law applied only to non-living systems but not to living things. So Helmholtz set

up an apparatus to measure the heat output of a living muscle preparation and

showed that it did obey the first law just as well as an ‘ideal’ or perfect machine

should. Amazingly, when he set up a demonstration of this at a scientific meeting

in Germany, even many contemporary physicists (including Joule and Carnot)

refused to believe the evidence before their eyes. Helmholtz wrote home to a

friend: ‘Despite the evidence staring at them, not a single member believed a

word of what I said. From the vehemence of their denials and denunciations I now

know I must be right.’ The vigorous opposition to Helmholtz was mainly the

result of the then prevailing strong anti-reductionist stance in biology —

especially the superstitious belief that living things were imbued with a mysteri-

ous ‘vital essence’ that rendered any application of scientific laws — such as the

first law — inapplicable to them. One discerns a touch of the same endearing

naivete in the responses of many social scientists to our essay on the science of

art, for example in the comments of Ruth Wallen (who describes herself as a

‘practising Buddhist’) who seems worried that a scientific analysis of art might

detract from its spiritual dimension! (See further on p. 22 below.)

Incidentally, I am amazed at the willingness of Dr Goguen to allow ‘Donnya

Wheelwell’ to publish her piece under a pseudonym — a practice that is unheard

of in respectable academic circles (and despite his telling me over the phone that a

pseudonym would not be used). If Ms Wheelwell really had something interest-

ing to say, why hide behind a psudonym?1

Freeman: You have said that you mainly wrote the article ‘for fun’, but I don’t

entirely believe that. I sense that you have a passionate and deeply serious under-

lying commitment to the project you have begun.

Ramachandran: In publishing ‘The Science of Art’ we had one main goal in

mind, namely to stimulate an interdisciplinary discussion and debate between

neuroscientists, social ‘scientists’, artists and art historians on the meaning of art

and the quest for artistic universals.

10 V.S. RAMACHANDRAN

[1] See pages 3–6 above for a discussion of editorial policy on unsigned contributions — Editor.

Freeman: Doesn’t the attempt to discover ‘artistic universals’ negate the very

essence of art, which is after all a celebration of individuality, not about what’s

common across people and cultures?

Ramachandran: My spelling out universal ‘laws’ of aesthetics does not negate

the originality of the artist. I am merely asking what rules of thumb the artist con-

sciously or unconsciously deploys. But which laws a given artist chooses to

emphasize and how effectively she does it is entirely up to her skill and original-

ity. Monet was a master of introducing peak shifts in colour space whereas Henry

Moore was probably tapping into and exploiting form primitives — analogous to

the stick with three stripes. By way of analogy consider two ‘universal laws’ in

poetry — metaphor and rhyme. Spelling out these laws does not detract from the

fact that Shakespeare was better at deploying them than any other human being.

(The only difference is that for universals in poetry, such as rhyme, it is hard to

find the evolutionary rationale or neural underpinnings in the manner I have tried

to do for the universal laws of visual art.) An even better analogy is with

Chomsky’s discovery that there is a ‘universal grammar’ underlying all human

languages despite their surface diversity. His ideas have enormously enriched our

knowledge of language, but no one would argue that the existence of such univer-

sals detracts from Shakespeare’s originality.

Freeman: How did you — a neuroscientist — first become interested in the

question of artistic universals? And didn’t you feel the need for a formal training

in art history or the philosophy of aesthetics before embarking on such a venture?

As things are, you surely run the risk of making simple mistakes that give your

opponents an excuse to brush aside your more incisive remarks.

Ramachandran: My current work is in behavioural neurology but most of my

earlier work in the last two decades was on human visual perception — especially

its biological basis. I also enjoy and study Indian art as a hobby. So it was quite

natural for me to become interested in artistic universals and the underlying neu-

ral substrate. I became especially intrigued by the word rasa that appears in

ancient Indian (Sanskrit) art manuals. The word is hard to translate but roughly

means ‘Capturing the very essence of something in order to evoke a specific emo-

tion or mood in the viewer’s brain’. I realized that to understand the neural basis

of art one needs to understand rasa.

My lack of expertise in art history is both a disadvantage and an advantage. A

disadvantage, perhaps, because it may lead me to ignore stylistic historical trends

and the important cultural dimensions of art. An advantage, because it is only by

deliberately ignoring the variations imposed by culture that one can begin to

approach the question of whether or not there are universals and it’s easier to

ignore something if you don’t know anything about it!

My point is not that culture is not important but that it’s what most people study

— it’s called art history. On the other hand almost no real progress has been made

in understanding universal principles that cut across cultures and that’s what

interests me as a scientist. Even if only 10% of art turns out to be lawful it’s that

SHARPENING UP ‘THE SCIENCE OF ART’ 11

10% that interests me, not the unlawful 90% that interests sociologists and

historians. After all there are thousands of stodgy, boring inaccessible tomes on

art history, styles, fads, etc., gathering dust on bookshelves, but very little that is

intelligible and interesting about the universals.2 So while I plead guilty to ignor-

ing the complexities imposed by culture, I would point out that my decision to do

so was deliberate. And I would argue that my ignorance of art history (especially

Western art) is an advantage since it has perhaps allowed me to approach the

problem with fresh unprejudiced eyes.

Freeman: But why is this debate so important to you?

Ramachandran: The history of science may be seen as a long succession of

attempts to banish vitalism from our understanding of Nature. People said that

living things could never be understood in terms of chemistry because there was a

mysterious élan vitale, ‘entelechy’ or vital spirit, but the discovery of DNA’s

structure changed all this. Likewise it was widely believed that even though we

may someday understand all the functions of the brain, we can never explain the

‘soul’ or consciousness. This challenge too has been taken up (e.g. Crick and

Koch, 1998) and few educated people now believe in a nonmaterial soul. Now

with regard to art, the easy way out is to say, ‘Well, it will always remain ineffable

and mysterious’ — as some of the commentaries on our essay imply. On the other

hand it requires guts to say, ‘Well, maybe that’s true and eventually we will come

up against an epistemological barrier, but even if that is so, I would prefer to forge

ahead in my quest for laws until I do confront that barrier.’ This is why I regard

Zeki’s efforts to understand art as so important, even though he doesn’t have all

the answers. Saying ‘it will always remain mysterious’ is the easy way out.

Prediction and Experimental Testability

Freeman: You lay great emphasis on the claim that your ideas entail predictions

that are experimentally testable. Can you elaborate on this with concrete examples?

Ramachandran: Yes, at least some of the ideas in our essay make very specific

counterintuitive predictions — and that’s what sets these apart from the vague

ideas of philosophers and even vaguer ideas of art historians.

Let me take up a challenge that would, at first sight, seem almost unapproach-

able by science: cubism. It has long been known that cubism liberates the viewer

from the tyranny of a single viewpoint by presenting two or more ‘views’ of an

object or face squished into a single plane — so as to be seen from a single

vantage point. So a Picasso portrait might show a profile face and a frontal view

simultaneously in a single image (see, e.g., Zeki, 1999). However what no art

scholar or scientist has asked is why such a simultaneous depiction of two views

should be actually more pleasing to the observer than any single view, even

though it looks ‘unrealistic’. One option would be to suggest that cubism is not all

that pleasing and that it’s the result of propaganda and culture — indeed Zeki

12 V.S. RAMACHANDRAN

[2] See below for some notable exceptions, e.g. work by Gregory, Arnheim, Gombrich, Humphrey, Zeki.

speaks of it as a ‘failure’. But following my line of reasoning about the ‘super

beaks’ that gull chicks respond to, I can make a specific physiological prediction

concerning cubism (Ramachandran, 2000a,b).

First I would emphasize that the ‘super beak’ idea is not the same as ‘carica-

ture’ or hyperbole. The key difference is that in the latter it is obvious what

dimension is being exaggerated (‘peak shift’) whereas in the former the super-

stimulus bears no obvious relationship to the original stimulus to which the

organism has evolved (or learnt) to respond. That is to say, from looking at a real

beak you couldn’t have predicted that a long stick with three stripes is the optimal

configuration for driving the neurons. And I suggest that this idiosyncratic

response is a sort of spandrel — it has something to do with neurons using certain

primitive rules or ‘shortcuts’ to detect beaks quickly and adequately with high

probability and without too many false positives. The system is adequate but not

optimal (and certainly not infallible). Indeed all that’s required is that the neurons

have a high enough ‘hit rate’ in detecting beaks that they leave behind enough

genes for the next generation of gulls. (We need a new name for this so lets call it

‘hypernormal stimuli’.)

Freeman: Is the response to the beak learnt or present at birth in the chick’s brain?

Ramachandran: It is present at birth. But my notion of ‘hyper-stimuli’ would

apply even to learnt visual patterns — e.g. specific faces. For instance, a large

proportion of cells in a certain part of the fusiform gyrus (inferotemporal cortex)

of the monkey respond to faces in a highly selective manner and not to other

things (e.g. pigs, chairs or bananas). Each cell will respond only to one view of a

particular face, e.g. ‘boss’ or ‘infant’ or ‘Monkey A’ or ‘Monkey B’, with differ-

ent cells responding to different faces. In addition there are many different cells

responding to a given face — with each cell responding to different view of that

same face, e.g. frontal view, semiprofile or full profile (Rolls, 1999). But at the

subsequent stage in processing, in the next visual area in the hierarchy, a new type

of cell is seen that will respond to any view of a given face — e.g. mother or mon-

key A or B (and different cells respond to different faces). These ‘master face

cells’ (as I call them) are of course what the monkey ultimately needs — for it

needs to detect the mother’s presence no matter what direction she is looking.

Now how do you ‘construct’ or wire up one of these master face cells? In all like-

lihood you pool the outputs of several regular face cells each of which responds to

a different view of the same face. So you take the output of all the regular single-

view face cells corresponding to a given face and have these output axons con-

verge on to a single master face neuron for that face. The net result of such a con-

vergence would be that the master cell in question responds equally well to any

view of that face — which is what you want.

Freeman: But what has all this to do with cubism and Picasso?

Ramachandran: In order to see that, first consider this fact. When a conver-

gence of axons from several ‘regular’ face cells occurs on a single master cell,

nature (or evolution) is not going to go through all the trouble of ensuring that the

SHARPENING UP ‘THE SCIENCE OF ART’ 13

convergence results in a perfect ‘OR-gate’. On the contrary it may well be that if

both views are simultaneously presented to the master cell then the converging

inputs from the two corresponding regular ‘single view’ cells may simply add lin-

early — until saturation. This means you would be hyperactivating the master

neuron in a manner that could never occur in nature (Ramachandran, 2000a,b). So

this master face neuron may scream out loud (so to speak) ‘WOW — what a face!’

and excite the limbic system correspondingly. Now the advantage with this expla-

nation is that it can be tested experimentally.

Neuroscientists at Oxford and Princeton are currently recording from both

types of cells in these very areas. My prediction is that if you find a regular face

cell, it should get excited by regular faces but not any more so by a Picasso face

(since only one of the views will excite the cell). But if you go to the master cell,

where convergence of many views occurs, then that cell will not only respond to

any individual view but even better to two views presented simultaneously as in a

cubist portrait! I am especially fond of this example because often when I give

talks a social ‘scientist’ will raise his (or her) hand and say, ‘You scientists are

reductionists. You may explain atoms and molecules using string theory, you may

explain heredity using DNA — maybe even aspects of the mind, such as memory

in sea slugs — but you can never explain a Picasso!’ Well what I have proposed

above is nothing less than a specific testable theory of Picasso, in neural terms.

Freeman: But it may turn out that your explanation is wrong.

Ramachandran: Yes — and that is the beauty of it! In that case it’s at least

wrong: one can show for sure that it is wrong! And you can’t say that for 99% of

the philosophical theories of aesthetics and art. To my knowledge this is the first

clear cut experimental prediction on cubism that has been ever proposed.

Freeman: Your enthusiam is contagious, but if your ideas about Picasso are

correct, then why doesn’t everyone like Picasso?

Ramachandran: This is an important question. The surprising answer might be

that everyone does ‘like’ Picasso but not everyone knows it! The key to under-

standing this is to realize that the brain has many quasi-independent ‘modules’

that can at times signal inconsistent information. It may well be that all of us have

the basic neural circuits in the fusiform (and especially at the next stage) i.e. the

‘master face neurons’ that would show a heightened response to Picasso, but per-

haps in many of us other ‘higher’ cognitive systems (e.g. the interpreter in the left

hemisphere) might kick in and censor or veto the output of the face neurons by

saying, in effect, ‘There is something wrong with this face — two eyes cannot

possibly occur on a profile — so ignore that signal from the master face cells even

though it is very strong’. In short I am saying all of us do like Picasso but many of

us are ‘in denial ‘about it, and as I have argued elsewhere ‘denial’ is much more

widespread than people realize. For instance polls show that roughly 95% of peo-

ple think their intelligence is ‘above average’ — a mathematical impossibility.

The finding implies that almost half of mankind is in denial about its stupidity!

Likewise I would argue that even people who claim not to like Picasso are closet

14 V.S. RAMACHANDRAN

Picasso enthusiasts and this could in principle be tested with brain imaging and

GSR — it’s not an untestable assertion.

Let me make a second prediction and that concerns the exaggerated ‘feminine

pose’ or sense of movement rhythm and dance seen in certain Indian sculptures

(see my article for examples) as well as Western sculptures, e.g. Degas. I suggest

these are hyper-normal stimuli for a system of cells called ‘mirror neurons’ in the

frontal lobes of primates. Certain cells in the monkey’s ventral premotor area

respond when the monkey performs actions such as ‘pulling’, ‘reaching for a pea-

nut’, etc., i.e. they are motor ‘command’ neurons for initiating and orchestrating

complex sequences of movement. Now Rizzolati has shown that a subset of these

— called ‘mirror neurons’ — will be activated even if the monkey doesn’t move

his hand but watches another ‘model’ monkey perform that very same action, i.e.

it’s a ‘monkey see monkey do’ neuron. Similarly, there may be neurons that

respond optimally to the adoption of certain postures by the other monkey. It was

as though the neuron was doing an internal VR simulation in order to ‘read’ the

model monkey’s ‘intention’. To create a work of art that takes advantage of these

principles, I suggest that one exploit two laws — ‘isolation’ and ‘peak shift’ — in

order to hyperstimulate E, the mirror neuron system in humans. To do this one

could create point light displays mounted on the ‘model’ human’s joints as she

performs complex movements (e.g. dance) or adopts certain poses (‘isolation’)

analogous to displays used by Johannson (1975) and then introduce postural peak

shifts or exaggeration in these displays. This would serve as a hyper-optimal

stimulus for the mirror neurons and I predict it would be aesthetically more pleas-

ing than watching a real woman adopt corresponding poses or movements.

Lastly I would emphasize that the idea that art involves exaggeration or hyper-

bole is hardly new. What is novel here is the idea that the more effective stimulus

that hyperactivates the visual cells need not bear any obvious resemblance to the

thing being depicted (e.g. stick with three stripes = beak) and may, instead, reflect

the manner in which that cell is ‘wired up’ to take certain short-cuts in processing

visual images. Indeed certain fish will respond most vigorously to a blue dot

painted on a fish of the opposite sex even though there is nothing resembling that

blue dot on the original! Perhaps abstract art taps into such form and colour primi-

tives in human vision (and the idea can be tested by experiments of the kind out-

lined above for Picasso).

Also, as we argue in our essay, both the idea of ‘peak shift’ and the idea of

‘hypernormal stimuli’ (like the stick with three stripes) can be extended to

domains other than ‘form’ — e.g. colour (Monet and Van Gogh) or movements

(in dance) — i.e. to domains to which the idea may initially seem inapplicable.

Indeed it is likely that there are visual areas (or neurons) in the brains of primates

that extract other visual primitives that are even less obvious, such as light and

shade, highlights, skin texture, etc., and one often sees renaissance paintings that

deliberately exaggerate or produce ‘hypernormal’ versions of these very charac-

teristics — e.g. an old man with anatomically impossible wrinkles, absurdly

accentuated highlights on a shiny face or eyes, rippling, grossly hypertrophied

muscles, or a pink flush that is almost too healthy (Boucher). Deeply accentuated

SHARPENING UP ‘THE SCIENCE OF ART’ 15

(and physically impossible) shading is also often used to artificially heighten the

sense of 3-D form and depth. Without these extra touches you just have realism —

not art. (And conversely if these laws are deployed inappropriately you get kitsch

art and tacky art.) Some people have argued that what is kitsch and what is sophis-

ticated is a matter of convention and culture but I disagree. The evidence against

this is that one can mature from initially liking kitsch to liking more sophisticated

art but one cannot usually slide backwards to kitsch (which implies that there is

something genuinely ‘better’ about the former). Other aspects of aesthetics,

besides art, may also exploit ‘peak shifts’. E.g. lip stick and rouge emphasize that

a potential mate is not anemic and is therefore fertile. Belladonna enlarges pupils

and glitter sprinkled on the skin looks like sweat — simulating sexual arousal.

Freeman: As an aside, your speaking of sea gulls and impressing chicks raises

the question, Do animals have art?

Ramachandran: I wouldn’t rule it out. Male bower birds in New Guinea and

Australia build elaborate and very ornate bowers decorated with brightly col-

oured berries, shiny bits of metal (Ramachandran and Hirstein, 1999), lustrous

feathers and pebbles, etc. Even a cursory examination reveals many of my laws at

work, e.g. grouping (similar coloured pebbles or berries are grouped close

together) and clusters of shiny metallic bits may be hyperactivating certain

neurons just like a Picasso face does for us (signalling the presence of water or

dew perhaps?). I don’t see all that big a difference between the creations of these

birds and much of what passes for contemporary art in Madison Avenue or Santa

Fe (although, in saying this, I do feel a bit like the little boy who said the emperor

has no clothes!).

Originality

Freeman: I have also played the role of that little boy in my own discipline of

theology, and one of the defences put up against me by the ‘establishment’ was to

claim my ideas were actually nothing new, and therefore could be ignored. You

seem to be vulnerable to a similar accusation. For instance, you open your essay

by saying that the goal of art is not realism, that is to say it’s not to copy some-

thing. Surely this is not an original observation.

Ramachandran: Of course it’s not original — we were not even trying to be

original — this was just a rhetorical opening line, nothing more. But you would be

amazed at how widespread this misconception of art was, and indeed still is. As

we say in our essay, Victorian Englishmen often criticized Indian sculptures

because of lack of anatomical realism (‘multi-armed monstrosities’ to quote one

eminent bard). For example, they thought that Chola bronzes and ‘mediaeval’

stone sculptures of women were too voluptuous — with enormous breasts and

hips — for their delicate Victorian sensibilities. How ironic, then, that it was at

just this time that corsets were introduced in England. Women even had their

lower ribs removed surgically to accentuate their bust and hips to impress those

16 V.S. RAMACHANDRAN

very men who were criticizing statues of Indian goddesses for having precisely

the same attribute — too narrow a waist!

Freeman: Does that mean you are not claiming any of your ideas are original?

Ramachandran: I am tempted to quote Gerald Edelman, who once said, ‘If I

have seen further, it is by standing on the heads of pygmies.’ On a more serious

note, our article of course builds on the earlier work of many eminent scholars,

especially Rudolf Arnheim, Ernst Gombrich, Richard Gregory and Roland

Penrose, all of whom have written extensively and eloquently about art. I was

especially intrigued by the ingenious recent speculations on art by the physiolo-

gist Semir Zeki — particularly his insight that neurons in the visual pathways

often seem to do exactly what the artist himself is trying to do, e.g. the extraction

of contrast. (Although I part company with him on certain specifics. For instance,

he says cubism was a failure whereas I present I very specific explanation of why

cubism is a success and offer a precise mechanism in terms on neural circuitry.)

But one also needs to address the question of why the artist’s ‘mimicking’ what

the neurons are doing should be pleasing to the organism. For this one needs to

borrow ideas from computational vision and evolutionary biology. A proper

answer to why requires teleology.

Freeman: We all owe a general debt to earlier research, but haven’t some of your

quite specific ideas been proposed before? For example, one of your laws is

‘grouping’. Wasn’t this well known to the Gestaltists?

Ramachandran: Yes, indeed we owe a major dept to the Gestaltists and much of

this has been expounded to modern audiences by Arnheim and Gombrich. Many

of the ‘laws’ I have described have indeed been foreshadowed by others and I cite

all those that I am aware of. One must be careful, however, not to read too much

(by hindsight) into these early glimmerings. For example one could argue that the

so-called Pythagoras theorem is already implied in the first axioms of Euclid, but

we still give Pythagoras the credit for making it explicit. (Incidentally, this theo-

rem was already known to Indian mathematicians a millennium earlier, but we

call it ‘Pythagoras theorem’ because the Indians did not have a proof — only

Pythagoras did.) Likewise the atomic theory was known to Jains in India several

hundred years before Christ, but we usually credit Dalton (who himself drew on

the ancient Greek Epicureans) because he was the first to marshall all the evi-

dence in its favour into a comprehensive scientifically testable scheme.

Secondly, we believe that some of the laws we propose (and certainly the spe-

cific experimental tests) are in fact completely new, to the best of our knowledge.

Thirdly, our goal is not so much to propose an entirely new theory of art as to

bring together strands of evidence from seemingly unrelated disciplines such as

single unit neurophysiology, ethology, perceptual psychology and evolutionary

biology. And judging from the volume of responses received, it seems to have

served as a forum for debate between scholars from all these diverse disciplines.

More specifically, I argue that in order to understand any complex mental attrib-

ute in humans — be it humour, art, dance, or sex, one needs to have in place three

SHARPENING UP ‘THE SCIENCE OF ART’ 17

cornerstones: First, the underlying functional logic (e.g. what I call ‘laws’). Sec-

ond, the evolutionary rationale, i.e., speaking teleologically, why do the laws

have the ‘form’ that they do? (e.g. evolution has wired into your brain the ‘rule’

that grouping is pleasing and attention grabbing). Third, an understanding of what

is the neural hardware in the brain that mediates the law in question. As a specific

example of these three ‘corners’ of an argument consider the following.

In the ‘law of grouping’, the functional logic is to link scattered fragments into

a whole. It’s evolutionary rationale is to help defeat camouflage and find objects

in noisy environments: vision evolved in our primate ancestors mainly to find

objects quickly and efficiently but not infallibly. And third, we suggest that as

soon as the fragments are bound, there is a synchronization of neuronal spikes of

those neurons (Singer and Grey, 1995) that fire for different parts of an object and

it is this synchrony that causes an ‘AHA’ reward signal to be sent to the limbic

system (Ramachandran and Hirstein, 1999). So while most artists, fashion

designers and art historians may be aware of the grouping law they may not be

aware of the evolutionary rationale nor of the neural mechanism (synchrony of

spikes causing a reward signal to be sent to the limbic system). And we attempt to

do this not just for grouping but for all eight of our laws.

Finally, as Richard Gregory points out in his commentary, we make novel pre-

dictions and propose specific experimental tests (e.g. our physiological experi-

ment on Picasso). Without such empirical tests, ‘theories’ of art are merely

intellectual exercises of the kind philosophers engage in.

Freeman: Aren’t some of your principles well known to practising artists? For

example, what you call ‘isolation’ or minimalism is surely very familiar.

Ramachandran: I don’t deny that some of my laws have been foreshadowed by

others and some have even been explicitly stated. But what I have tried to do is to

spell out each law in considerable detail, point out its logical consequences,

explore the evolutionary rationale of the law and (when possible) propose physio-

logical mechanisms and experimental predictions. Each of these approaches on

its own is insufficient. Let me illustrate what I mean using three of my laws,

isolation, symmetry, and ‘peekaboo’ (or perceptual problem solving), all of

which are well known to artists. We also need to avoid internal contradictions

among the ‘laws’ — e.g. isn’t the principle of isolation or minimalism (‘less is

more’) the very antithesis my other law of ‘peak shift’ or hyperbole or exaggera-

tion? No it isn’t, and I will explain why.

Artists have long known that an outline sketch can be very beautiful. Even pre-

historic cave artists knew this. Now the standard ‘explanation’ for this provided

by perceptual psychologists (see, e.g., Cavanagh, 1998) and physiologists is the

cells in the visual parts of the brain respond to change — not to uniform surfaces

— e.g. cells in area 17 and 18 respond optimally to sharp edges or boundaries but

not to homogeneous edge-free surfaces. So an outline drawing has most of the

essential information concerning the objects face and will ‘drive’ or activate the

cells just powerfully as the cells would be driven by a half tone photo, But surely

this explanation begs the question. It only explains why an outline drawing

18 V.S. RAMACHANDRAN

should be adequate or as effective as a half tone grey scale picture, it doesn’t

explain why the outline is actually more evocative. That’s where my isolation

principle comes in. Analogously, just because you find cells in the fusiform that

respond adequately to (say) a Rembrandt portrait it doesn’t follow you have

explained the neural basis of portraiture or Rembrandt. To do so you would have

to find cells that respond better to Rembrandt than to a real face and also explain

why the cell does this (in terms of both its circuitry and its evolution).

To understand the principle of isolation we need to consider the critical role of

attention, i.e. the fact that even though the brain is a massively parallel computer

with 100 billion cells, it has limited attentional resources — an attentional ‘bottle-

neck’ that allows only one stable neural representation to occupy centre stage at a

given instant. Since the critical information about (say) a nude or a face is in its

outlines, everything else on it, e.g. skin tone, hair, shading, etc., is irrelevant to its

being that particular face or nude shape and actually distracts your attentional

resources from where it is critically needed. Therefore, if the artist has introduced

‘peak shifts’ in the form domain, then he needs to throw away everything else so

that the viewer can allocate all his attention to this single domain that the artist has

exploited — and ignore the others. On the other hand if the artist is mainly intro-

ducing ‘peak shifts’ in colour space (e.g. Monet or Van Gogh) then it might be a

good idea to deliberately smudge the outline or form so that attention is spontane-

ously allocated to colour alone (and indeed that’s what Monet and Van Gogh do!).

Three experimental predictions can be made. First, people who can spontane-

ously allocate their attention to single (or a small subset of) dimensions should be

actually better at art! Thus Van Gogh’s epileptic seizures in his temporal lobes

may have actually strengthened neural connections between his visual object and

face area and the amygdala, nucleus accumbens and other brain regions involved

in gauging the emotional significance of what’s being viewed. Such a heightened

attention and emotional response to visual images may have made him a more

accomplished artist — his seizures enabling him to ‘attend’ to certain critical

dimensions more than you or I. Indeed it’s not inconceivable that there was a

selective enhancement of connections from V4 (the colour area) and the limbic

structures which would explain his preoccupation with the rasa of colour.

Freeman: Are you using rasa here to mean the same as ‘peak shift’?

Ramachandran: Rasa is at least partly that, but probably more. Newton showed

that all the colours we see can be produced by optically mixing just three ‘prima-

ries’ (red, green and blue) because we have three cone pigments in the eye that

respond optimally (but not exclusively) to these primaries. Any other colour

excites the cones in certain ratios unique for that colour. By artificially bleaching

out two of the cone pigments (e.g. green and blue) and presenting white light

(composed of all wavelengths) you can hyper-stimulate red in a manner that

you could never do in nature and the result is an abnormally heightened super-

saturated red. (One that may well signal danger and evoke a bigger GSR than any

red in the real world could.) We have no idea how ‘form’ is encoded but it’s not

inconceivable that there is a finite set of form primitives or ‘primaries’. So the

SHARPENING UP ‘THE SCIENCE OF ART’ 19

Chola bronze of Parvati is not merely a peak-shifted version of a woman but

somehow evokes the percept of an idealized ‘quintessential’ female. (The appeal

of the universal ‘Golden ratio’ may also be based on some such elementary

coding principles — as yet unknown.) One wonders if something like this could

also explain the erotic appeal of a Henry Moore statue or O’Keefe painting.

Freeman: Does the peak shift principle apply to subjects other than art?

Ramachandran: Sometimes you find yourself attracted to a woman (or man)

inexplicably — even though she may not be quite as pretty (by consensus) as oth-

ers. I think this may be because you carry ‘templates’ of your mother and perhaps

other women from your teenage years, and the new woman has a face that is a

‘hyperstimulus’ for those neurons — even though she may not resemble your

mother or other early sexual partners in any externally obvious manner. (Just as it

isn’t obvious why the stick with three stripes hyper-activates ‘beak neurons’.)

Of course there is a danger of circular reasoning here. Since we do not know

what the coding parameters of ‘form’ are for human vision, if you find something

inexplicably attractive you could at once invoke my principle — saying that the

pattern is ‘like the stick with three stripes’. In response to this I would say that my

theory may be difficult to test in practice (making it vulnerable to the charge that

it is circular) but not in principle. So once the coding parameters for human per-

ception are understood, one can begin to explain clearly why certain patterns are

pleasing to the eye (e.g. my suggestion about cubism).

Freeman: I interrupted your discussion of the enhanced artistic skills of people

who can spontaneously allocate their attention to a single dimension.

Ramachandran: Yes. Another piece of evidence for this comes from patients in

whom most or many brain areas (e.g. frontal and temporal lobes) other than those

concerned with artistic proportion (e.g. the right parietal) are functioning sub-

optimally so that attention is then spontaneously allocated to the right parietal

alone. Examples include retarded autistic savants as well as adult onset fronto —

temporal dementia — a degenerative brain disease in which the frontal and tem-

poral lobes are affected but the parietal is spared. Some of these patients show a

sudden efflorescence of artistic talent — a change that I would attribute to the

principle of isolation or minimalism.

Third, one might predict that if you record from master ‘face cells’ in monkeys

they might actually respond better to outline drawings than to half tone photos

even though the latter have ‘more information’ (assuming that the masking by

irrelevant attributes occurs earlier than the neuron’s response). The extra infor-

mation is not part of the defining attribute of the face — only the outline is. Lastly

if you obtain GSR (galvanic skin response) from humans or do non-invasive brain

imaging you should see a bigger signal (either on the palm or in the fusiform brain

area) for outline drawings of faces than would be the case with half tone photos —

and an even bigger response if the drawing is also subjected to peak shift, carica-

ture or exaggeration.

20 V.S. RAMACHANDRAN

So we started with a deceptively ‘simple’ law — isolation — and explored its

rationale and many of its logical consequences. I will now apply a similar

approach to two others laws, symmetry and ‘peekaboo’.

Symmetry is attractive, whether one is talking about a child playing with a

kaleidoscope or Shah Jahan’s immortal monument to Love — the Taj Mahal. (In

fact I would regard it as my ‘eighth law’ in addition to visual repetition or rhythm

which is the ninth law.) But why does symmetry grab our attention? There are two

possibilities. First, the goal of vision is often to discover and orient to (and iden-

tify and respond to) a camouflaged object in a cluttered visual scene. Now in allo-

cating attentional priorities the detection of biological objects is especially

important since such an object is usually a prey, predator, or mate — eliciting the

so-called ‘Four “F”s’ of behaviour — feeding, fighting, fleeing or f***ing. And

most biological objects (animals, not plants) are symmetrical, so we have an early

warning system hard wired into our brains — based perhaps on direct responses

by neurons in the early visual areas to symmetry. I suggest these neurons are

directly wired up to the limbic/reticular structures in such a way as to cause you to

‘orient’ and pay attention to the symmetrical object, i.e. your visual pathways are

hardwired to the limbic system is such a way as to make symmetry pleasing in

order to provide the incentive to direct your eyes and attention towards it. Second,

we shun asymmetrical mates because in nature asymmetry is often a sign of early

parasitic infestation and and you don’t want to mate with an anemic parasite

infested partner. For this reason, too, there may be an aesthetic preference for

symmetry. Little did Prince Shah Jahan realize that the reason he built the Taj is

the same reason he fell in love with the symmetrical, parasite-free face of his

beloved Mumtaz!

Third, consider the ‘peekaboo’ principle, a device well known to artists: a face

or a nude seen behind a shower curtain or a diaphanous veil is much more evoca-

tive than a fully visible one. Many authors (e.g. Gombrich) have noted its use but

the question is why is it beautiful and/or attention grabbing? Again, I suggest this

has to do with our ancestors having to detect prey, predator or mate in camou-

flaged, noisy backgrounds often in dim light. So when the visual system is con-

fronted with a perceptual jigsaw to link together scattered fragments to find

‘hidden’ objects — you have to wire up the visual pathways in such way as to

ensure that the visual ‘hunt’ itself is rewarding, lest the visual system ‘gives up’

too easily. I suggest that just as problem solving is rewarding to the cognitive sys-

tem (which explains things like ‘curiosity’) so also visual ‘puzzle solving’ has

evolved to be rewarding. Without this neural wiring in place your ancestors

wouldn’t have pursued a mate or game in fog! To understand this more clearly we

need to abandon the ‘standard’ model of vision as proposed by Marr (1981), i.e.

the notion that it is a sequential, hierarchical ‘bucket brigade’. (Visual images are

first decomposed into elementary features, then the features are grouped into

clusters based on Gestalt laws, the cluster is segregated from background clutter

and finally the object is recognized and an ‘AHA’: rewards is sent to limbic struc-

tures.) Instead of this we suggest that at each stage in the processing hierarchy an

‘AHA’ signal is sent to limbic structures which then feeds back to the early stages

SHARPENING UP ‘THE SCIENCE OF ART’ 21

and encourages further binding of features. And through such progressive boot-

strapping the object finally emerges (Ramachandran and Hirstein, 1999;

Ramachandran et al., 1998). It is these multiple ‘AHA’ signals that the artist titil-

lates in order to excite your visual-limbic structures more optimally than she

could using real images — and I believe this to be a novel conception of art. (One

prediction from this is that patients with Capgras syndrome resulting from a

visual–limbic disconnection should no longer be able to enjoy art — even though

they might still be able to appreciate it intellectually like an art critic or art histo-

rian might.)

Reductionism

Freeman: One of the consistent criticisms I hear is that what you are giving is a

highly reductionistic theory of art. I suspect that you welcome that assessment,

but how do you answer the allied charge that by doing this you belittle art?

Ramachandran: I have three responses to this. First, reductionism is the most

powerful strategy known to science. Second, although I provide reductionist

explanations (e.g. neurons responding to Picasso), I also repeatedly emphasize

the need for two other ‘levels’ forming the other two apices of the triangle — i.e.

the level of functional or computational logic (as advocated by Marr, Pinker,

Tooby and Cosmides, Wilson, and others) and the level of evolution. Where I part

company with functionalists like Marr and with evolutionary psychology is in my

insistence on reductionist neurophysiology as being equally important (see my

critique of functionalism, pp. 25–6 below, and the analogy with digestion). Third,

a reductionist explanation of a complex phenomenon does not eliminate the phe-

nomenon — it only explains it. For example, if I explain the ‘wetness’ of water in

terms of the physical properties of its constituent molecules, would that mean

‘wetness’ no longer exists? Of course not. Likewise, a reductionist view of what

happens in your brain when you look at art, will not eliminate your enjoyment of

art (any more than an account of the neural mechanisms of sex would eliminate

your ability to experience orgasms).

Freeman: But what about the spiritual dimension of art? Ruth Wallen was espe-

cially concerned with that in her commentary.

Ramachandran: Exactly the same answer applies. Despite a common miscon-

ception among social scientists to the contrary, explaining a phenomenon in terms

of the behaviour of its constituent components (‘reductionism’) doesn’t explain it

away. As someone who collects Indian art I am well aware of the capacity for the

highest art forms to evoke transcendental or spiritual experiences. But often this

involves the clever use of hyperbole of expression and posture and certain visual

metaphors (e.g. the dance of the Chola bronze Shiva, Nataraja, in a circle of fire

evokes a profound sense of grandeur and awe and a strong sense of the cyclical

nature of the creation and destruction of the cosmos — his dance symbolizing the

dance of the cosmos itself). I would refer Ms Wallen to the works of the art

22 V.S. RAMACHANDRAN

historian Heinrich Zimmer who has written eloquently about the imagery and

metaphor of Indian and Buddhist art.

Freeman: Even so, can something as ineffable as art really be reduced to the

skin’s resistance (GSR)?

Ramachandran: Of course not — I never said that something as complex and

multidimensional as art can be measured entirely by GSR. My point was that it

provides a starting point and Berlyne (1971) has made the same point. By way of

analogy consider IQ tests as a measure of the heritable component (little g) of

general intelligence. I myself have argued that this measure provides only a very

crude partial measure of the richness of a human beings intellect Yet no one would

deny that the test does measure something — and in fact is quite useful if you want

to get a quick and approximate measure of intellect. A person with IQ 70 is

unlikely to excel intellectually whereas a person with 130 is. Or an even better

analogy is pulse and temperature as measures of one’s ‘general health’. These two

indeces go all the way back to Ayurvedic and Greek Medicine but even in this day

and age, when we have 25 tests of liver function alone, they continue to provide a

quick and ready ‘index’ of the person’s health as he/she is wheeled into the emer-

gency room. They are crude and incomplete tests but are useful nonetheless. And

I would argue that the same holds for GSR in response to art and beauty.

Freeman: A related question — doesn’t GSR merely measure ‘arousal’ as

opposed to aesthetic response?

Ramachandran: Obviously the GSR measures only emotional arousal — it

can’t discriminate positive and negative arousal. But even negative arousal has a

place in art — as witness all the bruhaha surrounding the ‘disgusting’ art in

Manhattan recently. My point was only that the GSR provides a starting point for

measuring responses to art, not that it’s a complete measure. It’s also a more direct

measure than asking someone for a verbal description because such descriptions

are often not an accurate measure of what the person truly believes — it is

‘tainted’ and filtered through many layers of subsequent processing (e.g. your

visual centres may signal a strong positive signal from seeing a Picasso, but the

signal may be vetoed or censored by subsequent high level thought processes on

the logical grounds that ‘A woman can’t have two eyes on a profile’). Also, the

use of GSR allows us to make certain testable counterintuitive predictions, e.g.

response to a caricature of a familiar face should be greater than response to a reg-

ular undistorted drawing. One can readily distinguish this from the effects of

mere arousal due to surprise (or horror) by comparing the increase seen with the

caricature with the increase seen with a random distortion (which would be

equally surprising to look at). So, contrary to Ms Wheelwell’s ‘common sense’

expectation, our ideas predict that a simple outline doodle of a nude by a brilliant

artist — cleverly exploiting peak shifts, isolation, etc. — will, paradoxically elicit

a bigger GSR (or response in single neurons encoding such stimuli) than a half

tone photo, e.g. a playboy pinup. Deliberately simplifying a problem as a starting

point is an immensely successful strategy in science, but social ‘scientists’ like

SHARPENING UP ‘THE SCIENCE OF ART’ 23

Ms Wheelwell simply don’t understand this. (I am reminded of Jim Watson’s

derisive but brilliant and pithy quip, ‘There are only molecules — everything else

is sociology’.)

Also, words such as arousal vs aesthetic response are used very loosely by psy-

chologists when, in fact, no one has the foggiest idea of how distinct and separa-

ble they really are. Indeed I would argue that arousal is an important part of

aesthetics — although not the only part. And surely no one but a prude or feminist

like Wheelwell would deny that aesthetic appeal and sexual arousal might be

inextricably linked when viewing beautiful renderings of the human nude,

whether female or male. (And, more generally, who can deny the link between art,

Eros and creativity?) I suggest Ms Wheelwell also look at a Henry Moore sculp-

ture or a Georgia O’Keefe painting.

Finally, to get around the arousal problem altogether, one could, if necessary,

stay away from nudes and stick to faces. It is well known that everyone shows a

bigger GSR to (say) his or her mother’s photo than to a stranger. Now using a

computer it is possible to produce a caricature of the mother (by exaggerating the

difference between her and the average female face), an ‘anti-caricature’ that

takes the difference and reduces it, a skilled artist’s portrait and lastly, a random

(but anatomically plausible) distortion equal in magnitude to the caricature. If the

GSR was only being caused by a distortion-induced ‘arousal’ then all four should

produce the same GSR (since the level of distortion is the same). But if I am right

then the caricature and portrait should yield a bigger GSR (because the arousal

component of GSR would be the same for all four but the ‘hyper stimulation by

recognition’ component should be higher for caricature and portrait and the net

result might be an even higher overall response).

Freeman: It has been suggested that your ideas smack of ‘sociobiology’ and

‘evolutionary psychology’ of the kind popularized by Edward Wilson. How do

you feel about being linked to these notions, which have been the subject of sharp

— you might say vicious — criticism?

Ramachandran: If you regard art as a product of the human brain and the brain

as (at least partly) the product of evolution then there is no escaping the impor-

tance of natural selection. I would be the first to admit that untestable arm-chair

teleology and evolutionary psychology can easily lead one astray, and I myself

have criticized evolutionary psychology for these reasons. Indeed, I once wrote a

satirical essay, ‘Why do Gentlemen prefer blondes?’, which was so convincing

that many in evolutionary psychology took it as a serious candidate theory rather

than a spoof! (see Ramachandran and Blakeslee, 1998). But there is no way you

can get around the question of why grouping evolved, or why it is aesthetically

pleasing, without bringing evolution into the picture. This is where the Gestaltists

failed. They attributed grouping to mysterious autonomous ‘fields’ in the brain

instead of thinking about the neural and evolutionary basis.

Freeman: Speaking of evolutionary psychology, I recently heard a theory that

the reason why humans (especially women) like art is because it serves as a

24 V.S. RAMACHANDRAN

marker or ‘index’ of the potential mate’s superior eye–hand coordination skills,

whether to wield a paint brush or a sword. Have you come across this?

Ramachandran: Yes, this is an interesting idea. It’s been dubbed the ‘Come and

see my etchings’ theory of art. It’s as if the guy is telling the gal, ‘See how good

my eye–hand coordination is — I have good genes, come mate with me.’ In my

mind this theory beautifully illustrates the pitfalls of the claims of evolutionary

psychology — namely that either they are obvious and banal (e.g. the fact that

faeces smell disgusting because they have pathogens and its bad to eat them) or

they are untestable. If the main selection pressure came from the need to ‘display’

eye–hand coordination, then men should have evolved to be excellent at knitting

(and not many women, not even feminists like Ms Wheelwell, are likely to find

knitting attractive in a man). I guess what I am trying to say is that this theory

doesn’t explain why the ‘index’ or marker for eye–hand coordination should take

this particular form, namely art. Why not use more direct indeces like archery or

javelin throwing (which, to be sure, are attractive)? Let me add, though, that even

though I disagree with this particular idea, the author does introduce many other

ideas in his book that I do find convincing. I think the real reason we engage in art

is because detecting and recognizing objects is pleasurable to the organism and

the visual system is wired up to embody this rule. And the artist enhances his

images to more optimally titillate these rules than he could with a mere snap-shot.

A second reason might be that humans have to generate internal simulations

(what we call mental images) in order to do a sort of VR simulation of forthcom-

ing future actions, such as rehearsing a bison hunt. Now nature (or evolution) has

seen to it that such internal simulations are not perfect. They cannot actually sub-

stitute for the real thing, because if they could an organism would soon starve to

death (by imagining banquets) or fail to reproduce (by imagining orgasms). As

Shakespeare says, you cannot cloy the hungry edge of appetite with bare imagina-

tion of a feast! But if this were true how did our ancestors rehearse forthcoming

hunts or instruct their young ones about it? I suggest this is how cave art origi-

nated. The caricature-like renderings served as a less dangerous and less energy

consuming substitute for the real thing — whether for rehearsal or for instruc-

tion.3

A third reason for the emergence of art might be status and the assertion of

individuality, as eloquently argued by Pinker (1999).

Freeman: This will probably sound like heresy to a neuroscientist, but part of

me wants to ask, What does it matter what the neurons are doing? Can’t we simply

follow the functional and evolutionary logic of art — or any phenomenon — and

wouldn’t that constitute a complete and intellectually satisfying account?

Ramachandran: This stance, as you know, is called ‘functionalism’ and although

it sounds logical it doesn’t work in practice. To see why this is, instead of the

visual system consider another biological system such as digestion. If in the last

two hundred years we had only looked at the ‘output’ of the digestive system, i.e.

SHARPENING UP ‘THE SCIENCE OF ART’ 25

[3] This idea emerged from a conversation with to Diane Rogers, whom I thank.

faeces, and its ‘evolutionary logic’, how much progress would we have made?

Hasn’t our understanding been tremendously enhanced by our knowledge of the

anatomy, i.e. the liver, pancreas, salivary glands, intestinal mucosa, etc.? I would

argue that, analogously, our understanding of visual perception and art will be

enormously enriched by understanding the detailed neural circuitry that mediates

it (just as our understanding of heredity was enhanced a hunderedfold when the

structure of DNA was unravelled).

As a specific example consider our evidence about gull chicks’ responses to a

long stick with three red stripes. You cannot predict or understand this in terms of

functional logic or evolutiuonary rationale. It works because evolution often

takes ‘short cuts’, i.e. the goal is not to produce a neuron that optimally responds

only to a beak (in which case you couldn’t fool it with anything) but to wire up the

system with minimum computational cost so that it is adequate (rather than opti-

mal) for the job on hand i.e. detecting beaks — as quickly as possible. Now all of

this (and my argument concerning its relevance to abstract art and cubism) was

discovered, not by arm chair evolutionary psychology or functionalism but by

ethological experts. And a more detailed understanding will require additional

physiological expertese, e.g. recordings from the chick’s tectum, rotundum or

hyperstraiatum may reveal aspects of the neural circuitry that causes the neurons

receptive field to prefer a stick with three stripes to a real beaks, e.g. the receptive

field may be wired up to embody the rule ‘The more red contour the better, etc.’.

As a result the neuron isn’t very ‘fussy’ in its requirements and can be easily

fooled by the experimenter (just as we are ‘fooled’ by Picasso or Henry Moore)

but in nature the neuron works perfectly adequately since its never likely to

encounter a mutant seagull with a beak with three stripes or indeed a malicious

ethologist waving a stick with three stripes! The point is that one cannot predict

the efficacy of this peculiar stimulus configuration — long thin stick with three

red stripes — simply based on the organism’s functional requirements. The third

cornerstone of our triangle (neural circuitry) needs to be taken into account.

Comprehensiveness

Freeman: A number of critics have argued that your focus is very limited. For

instance, isn’t your theory largely restricted to artistic representations of the

human form, and in particular Indian art?

Ramachandran: No it isn’t. I use the human form as a convenient starting point,

mainly as a rhetorical device because the ‘laws’ are easiest to demonstrate along

this dimension. But throughout the essay I refer to (and try to explain) many other

types of art. It’s valid for all — or at least for most — of them (although perhaps

not so much for Dada, surrealism, ‘conceptual art’, etc.). I chose Indian art to

illustrate my ideas only because I am most familiar with it, and also to dispel mis-

conceptions about Indian art, widely prevalent in the West. (Many artists and con-

noisseurs have pointed out to me that after reading my essay they now enjoy

Indian art much more than before.) But I also discuss Picasso, Rodin, Henry

Moore, Van Gogh and Monet.

26 V.S. RAMACHANDRAN

It is worth noting, though, that despite its initially unfamiliar appearance,

Indian art has much more in common with Western art than people realize. If you

look at the history of Western art it went through three stages of ‘evolution’ —

with lots of overlap. Initially there was emphasis on realism — e.g. every muscle,

sinew and tendon was ‘copied’ perfectly, as in Gray’s text book of anatomy (e.g.

renaissance sculptures or ancient Greek sculpture). Later, the emphasis shifted

away from realism to semi-abstract (e.g. Rodin, Klimt, Picasso) and lastly, in this

century, to completely abstract forms (e.g. Kandinsky or Pollack). Few people

realize, though, that Indian art went through the same three stages of maturation

— Terracotta busts in the Indus valley (2000BC) are superbly realistic renderings

of the human male figure. Later in ‘mediaeval’ times (800–1100AD) artists said

‘to hell with realism’ and resorted to pleasing and whimsical exaggeration and

‘peak shifts’ (as in the erotic temple sculptures of Kajuraho and in Chola bronzes

of the south). And finally there emerged a movement called ‘Tantric art’ that used

just jumbles of abstract shapes to ‘align fields and forces’ in the brain to connect

you to God, words reminiscent of those used by Kandinsky. So Indian art went

through the same stages — realism, semi-abstract and completely abstract — as

Western art did, but a millennium earlier.

Unfortunately the ‘feminist’ Ms Wheelwell (whose intellectual cowardice for-

bids her from revealing her true name) conveniently ignores this fact and she pur-

ports to ‘criticize’ our theory on the grounds that the paper has too many pictures

of naked Indian Godesses! (I wonder if she also regards Raphael, Rubens,

Gaugin, Boucher and Rodin as pornography, or whether she views only Indian art

as pornography — if so she could be accused of ethnocentricity.)

Freeman: Why do you make a partial exception in the case of surrealism?

Ramachandran: I think there are two types of surrealism. First there is the

Magritt type that involves deliberate and whimsical violation of the rules of per-

ception (such as my law of avoiding ‘suspicious coincidences’), e.g. the rules of

occlusion, opacity. etc., as an attention grabbing ‘arousal’ device. To borrow a

linguistic analogy, the artist is playing with the syntax — rather than semantics —

of visual images. The second type is the Dali kind, which involves violation of

high level visual semantics or meaning rather than early syntactic rules. The

objects he creates are not literally impossible physically (e.g. melting clocks) but

are highly improbable and dream-like. My theory has very little to say about all

this except to point the way, since its really a form of conceptual art — one is

really talking about violation of high-level concepts rather than low-level, purely

visual rules. (And the same argument holds for conceptual art whose goal is to

visually intrigue or even shock the viewer.)

Freeman: A related criticism is that your theory is mainly about aesthetics rather

than art. How do you answer that?

Ramachandran: I have two responses to this. First, even if this were true, it

wouldn’t bother me, because a biological theory of aesthetics would be just as

valuable as (indeed more general than!) a theory of art. Second, the boundary

SHARPENING UP ‘THE SCIENCE OF ART’ 27

between art and aesthetics is not very clearly defined — I bet there is a lot of over-

lap between the two concepts — so understanding one will enhance our under-

standing of the other. (The same holds for the distinction that sets ‘design’ in

opposition to ‘high’ art.)

Freeman: At the start of this interview you said your proposals were never

intended to be ‘a complete theory of art’. So just how comprehensive is your paper?

Ramachandran: We have barely scratched the surface of the problem — it’s

really only a beginning. But I do believe that it points the way to the general form

that a future more mature theory should have. (In much the same way that Crick

and Koch’s ideas do for the study of consciousness and qualia.) More importantly

our goal was to get the wheels turning in people’s minds and to initiate a dialogue

between ethologists, artists, neuroscientists and perceptual psychologists. In this

regard, I would especially emphasize the need for the ‘three cornerstones’ —

functional logic, evolution and neurophysiology. If in addition, our ideas lead to

new testable empirical claims (e.g. in cubism, GSR, etc.) then we would be amply

rewarded. But we do have a long way to go.

Consider a simple question such as what makes a female (or male) face beauti-

ful? Galton showed that an average of many female faces tends to be prettier than

any one exemplar since it tends to average or iron out any imperfections. But we

would predict that according to our scheme a beautiful female face will result if

you subtract an average female face from an average male and amplify the differ-

ence. It makes evolutionary sense that blemishes are unattractive and would be

smeared out by the averaging process (as in Galton’s pictures). But if so, why are

some blemishes — such as beauty spots — attractive? I recently found that if

Cindy Crawford’s famous mole (near her left upper lip) is moved to the central

midline, e.g. on the forehead or tip of the nose, it looks hideous. It needs to be

asymmetrically placed. It’s also prettier if placed near a sharp facial feature e.g.

the angle of a lip or near the angle of the eye, but not so much if its far from a fea-

ture — e.g. middle of the cheek. And lastly if two moles are symmetrically placed

on either side of her nose just above her upper lip, it again looks bad.

I am saying all this only to emphasize that the aesthetic response to the precise

positioning of the beauty spot is quite lawful yet we have no inkling why such a

lawful response function exists. Or consider an even simpler question: why is it so

unpleasant to look at a picture frame on the wall that is even slightly tilted? Why

is such a tiny departure from rectilinearity so disproportionately disturbing? If we

cannot answer even such simple questions we are surely a long way from under-

standing the ‘Science of Art’. But however long it takes us, ‘the journey is sure to

be gorgeous’ (Kindy, 1999).4

28 V.S. RAMACHANDRAN

[4] I thank Diane Rogers, Francis Crick, Patricia Churchland and Julia Fuller Kindy for stimulating dis- cussions on the many fascinating topics that straddle the huge divide between art and science. I thank All Souls College, Oxford, for a fellowship that enabled me to devote time to some of these questions.

References

Berlyne, D.B. (1971), Aesthetics and Psychobiology (New York: Appleton-Century-Crofts). Crick, F.H.C. and Koch, C. (1998), ‘Consciousness and neuroscience’, Cerebral Cortex, 8 (2),

pp. 97–107. Johansson, G. (1975), ‘Visual motion perception’, Scientific American, 232, pp. 76–8. Marr, D. (1981), Vision (San Francisco, CA: W.H. Freeman). Kindy, Julie (1999), ‘Of time and beauty: Commentary on the “Science of art”’, Journal of Con-

sciousness Studies, 6 (6–7), pp. 51–3. Pinker, S. (1999), How the Mind Works (New York: Norton). Ramachandran, V.S. (1998), Paper presented at AAAS meeting on the Science of Art Ramachandran V.S (2000a), ‘The science of art: How the brain responds to beauty’, in Understand-

ing Wisdom, ed. W.S. Brown (Philadelphia and London: Templeton Foundation Press). Ramachandran, V.S. (2000b), ‘The artful brain: A neurologist looks at Indian art and rasa’, Lecture

given at the Getty Center Los Angeles California. Ramachandran, V.S. and Blakeslee, S. (1998), Phantoms in the Brain (William Morrow). Ramachandran, V.S., Armell, K. and Foster, C. (1998), ‘Object recognition can drive apparent

motion perception’, Nature, 395, pp. 852–3. Ramachandran, V.S. and Hirstein, W. (1999), ‘The science of art’, Journal of Consciousness

Studies, 6 (6–7), pp. 15–41. Rolls, E. (1999), Emotion and the Brain (Oxford: Oxford University Press). Singer, W. and Grey, C. (1995), ‘Visual feature integration and the temporal correlation

hypotheisis’, Annual Review of Neuroscience, 198, pp. 555–86. Zeki, S. (1999), Inner Visions (Oxford and New York: Oxford University Press).

SHARPENING UP ‘THE SCIENCE OF ART’ 29