This article was downloaded by: [The University of British Columbia]On: 18 March 2012, At: 23:41Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number:1072954 Registered office: Mortimer House, 37-41 Mortimer Street,London W1T 3JH, UK

Australasian Journal ofPhilosophyPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/rajp20

Creativity and constraintDavid Novitz aa University of Canterbury

Available online: 02 Jun 2006

To cite this article: David Novitz (1999): Creativity and constraint, AustralasianJournal of Philosophy, 77:1, 67-82

To link to this article: http://dx.doi.org/10.1080/00048409912348811

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private studypurposes. Any substantial or systematic reproduction, redistribution,reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or makeany representation that the contents will be complete or accurate orup to date. The accuracy of any instructions, formulae, and drug dosesshould be independently verified with primary sources. The publishershall not be liable for any loss, actions, claims, proceedings, demand, orcosts or damages whatsoever or howsoever caused arising directly orindirectly in connection with or arising out of the use of this material.

Australasian Journal of Philosophy Vol. 77, No 1, pp. 67-82; March 1999

CREATIVITY AND CONSTRAINT

David Novitz

What I regard as a slight and pretentious work of art, others may regard as highly creative.

Unless we can discover some way of resolving this sort of disagreement, our claims about

creativity will turn out to be empty. They will offer praise, certainly, but they will do so

without any defensible warrant.

In her own refreshing way, Margaret Boden has argued that there are criteria in terms

of which to identify radically creative acts; that an awareness of them enables us to

resolve the disputes that sometimes surround creativity in the arts and the sciences)

Taking my cue from her, I argue that there are indeed such criteria, but I argue as well that

Boden has not properly identified them. While her core account captures an important

class of creative acts, it does not do justice, I contend, to the full range of human

creativity. The aim of this paper, then, is to offer an alternative theory, which, I argue,

holds promise of a more comprehensive account of human creativity than that offered by Boden.

I. Tolstoy, Technique, and Creativity

There is, in Anna Karenina, an interesting excursion into questions of artistic creativity.

After Vronsky and Anna settle in the palazzo in Italy, Vronsky finds himself

unaccountably bored and takes to painting in order to fill his days: 'He had a talent',

Tolstoy writes, 'for understanding art and imitating it with accuracy and good taste and he

imagined that he possessed the real power an artist needs. '2 However, since Vronsky 'was

not inspired directly by life but indirectly by life already embodied in art, he found

inspiration very readily and easily, and equally readily and easily produced paintings very

similar to the school of art he wished to imitate' (AK, p. 463).

Vronsky's work, Tolstoy suggests, is derivative and banal, not creative. By contrast,

we are told that in the same Italian town, there lives the poor, uneducated Russian artist,

Mikhaylov, whom we first come across while he is trying unsuccessfully to sketch the

figure of a man. Dissatisfied, he searches irritably for an earlier rendition of the same

figure, believing his first attempt to have been better. Tolstoy writes:

Margaret Boden, The Creative Mind." Myths and Mechanisms (Reading: Cardinal, 1992); and 'What is Creativity' in Dimensions of Creativity, ed. Margaret Boden (Cambridge, MA: The MIT Press, 1994), Ch. 4. Wherever possible, these works will respectively be referred to as CM and 'WlC' in the text. Leo Tolstoy, Anna Karenina, trans. Louise and Aylmer Maude (Oxford: Oxford University Press, 1995 [1918]), p. 463. Further page references will be given to AK in the text.

67

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

68 Creativity and Constraint

The paper with the drawing he had thrown away was found, but it was dirty now and spotted with candle grease. Nevertheless, he took it, put it on his table, and, stepping backward and screwing up his eyes, began examining it. Suddenly he smiled and flung up his arms joyfully.

'That 's it! That's it!' he said, and taking up his pencil he began drawing rapidly. A grease spot had given the figure a new pose. (AK, p. 467)

The grease spot plainly suggests a new, let us suppose a very different way, of rendering the figure. Unlike Vronsky, it turns out that Mikhaylov is not bound by convention---by past

ways of rendering figures---but is sensitive to new, sometimes radically new possibilities, and, most importantly, can tell almost at a glance what will and will not work.

Mikhaylov's creativity, Tolstoy seems to say, resides in his refusal to be bound by

technique---by the set painterly formulae that any given style prescribes. Indeed, Tolstoy--or at least the implied author--is singularly disparaging of those painters, like Vronsky, who emphasize established techniques. Writing of Mikhaylov, he tells us that 'He knew it meant a mechanical capacity to paint and draw, quite independently of subject matter' but on his (Mikhaylov's) view 'the most experienced and technical painter [could] never paint anything by means of mechanical skill alone, if the outline of the subject-matter did not first reveal itself to his mind' (AK, p. 472). The motif, Mikhaylov appears to think, should fire the imagination, which in its turn should determine the forms and the strategies that are used in any creative work of art.

II. Boden on Psychological Creativity

What this excursion into Anna Karenina suggests is that the established techniques of painting do not only afford settled ways of representing, hence of perceiving and thinking about particular motifs; they also offer certain limitations and constraints. If we are bound by well-structured painterly conventions, we are allied to what Boden calls a conceptual

space and are bound, in consequence, to represent, see, and think about the world in specific, well-established ways that may suit some purposes but not others (CM, pp. 46-9, 63-74, 139-46; 'WIC', pp. 79-84). 3 'The dimensions of a conceptual space', Boden writes in what is a disappointingly sparse account of the notion, 'are the organizing principles that unify and give structure to a given domain of thinking' and so 'define a certain range of possibilities' ( 'WIC', p. 79).

Here Tolstoy and Boden arguably agree, for both suggest that genuinely or radically creative acts--what Boden calls psychological or P-creativity--reside in the ability of the individual to transcend and radically transform these constraints and, in the process, discover new possibilities. Although Boden allows that there are other ways of being creative, her interest is with those creative ideas that 'are surprising in a deeper way', and such radical creativity (what she calls P-creativity) is invariably explained as the transformation of the organising principles that unify a conceptual space.

3 See, as well, Boden, 'Introduction' to Dimensions of Creativity, pp. 2-3.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 69

On her view, not any alteration to a painterly convention constitutes such a transforma- tion. It is possible to alter or adjust or 'tweak' what E.H. Gombrich calls a schema for depicting a kind of object (a face, a tree, a torso, a foot), 4 and to do so creatively yet

without changing the broad methods or style of depicting ( 'WIC', pp. 79-82). 5 Such creativity, though, is never radical or revolutionary. It merely involves the exploration of an existing conceptual space, and results in the application of its fundamental principles in more or less imaginative and novel ways. But according to Boden, there is nothing very original about this; nothing that surprises or astounds. In order to be highly or radically creative, one would need to invent new principles of depicting--one would need to move, say, from bare outline drawing to perspectival projection, or from this to orthographic projection, or (unlikely as it may seem) from orthographic projection to impressionist painting. In each case, there is a break with the fundamental rules that constitute a given

conceptual space. Put differently, one might say that the generative principles involved in depicting are radically reconstituted in ways that enable the artist to overcome limitations in the previous system of depiction. Thus, for instance, perspectival projection, unlike bare outline drawing, enables one to represent depth visually. However, it does not permit accurate measurement in the way that orthographic projection does, nor does it emphasize shades and nuances of colour in the way that impressionist painting does. And

so it goes. Each style or method of depicting---each new conceptual space or generative system--furnishes the artist with ideas or techniques that are subtended by certain basic principles or rules that unify and organise the method in question, make the style

distinctive, and enable the artist to convey a different order of visual information about the same object.

Such developments in painting, Boden regards as 'P-creative'. The grease spot on Mikhaylov's sketch, we are meant to believe, suggests a radically new way of rendering

the figure of a person; one that answers to the subject matter in ways that pre-existing techniques could not allow. In the end, Mikhaylov's refusal to be bound by, and his determination to transcend and transform, established techniques results in a portrait of Anna that captures and conveys her beauty in an entirely new way, vastly different from,

and in this case manifestly superior to, Vronsky's stylized effort. As Tolstoy tells us:

After the fifth sitting the portrait struck everyone not only by its likeness but also by its beauty. It was strange that Mikhaylov had been able to discover that special beauty. 'One needed to know and love her as I love her, to find just that sweetest spiritual expression of hers', thought Vronsky, though he himself had only learned to know that 'sweetest spiritual expression' through the portrait. (ilK, p. 475).

As Boden would have it, Mikhaylov's act is P-creative because he transforms 'a conceptual space'. It is not that he is necessarily the first person to have done so. Were that the case his psychological or P-creativity would also be an all-time first, and so, as she calls it, historically or H-creative. But while an individual's H-creativity entails

4 E.H. Gombrich, Art and Illusion: .4 Study in the Psychology of Pictorial Representation, 3rd edn. (London: Phaidon, 1968), pp. 131-152.

s For more on this, see my 'Conventions and the Growth of Pictorial Style', The British Journal of Aesthetics 16 (1976), 324-337.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

70 Creativity and Constraint

P-creativity, the converse is not true, and, of course, whether a highly imaginative idea has

ever been thought before is entirely contingent and so of little philosophical interest here. 6

Boden is primarily concerned to explain the conditions under which an act that results in

an idea, a strategy, or a technique that is entirely new to the individual who thinks it, may

be said to be P-creative (CM, pp. 32-5; 'WIC' , p. 77). Such transformations, we have

seen, are revolutionary, and consist of deep organisational changes that, in Mikhaylov's

case, make it possible to convey visual information of a type that could not possibly have

been conveyed before. And it is also because such transformations are of value to people

that Boden considers them radically or deeply creative.

III. The Context of Discovery

To this point, Boden's theory seems adequate to Tolstoy's intuitions about creativity. But

she goes much further than Tolstoy, for, not satisfied with the metaphor of transforma-

tion, she seeks to tell us precisely what it is that makes an alteration to a conceptual

space P-creative. On her view, what distinguishes P-creative transformations of

conceptual spaces is that they result in ideas that couM not have been thought before

(CM, pp. 37-41; 'WIC' , p. 77). For, she says, to ask whether the thought process that

resulted in a particular idea or a technique was P-creative is to ask questions like 'Could

Constable, using the conceptual space available to him, have painted in the Impressionist

style?' or 'Could Mozart, using the generative principles available to him, have composed

in an a-tonal scale?' (CM, oh. 3). The 'could' in question is a computational 'could'; it

deals in possibilities, for what one wants to know is whether it was possible for

Constable or Mozart, using the techniques and 'generative principles' at their disposal,

respectively to have produced an Impressionist painting or atonal music (CM, p. 38). If

they could not possibly have done so then the transformations that led to Impressionist

painting, like the transformations that led to Arnold Schoenberg's atonal music, were

P-creative relative to the conceptual spaces within which Constable and Mozart were

working.

A conceptual space, then, is not just a cluster of ideas or techniques. It is a unified and

organised cluster; unified and organised, that is, by the basic rules, principles, conventions

that are presupposed by the ideas or techniques in question. Only if it is unified in this

way, are we able to determine what is and is not possible relative to it, and that a given act constitutes a transformation of it.

Compare this way of drawing the distinction with F. D'Agostino's distinction between subjective and intersubjeetive creativity in F. D'Agostino, Chomsky's System ofldeas (Oxford: Clarendon Press, 1986), pp. 174-5. Subjective creativity correlates with Boden's P-creativity; inter- subjective creativity roughly with her H-creativity. For reasons that are not apparent, D'Agostino associates intersubjective creativity with the product of a creative act; subjective creativity is associated with the creative act itself. This appears arbitrary since every creative product presupposes a creative act. My discussion focuses on the creative act, which is primary in any discussion of human creativity. Hence, the distinction between H-creativity and P-creativity finds its primary application with creative acts; its secondary and derivative application with creative products.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 71

On Boden's view, such transformations cannot be accidental. For it is only after an

artist or inventor has explored the conceptual space in question, only after she or he has

noticed its weaknesses and shortcomings, that the P-creative transformation of its

generative principles can occur (CM, pp. 46q59; 'WIC', pp. 80-2). It follows, Boden says,

'that the ascription of creativity always involves tacit or explicit reference to some specific

generative system' or conceptual space, and that one cannot decide whether or not an act

is genuinely creative without reference to that space. In Boden's view, this is the one

pervasive constraint that makes P-creativity possible ( 'WlC' , pp. 78-9). I shall call this

Boden's constraint.

It is a constraint that makes P-creativity much more than the 'chance' or 'irrational'

affair that E.H. Gombrich and Karl Popper think it is. 7 For on her view, it involves, and

must involve, hard systematic work, and always requires 'the mapping, exploration, and

transformation of structured conceptual spaces'. 8 And this, she thinks, will enable her to

show 'how computational ideas can help us to understand our own creativity' (CM, p.

135). But, as I shall now argue, the theory that she offers of P-creativity leaves certain

crucial dimensions of human creativity out of account.

IV. Computers and Human Beings

Appropriately programmed computers are well-placed to explore conceptual spaces, to

discover their limitations relative to certain goals and tasks, and to transform the

prevailing generative principles in ways that will allow them to achieve specific goals and

perform the requisite tasks. For this reason, Boden's account of human P-creativity

applies quite well to some computer programmes. What is not nearly so clear is that her

explanation applies as neatly to human beings. People are less able than computers to

abide by Boden's constraint, not least because of the difficulty they experience in

absorbing and integrating endless flows of information, but also because the flair and the

genius that we associate with human creativity seems to be voided by Boden's

requirement that humans need to be acquainted with all the possibilities and limitations of

the domain of thought relative to which they are creative. This would make of human

creativity the plodding, laboured and monumentally dull affair that it plainly is not.

The fact, too, that humans are social animals and that close familiarity with a complex

conceptual space sometimes involves considerable investment in an associated social

space, results in a range of rather speeial constraints on human creativity; constraints to

which computational programmes are not Subject. As social animals, human beings

depend on others for their well-being and have a strong interest in maintaining their

7 See, Gombrich, Art and Illusion, pp. 159-161. Like Leonardo, Gombrich believes in 'the power of "confused shapes", such as clouds or muddy water, to rouse the mind to new inventions' (p. 159). Karl R. Popper is even more explicit. On his well-known view, stated in The Logic of Scientific Discovery (1959) (London: Hutchinson, 1972), pp. 31-32, the process of invention has nothing to do with epistemology since questions about how we invent or discover new ideas are not concerned with the justification of our knowledge claims. On his view, 'there is no such thing as a logical method of having new ideas . . . . My view may be expressed by saying that every discovery contains "an irrational element" or a "creative intuition".'

8 Boden, 'Introduction' to Dimensions of Creativity, p. 6.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

72 Creativity and Constraint

networks of support. In effect, this furnishes a major disincentive to human creativity that

can be overcome only if humans contrive to distance themselves from some of the social

consequences of their creative endeavours. Boden's account, we know, requires that

P-creative thinkers should have explored, and should be maximally acquainted with, the

conceptual spaces that they transform, but this ignores the fact that people are reluctant to

concede that the body of knowledge that they have acquired often with difficulty, and that

now affords them status in a community, could be importantly deficient. Still more, they

are often sentimentally attached to their teachers and collaborators to whom they have a

range of obligations that arise out of past assistance. Close acquaintance with a complex

conceptual space thus carries with it very definite social expectations, some of which

demand one's intellectual allegiance and one's loyalty. This, according to Jonathan

Schooler and Joseph Melcher is why humans are often P-creative precisely when they

have not fully explored, and are not steeped in, a particular conceptual space?

Contrary to what Boden says, it just is not true that radically creative human beings

must always have explored and will always be familiar with the conceptual spaces that

their ideas transform. Sometimes a chance remark, an image, a shape, a dream, encourage

people to entertain new possibilities that cut across domains of knowledge and expertise

that they have not fully explored; sometimes, too, the weight of those domains, the

pressure of orthodoxy, prevent them from noticing new possibilities, new ways of doing

and conceiving. Close acquaintance with conceptual spaces may actually inhibit

P-creativity.

Much worse for Boden's theory, not all radically creative acts involve deliberate

attempts to transform conceptual spaces. Think here of Henri Matisse, who, it is well

known, developed a radically new approach to the use of colour in painting. His Woman

with the Hat contained bold areas of flat, pure colour that outraged French critics when it

was exhibited at the 1905 Salon d'Automne. In his essay, 'Notes of a Painter', he explains

the way in which he paints not as an attempt to transform some prevailing style, some

unified conceptual space with which he is intimately acquainted, but to achieve effects that he finds satisfying. He asks us to suppose that he has

to paint an interior . . . . and I put down a red which satisfies me.... Let me put a green

near the red and make the floor yellow; and again there will be relationships between the green or the yellow and the white of the canvas which will satisfy me. But these

different tones mutually weaken one another. It is necessary that the various marks I

use be balanced so that they do not destroy each other. To do this I .... 10

It would seem from what he says here and elsewhere that Matisse's aim in what is widely

regarded as extremely surprising and highly creative work, was not to transform a

pre-existing conceptual space with which he was closely acquainted, still less to solve

See Johnathan W. Schooler and Joseph Melcher, 'The Ineffability of Insight' in The Creative Cognition Approach, ed. Steven M. Smith, Thomas B. Ward, and Ronald A. Finke (Cambridge, MA: MIT Press, 1995), esp. p. 99 where the authors argue that not having excessive experience with a problem or in a field is sometimes helpful. I am indebted to Kevin Melchione for drawing my attention to this article. See Jack D. Flare, Matisse on Art (Oxford: Phaidon, 1978), p. 37.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 73

problems that could not be solved within it. His aim, rather, was to relay colour to the

canvas in ways that, given his present inclinations and moods, he found satisfying.

Although widely considered to be profoundly creative, Matisse's creativity, in this case, is

manifested in something akin to play, and not, as some have argued, in problem-solving. 11

Even if, in so doing, he effectively transformed a given conceptual space, there is no a

priori reason to suppose that he must first have explored this conceptual space, still less

that it was his aim to transform it.

In just the same way, there can be no doubt that Pablo Picasso did not deliberately set

out to transform a well-explored Set of painterly conventions with his groundbreaking Les

Demoiselles d'Avignon (1907) even though there undoubtedly was a t radi t ion--a

conceptual space-- that he did transform. It was his fascination with tribal carvings,

especially those of African origin, that led to this painting, and so to the introduction of

forms that so impressed Georges Braque and signalled the advent of Cubism. If this is

right, and there is no reason to doubt it, at least one paradigm example of P-creativity did

not involve any attempt to transform an existing conceptual space or to solve problems

posed by it; nor did it require the prior exploration of such spaces.

More damaging yet, some cases of what we would normally regard as radical creativity

do not even require the existence, let alone the exploration and transformation, of a

conceptual space. Think here of Edward Jenner. Prior to his development of the smallpox

vaccine, there was no well-structured and unified body of knowledge or belief, no

conceptual space, that dealt specifically with vaccinations and immunity. Certainly

doctors wanted a cure for smallpox, and they undoubtedly dreamed of preventing the

disease. It was known that one could get smallpox only once, and as a result some people

had tried, with a notable lack of success, to induce mild forms of the disease through

deliberate infection. But there was no well-organised body of theory or unified set of ideas

according to which it was possible to induce the body to protect itself against a disease.12

Rather, the story goes, Jenner overheard a young farm girl telling someone that she could

not contract smallpox because she had once had cowpox, and it was this that set Jenner

thinking and experimenting. The resultant ideas and medical procedures were startlingly

new, highly suggestive, and plainly encouraged others to explore the mechanisms of

infection and immunity. Indirectly it brought Robert K o c h t o the View that disease was

H Cf. L. Briskman, 'Creative Product and Creative Process in Science and the Arts', The Monist 22 (1980), esp. p. 95, and T. Nickles, 'Can Scientific Constraints be Violated Rationally' in Scientific Discovery, Logic, and Rationality, ed. T. Nickles (Dordrecht: Kluwer, 1980), p. 288, both of whom see creativity as conceptually linked to problem-solving. This claim brings D'Agostino, Chomsky's System of Ideas, p. 171, to the view that 'the degree of creativity manifested in a particular product is inversely proportional to the "strength" of the constraints on the problem for which that product is a solution'. Where the constraints strongly determine the solution to a problem, he argues, the solution is less creative than where the constraints allow the agent considerable latitude. While this is plausible for many cases of P-creativity, the strong connection made between creativity and problem-solving overlooks the fact that the explorative exercise in which Matisse engages when he paints is not directed at finding a solution to any particular problem---even in the very weak sense of problem-solving that D'Agostino isolates on p.170. It is not as if be sets out, for instance, to solve the problem of satisfying himelf. Rather, his is a playful activity performed for its own sake, not for the sake of solving a problem. Even so, it may properly be described as P-creative.

12 Some semi-mystical ideas of inoculation had existed in ancient China and India, but either were not available to Jenner or were not seriously entertained by him. Homeopathic theory, it is worth noting, dates from the 1790s--just a little later than Jenner.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

74 Creativity and Constraint

caused by microbes, and it would eventually enable Louis Pasteur to develop different and highly effective vaccines for other diseases. What we have, it seems to me, is an example of a development that most people would regard as radically, hence in Boden's terms psychologically, creative, but one which was starkly innovative rather than transformative or revolutionary; one that succeeded in developing a conceptual space ab initio rather than transforming an existing conceptual space.

It is open to Boden to defend her theory by pointing out that since Jenner knew of the

medical impossibility of contracting smallpox twice, since he knew that people were

trying unsuccessfully to protect themselves from the ravages of the disease by inducing mild forms of it in themselves, there was a conceptual space well-known to him, which he subsequently transformed. But this defence does not work. There is no suggestion in my argument that Jenner's discovery did not rely on earlier ideas; only that these ideas did not

constitute what Boden regards as a conceptual space. For on her view, as we have seen, a conceptual space requires 'organizing principles that unify and give structure to a given domain of thinking' and so 'define a certain range of possibilities' ( 'WIC', p. 79). She

does not tell us how unified such principles and ideas must be in order to constitute a conceptual space, but even so it is clear that if she allows very loosely related, disorganised clusters of ideas to count as conceptual spaces, she will find that such spaces cannot be systemically transformed in a way that will allow computational ideas to shed light on human P-creativity. If, on the other hand, conceptual spaces are always structured

sets of ideas that arise out of and depend on certain basic principles or general rules--like

perspectival projection, or Mendeleyev's periodic table, or Newtonian mechanics--then Jenner was not operating within, and did not transform, a conceptual space, even though his discovery was radically creative.

Boden could, I suppose, respond by maintaining that Jenner's development of a vaccine for smallpox, since it was brought about without exploring and transforming a conceptual space, simply was not psychologically, and, in her sense, radically creative.

But any such suggestion would have the wholly untoward effect of making her account of radical or P-creativity too exclusive. Jenner, after all, was responsible for developing a conceptual space pertinent to immune systems, where none had previously existed--and it seems arbitrary and deeply counter-intuitive to suggest that this was not a central instance of radically creative thinking.

The same considerations apply to Thomas Edison's invention of the phonograph in 1876. Although Edison's one thousand and ninety-three patented inventions were, for the most part, transformations of existing techniques, artefacts, and conceptual spaces, his invention of the phonograph does not seem to have involved such a transformation. The idea of developing some way of recording and reproducing the human voice was most probably occasioned by the fact that while people were able mechanically to record visual experiences with the help of the camera, there was no similar way of

recording our experience of sound. But the principles and conventions of photography cannot plausibly be regarded as the conceptual space transformed by Edison's invention, and while he certainly applied some of the laws of movement and acoustics in order to invent the phonograph, it can hardly be said that he transformed these. He would almost certainly have known that the human voice had been transmitted along an electric wire just the previous year by Alexander Graham Bell, but the phonograph did not employ, let alone transform, the electro-magnetic laws that Bell had exploited. So far as I can

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David No vi tz 75

see, there were no existing conceptual spaces pertinent to the phonograph that he could

transform; his invention, if the pun can be excused, was an invention from scratch. 13

It is interesting to contrast Jenner's and Edison's inventions with Charles Goodyear's

attempts to find a way of making rubber durable by preventing it from becoming brittle

with cold or viscous with heat. Perhaps because he was in prison for debt at the time,

Goodyear's researeh was surprisingly unsophisticated, involving only the successive

combination of raw rubber with a vast array of randomly chosen additives--from witch

hazel and cream cheese to black ink. It was only when he eventually stumbled on a

heat-sulphur treatment for rubber (by accidentally dropping rubber mixed with sulphur on

to a hot surface) that the process of vulcanization was discovered in 1839.14

Here, I think, we would normally resist describing Goodyear's development of the idea

of vulcanization, or his subsequent discovery of vulcanization, as radically creative. For at

that time, almost, anyone would have known that one could alter the physical properties of

a substance by combining it with other substances. Then, too, provided that there are no

life-extinguishing explosions along the way, anyone is bound to come across some new

and useful results by randomly combining enough different substances with rubber.

Even so, the idea and the subsequent discovery was deeply transformative of what

Boden would have to treat as an existing conceptual space. Up until then, there were

certain basic, law-like ideas that governed thinking about rubber, namely that rubber

becomes viscous with heat and brittle with cold. And this, in its turn, affected almost

everyone's ideas about what could be done with rubber. Here we have a conceptual space,

where ideas about what can be done with rubber are not just interrelated, but are

subtended, organised, and unified by some fundamental principles. And it is plain that

Goodyear transformed that space, since he developed the idea that rubber could take a

different form, need not become sticky or brittle with heat or cold, and this, in its turn,

radically changed the ideas that people had about rubber and the uses to which it might be

put.

Given her theory, Boden would have to regard the idea and discovery of vulcanization

as radically creative. On her view, as we will see later, the relative 'depth' of

transformation of a conceptual space, gives her a rough measure of degrees of P-creativity

('WIC', p. 113). Since the prevailing beliefs about rubber before 1838 constituted a

conceptual space, and since the idea of vulcanization could not have been thought in terms of it, Goodyear's idea and subsequent discovery would, by her own lights, have

'transformed' the fundamental organizing principles of that space. On Boden's view,

therefore, Goodyear's discovery was highly creative. 15

t3 See Glyn Alkin, Sound Recording and Reproduction (New York: Focal Press, 1981). Edison's phonograph consisted of a revolving cylinder wrapped in tinfoil. A sharp point was pressed against the foil-wrapped cylinder. Attached to the point were a diaphragm and a large mouthpiece. The cylinder was rotated by hand. When Edison spoke into the mouthpiece, his voice made the diaphragm vibrate. This caused the sharp point to cut a trace in the tinfoil. When a needle replaced the cutting point, the talking machine reproduced Edison's original words.

14 The discovery was not H-creative. The process of vulcanization was first discovered by American Indians. See Jack Weatherford, lndian Givers (New York: Fawcett Columbine, 1988), pp.46-49.

15 D'Agostino, Chomsky's System ofldeas, p.171, is bound to agree with Boden on this point. He, too, would have to treat Goodyear's discovery as highly creative since the constraints on the problem that Goodyear wishes to solve do very little, if anything, to determine the appropriate solution to it. See note 11 above.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

76 Creativity and Constraint

But this is odd. Goodyear's invention, if invention it was, was pretty laeklustre, and required very little in the way of imagination, intelligence, or endeavour. To describe it as highly creative in the way that Einstein's transformation of Newtonian physics was highly creative, and to suppose that it is somehow more creative than Jenner's development of the smallpox vaccine, or Matisse's painting, appears to be much too generous to Goodyear.

It is now clear, I hope, that where human creativity is concerned, Boden's theory is deficient in a number of respects. First, it underestimates the social constraints on human

P-creativity, and so fails to recognize that people may be P-creative without exploring the conceptual spaces that they transform; indeed, that the close exploration of a complex

conceptual space may on occasions inhibit rather than encourage human creativity, and is, in any event, at odds with the flair and spontaneity that is sometimes a feature of human creativity. Second, i t is maintained, incorrectly, that all cases of radical, hence psychological creativity require a pre-existing conceptual space. But this tends to make the account too exclusive--as the Jenner and Edison examples show. Third, the theory is in some respects too inclusive, for while Goodyear's discovery (or invention) of vulcanization seems to satisfy all of Boden's conditions for radical creativity, it hardly warrants this praise when compared with the achievements of Edison, Jenner, Picasso, or Mozart.

V. The Recombination Theory of Creativity

How are these defects to be remedied? Only by coming to realize that there is no single process that exhausts the notion of radical creativity. While Boden offers us an excellent account of one way of being radically creative, she neglects others, and offers us conditions that are not necessary and are, at best, only generally (but not invariably) sufficient for radical or, as she calls it, psychological creativity. On my view, people may be radically creative even when they do not transform anything as well-defined as a conceptual space. They may, for instance, combine ideas in accordance with well-established rules and techniques, and they may do so in ways that turn out to be wonderfully insightful, that shift our understanding of certain phenomena, that cast things

differently--in a different light, with new emphases, altered alignments and surprising new connections.

When Antonia Byatt combines words, phrases, images, and ideas in accordance with

existing semantic and grammatical rules, and in so doing produces a novel that is resplendent with insight and new understanding, I can see no reason to deny that her work

is radically creative, even though there are no obvious conceptual spaces that she transforms. And when Edison combines pieces of tin with a crank handle, a needle, a cylinder, a diaphragm, and a mouthpiece to make a phonograph, his invention is momentously surprising and deeply creative even though it does not obviously fit the requirements of Boden's theory.

Is there a theory about the nature of P-creative acts that will accommodate Byatt and Edison, Jenner, Picasso, and Matisse, while also allowing that Goodyear's invention was

too pedestrian to qualify as radically creative? I think that there is. It is a modified form of the combination theory of P-creativity that Boden dismisses as inadequate early on in her book (CM, oh. 3). According to her, such theories 'fail to capture the fundamental novelty

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 77

that is distinctive of creative thought' (CM, p. 30). But we already know what this

'fundamental novelty' amounts to, and we have seen that there are reasons for doubting that it is distinctive of all radically creative acts.

The theory that I favour allows that the starting point of human creativity is always an existing cluster of ideas or objects or techniques, but insists that such clusters need not constitute or form part of conceptual spaces. Allowing that the term 'object' may include sensations and qualities as well as physical objects, my claim is that all creative acts require

(1) the intentional or chance recombination of such ideas, techniques, or objects--where this recombination is subsequently deliberately used or deployed

(2) in ways that result in something that is (or would have been) surprising to---hence, not predicted b y - - a given population, and

(3) in ways that are intended to be, and are potentially, of real value to some people.

These three conditions are each necessary and jointly sufficient for creativity. However, since (as we shall see) each condition may be satisfied to a greater or lesser degree, a high

degree of creativity, like radical creativity, is not assured just by satisfying these three conditions.

There doubtless are specific psychological mechanisms that facilitate very valuable and surprising recombinations--hence a high degree of creativity. But the job of discovering these I leave to psychologists and cognitive scientists. 16 My aim in developing

this theory is only to uncover the framework within which such an investigation must be conducted, and to rule out the idea that the transformative process described by Boden exhausts the notion of radical creativity.

As it stands, then, the recombination theory of creativity allows that there is an indefinitely large number of processes that constitute creative recombinations, only some

of which transform the basic rules that define conceptual spaces. Others may involve recombining words or phrases according to existing grammatical and semantic rules, yet in ways that nevertheless turn out to be valuable and surprisingly insightful. Or composers may recombine notes and musical phrases in accordance with the prevailing rules of counterpoint and harmony, and may do so in ways that result in the brilliance of Mozart's music. In this way, the recombination theory can allow that many of Mozart's compositions were P=creative even though he only elaborated and explored but never actually transformed the conceptual spaces that he had inherited from Joseph Haydn. Then, too, there are recombinations that, far from being the result of detailed exploration or any great learning, are simply a matter of serendipity--sometimes the freak chance and happy result of simple play; sometimes, too, the result of comparatively uninformed trial and error strategies that are subsequently put to valuable and surprising use.

According to this theory, recombinations do not always involve the actual

manipulation of words, phrases, lines, eolours, notes, and physical objects. Sometimes they are purely ideational. Thus, for instance, aspection of the sort that allows me to see a

16 See, for example, Ronald A. Finke, Thomas B. Ward and Steven M. Smith, Creative Cognition: Theory Research and Applications (Cambridge, MA: MIT Press, 1992). See, as well their edited collection The Creative Cognition Approach (Cambridge, MA: MIT Press, 1995).

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

78 Creativity and Constraint

duck-rabbit figure first as a duck, then as a rabbit, does not involve the actual

recombination of any parts of this arrangement of lines. Rather, by considering aspects of

the figure (or of any scene, state of affairs, or visual image) in terms of different beliefs or

theories, and so with different emphases, the beak of the duck is thought of, or is seen as,

a rabbit's ears and, as a result, everything else 'falls into place': the indentation at the back

of the duck's head becomes the rabbit's mouth; the eye shifts its gaze; the lump at the

back of the rabbit's head becomes part of the duck's chin. But the rearrangement, the

recombination, does not involve the physical manipulation of any parts of the figure. Such

manipulation as there is, is ideational, and involves looking differently, by considering

aspects of the figure under different descriptions. To do this is to recombine these aspects

with different ideas or beliefs in ways that may breed new, sometimes wonderful insights,

innovative ways of seeing or thinking or understanding--in much the way, I think, that

the Copernican revolution did.

The second condition--the surprise condition--captures the intuition that creativity

involves producing something novel or innovative from the point of view of a given

population whose members were acquainted with some of the ideas or objects prior to

their recombination. In other words, it captures the intuition that such an act and its

outcomes could not easily have been predicted. Of course, accidents do happen, and it is

well known that some recombinations will not live to see the light of day. On my view, such recombinations may nonetheless be properly regarded as creative if it is ~ue that

they would have been surprising (and of real value) to a population had they survived.

The third condition requires that a creative act be of real value to some people. Let us

say that an act or an object is of real value if it possesses properties that are of actual or

potential benefit to sentient beings: that either do or can increase enjoyment of life,

enhance security, health, prosperity, and so on. It follows that a recombination that

appears to be valuable, yet is later found to be thoroughly harmful and of no lasting

benefit to anyone, will not be of real value and so will not be creative. This is why one can

create something--mayhem or a mess, for instance--yet not be creative. The mad

scientist who creates nothing but harm is ingeniously destructive but his actions are not

properly described as creative.

It is true that an evil scientist may delight in his malevolent acts, and if he does it

seems that his inventiveness is of real value to at least one person, namely himself, so that

his inventions, although enormously harmful and destructive, would seem, by my account,

to be creative. And this, if true, is a perfectly unwelcome consequence of the present

account.

However, this sort of objection overlooks the fact that there are degrees of creativity,

and that a recombination that benefits only one person is of much less value, hence much

less creative, than one that benefits many. So it may be the case that such an invention is

minimally creative, but this is a point that I take up presently. For the time being, we

should note that the objection also overlooks the fact that, according to my third condition,

a recombination, or the use made of it, must be intended to be of beneficial value. If, as

could be the case in this example, the evil scientist intends primarily to harm people, his invention fails to satisfy my third condition.

It is important to emphasize both the inventor's intention and the potential

consequences of the invention. For it is clear that an act may harm some people and yet be

creative, provided that it is designed to benefit such people and has the clear potential to

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 7 9

do so. On the account that I have given, it would then possess real value. For the most

part, we judge the creativity of an act not so much on its actual but on its potential

benefits. The invention of a new wonder drug to cure full-blown AIDS will be judged to

be of value, and so considered creative, long before it has had time to work its benefits.

Even when a recombination is entirely due to chance, we expect the person who

deploys or uses it creatively to do so while intending to put it to a beneficial use. On my

view, then, if an act is designed to harm but quite fortuitously benefits large numbers of

people, we might regard it as a happy accident, but not as a creative act. So, for instance,

were Dr Strangelove to develop a bomb with which to destroy his least favourite planet,

but find to his enormous surprise and dismay that instead of achieving his aim, it greatly

increases plant growth, restores forests, and wholly rejuvenates the environment, we still

would not describe his act as creative. At no stage did he envisage, let alone plan, this

consequence, and partly because of this we decline to think that he has acted creatively.

What happened, we say, was wholly accidental and certainly not something for which he

can be praised.

Suppose, though, that Strangelove's invention is confiscated by the police, handed to

a different scientist who, after a good deal of thought, comes to the view that, although

never intended as such, the device can be used to improve the environment. Suppose,

too, that the second scientist subsequently uses it to this end. In this case, a creative act

is performed, although not by Dr Strangelove. It is the scientist who painstakingly

discovered the beneficial worth of Strangelove's findings and who subsequently

deployed them in ways that were intended to be of benefit, who has performed a creative

act.

There have been some inventions--nuclear weapons are one--that have been

obviously beneficial to some people while harming, even destroying others. Here it can be

difficult, perhaps impossible, to determine the precise value of the invention, but there can

be no doubt that the invention of nuclear weapons was intended to benefit and harm a

considerable body of people, and that it has the undoubted potential to do both. It would

seem, then, that on my theory, the invention both was and was not creative, and this is

plainly untenable.

But the problem can be avoided. It is part of our shared knowledge of modem history

that nuclear weapons were invented in time of war and only after Franklin D. Roosevelt

had been advised by Albert Einstein that the Germans were trying to do the same, that

were they to be successful the United States and the rest of the free world would almost

certainly succumb to the might of Germany. It was in order to avert this fate that a group

of scientists under Oppenheimer attempted to, and eventually did develop the weapon. If

we believe the story, their aim was to serve the American people beneficially, and it was

always a matter of regret to them that the weapon would also harm many others.

When once we know the motives of its inventors, and when once we have been

informed about the situation in which it was developed, we have little difficulty describing

the invention of nuclear weapons as creative. The primary motivation of the inventors was

beneficial since it was directed at the survival of their country and the free world. Since

this could be achieved only by potentially harming others, the aim of harming the enemy

was at best secondary--a conceptually unavoidable consequence of entertaining the

primary and beneficial aim. The invention thus satisfies the third condition of my definition.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

80 Creativity and Constraint

What is of real value under one set of conditions can, of course, have no value at all

under other conditions. It follows from this that what is properly described as creative

changes from one set of circumstances to another. But this, it might be objected, amounts

to a relativistic account of creativity since it now turns out that any creative act depends

for its creativity not on the nature of the act itself but on on the value that attaches to it in a

given situation--and that this will change depending on the social situation of the valuer.

The objection, though, is misleading and trades on a confusion between (a) the fact that

the real value of an act depends on circumstance, with (b) the supposed fact that the real

value of an act depends on the value ascribed to it in any given social circumstance. The

two are not the same. There certainly are some cases in which a socially ascribed value is

of real value--the knighthood conferred on me by the Queen of England may be of real

value to myself and my heirs. But the value ascribed to skin eolour in a society may be of

no real value at al l-- in the sense that it causes widespread harm or bestows no lasting

benefit. Plainly, then, the fact that people in a society regard something as of value, and so

as creative, does not entail that it really is creative. Rather, what is regarded as creative

will not actually be so unless each of the three conditions for creativity is satisfied--that

is, unless there is a recombination that is both surprising, is intended to be, and is of real

value to a given population. Of course, one's social interests may prevent one from

perceiving the value that an act has---or, alternatively, may encourage one to see value where there is little or none at all.

VI. Creativity by Degrees

By now it will be clear that the recombination theory of creativity allows that there are,

indeed that there have to be, degrees of creativity. What is more, it allows for their

measurement in ways that are not available to Boden. On her view, there can be no

numerical measure of creativity, but even so, since some features of a conceptual space

'are "deeper", more influential, than others', the relative 'depth' of transformation gives

her an impressionistic measure of degrees of radical or P-creativity ( 'WIC', p. 113). But,

as I have remarked, what she fails to see, is that on this measure Goodyear turns out to be

highly creative since he successfully transforms the deep principles that govern a

conceptual space. Mozart, however, does not do anything of the sort, and so, on her

measure, should be considered much less creative than Goodyear.

The recombination theory can do better than this. The measures that it suggests are on

three different axes, each corresponding to one of the conditions for creativity. The first is

measured in terms of the difficulty of the achievement--given in terms either of the

likelihood of just anybody (that is, of people-in-general) effecting precisely the same

recombination in similar circumstances, or else of putting a chance recombination to

(beneficial) use. If there is a high likelihood (approaching, say, a probability of 1) that any

person with an average IQ can do the same, we would regard the achievement as much

less creative than if there is only a negligible likelihood of this. In the latter case, we

would think the new ideas highly creative; perhaps as requiring special talents, skills,

insights, knowledge, and the right degree of motivation. For all of these reasons, we

would normally think of it as something that is difficult to achieve--in the way that

winning a gold medal at the Olympic Games is difficult to achieve, even if the winner, on occasions, finds it a relatively easy thing to do.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

David Novitz 81

The second axis is concerned with the degree of surprise generated by the recombination, and can be measured by inverting standard information theory; which is to say that the more informative the resultant combination of ideas or objects is to those rational individuals who were acquainted with them in their original (non-reeombined) state, the more unusual or surprising (hence the newer, more innovative) they are.

It might be thought, though, that the second axis is just a version of the first. After all, it is said, the difficulty of an achievement will be related to the surprise occasioned by it. But not so; a recombination may be very difficult to think up, yet result in nothing

particularly new or surprising. And sometimes, very surprising effects are achieved quite easily--through happy accidents perhaps, whose relevance to some or other problem is immediately obvious.

The third axis has to do with value. The problems attendant on the measurement of

value are, of course, notorious, and I do not propose to settle them here. However, as I have explained it, the value that attaches to creativity is closely related to the benefits to be derived from any particular recombination of objects, ideas, techniques, et cetera. Suffice it to say that utility, very broadly construed, offers one promising measure of value, and that on this measure, the greater the number of people to whom a recombination is (or would be) beneficial, the more valuable the reeombination. Even if, when all is said, we cannot find an uncontroversial measure that allows for proper comparisons of value where different kinds of benefit are involved, it does seem clear that

if the degrees of difficulty and surprise are held constant for any recombination, an increase in value--say, in usefulness or in pleasure--will always increase the degree of creativity of that recombination.

It is important to see, though, that all these measures, although capable of refinement, must in the end result in a disappointingly uninformative global measure of P-creativity. The reason for this is that the same innovation may rate highly on one axis, without rating highly on another. Using inverted information theory, a new insight or idea may be shown to be surprising to a certain quantifiable degree, but since the idea may not have been very difficult to come by, and may be only moderately valuable, it is impossible, on this account, to affix an informative global score to any one creative act-- in the sense that a global score will not tell us which of the three conditions bears the weight of any

particular creative act. In such a case, the global score indicates the degree of creativity, but reasons corresponding to each of the axes will need to be given in explaining and defending this score.

We are now in a position to show that the recombination theory does much greater justice to our intuitions about P-creativity than Boden's theory does. Like Boden's theory,

the recombination theory can easily explain why the transformation of conceptual spaces is creative; but unlike her theory, it allows that the works of Mozart, like Jenner's discovery of the smallpox vaccine and Edison's invention of the phonograph, are highly

creative--even though they do not transform conceptual spaces. And it explains them all in the same way. Creativity in each case is achieved through the recombination of objects, ideas, or techniques, where such recombinations turn out to be both surprising and valuable. But whereas Boden is bound, on her theory, to treat Goodyear's invention of vulcanization as highly creative since it transforms the basic or 'deep' 'organizing principles' that 'unify' a given conceptual space, on my view the invention is, at the very best, only slightly, perhaps moderately, creative since, although it is of high beneficial

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012

82 Creativity and Constraint

value, it was neither difficult to achieve nor very surprising. Goodyear, after all, had

confidently predicted that, i f combined with some or other substances, raw rubber would

cease to be brittle when cold and viscous when hot. It was just a matter of finding the

substance. If one asks precisely how surprising Goodyear's idea or innovation has to be in order to

satisfy the second condition, the answer must be imprecise. At best we can say that it need

not be very or highly surprising; that, at the very least, it must provide some new

information to some reasonable people who were previously acquainted with the

constituent objects, ideas, or techniques. And we can also say that i f the idea is highly

informative, it will be deemed more surprising--(and more creative) than if it is only

moderately informative. And much the same for the first condition; there is, and can be,

no precise point at which the difficulty condition is or is not satisfied.

This imprecision is what we should expect. We should expect it because creativity is,

in part, a cultural phenomenon, often embedded in vaguely-formulated, partly-shared

beliefs, expectations, and values. This is why we find it difficult at times to decide just

how creative a new set of ideas is, and it is why our reasons for describing something as

creative may shift in emphasis, in accordance with the conditions that I have outlined. Of

course, this does not mean that there are no criteria for creativity; only that they are fuzzy

at the edges, and that they do not always fit or fail to fit the phenomena with the precision that scientists sometimes seek) 7

University of Canterbury Received: September 1997

Revised: July 1998

My thanks to Derek Browne, Philip Catton, Jack Copeland, Stephen Davies, Stan Godlovitch, Richard Keshen, Graham Oddie, Roy Perrett, Diane Proudfoot, Roy Sorensen, Alex Wynne, and two anonymous referees of this journal for comments on an earlier draft of this paper.

Dow

nloa

ded

by [

The

Uni

vers

ity o

f B

ritis

h C

olum

bia]

at 2

3:41

18

Mar

ch 2

012