CREATIVITY IN A NUTSHELL*
Margaret A. Boden
university of Sussex
*This article is reprinted from pp. 1-10 of M. A. Boden, The Creative Mind: Myths and
Mechanisms (London: Routledge, 2004).
Creativity and computers: what could these possibly have todowith one another? "Nothing!,"
manypeople would say.Creativity is a marvel of the human mind. But computers, with all due
apologies to Mario, Sonic, and friends, are basically just tin-cans. It follows -- doesn’tit? -- that
the twoare related only by utter incompatibility.
Well, no. Computers and creativity makeinteresting partners with respect to twodifferent
projects. One, which interests me the most, is understanding human creativity.The other is trying
to produce machine creativity -- or anyway,machine "creativity" -- in which the computer at
least appears to be creative,tosome degree.
What is Creativity?
First things ﬁrst. Human creativity is something of a mystery,not to say a paradox. One newidea
may be creative,while another is merely new. What’sthe difference? And howiscreativity
possible? Creative ideas are unpredictable. Sometimes, theyevenseem to be impossible -- and
yet theyhappen. Howcan that be explained? Could a scientiﬁc psychology help us to understand
howcreativity is possible?
Creativity is the ability to come up with ideas or artefacts that are new, surprising,and
valuable. "Ideas," here, includes concepts, poems, musical compositions, scientiﬁc theories,
cooking recipes, choreography, jokes ... and so on, and on. "Artefacts" include paintings,
sculpture, steam-engines, vacuum cleaners, pottery,origami, penny-whistles ... and you can name
As these highly diverse examples suggest, creativity enters into virtually every aspect of life.
It’snot a special "faculty," but an aspect of human intelligence in general. In other words, it’s
grounded in everyday abilities such as conceptual thinking, perception, memory,and reﬂective
self-criticism. So it isn’tconﬁned to a tinyelite: every one of us is creative,toadegree.
Nor is it an all-or-none affair.Rather than asking "Is that idea creative,Yes or No?," we should
ask "Just howcreative isit, and in just which way(s)?" Asking that question will help us to
appreciate the subtleties of the idea itself, and also to get a sense of just what sorts of
psychological process could have brought it to mind in the ﬁrst place.
Creative ideas, then, are new. But of course, there’snew --and there’s new. Ask a teacher,for
instance. Children can come up with ideas that are new to them, ev e nthough theymay have been
Dagstuhl Seminar Proceedings 09291
Computational Creativity : An Interdisciplinary Approach
in the textbooks for years. Someone who comes up with a bright idea is not necessarily less
creative just because someone else had it before them. Indeed, if the person who had it ﬁrst was
Shakespeare, or Euclid, we’dthink evenmore highly of the achievement.
Suppose a twelve-year old girl, who’dnev erread Macbeth, compared the healing power of
sleep with someone knitting up a ravelled sleeve.Would you refuse to say she was creative,just
because the Bard said it ﬁrst? Perhaps, if you’dbeen talking around the topic with her,
encouraging her to come up with non-literal ways of speaking, and evenputting one or more of
the three key ideas into the conversation. Otherwise, you’dhav e to acknowledge her remark as a
truly imaginative one.
What you might do, and what I think you should do in this situation, is to makeadistinction
between "psychological" creativity and "historical" creativity.(P-creativity and H-creativity,for
short.) P-creativity involves coming up with a surprising, valuable idea that’snew to the person
who comes up with it. It doesn’tmatter howmanypeople have had that idea before. But if a new
idea is H-creative,that means that (so far as we know) no-one else has had it before: it has arisen
for the ﬁrst time in human history.
Clearly,H-creativity is a special case of P-creativity.For historians of art, science, and
technology -- and for encyclopaedia users, too -- H-creativity is what’simportant. And in daily
life, we appreciate it too: it really isn’ttrue that "The old jokes are the best ones". But for
someone who is trying to understand the psychology of creativity,it’sP-creativity that’scrucial.
Nevermind who thought of the idea ﬁrst: howdid that person manage to come up with it, given
that they had neverthought of it before?
If "new,"inthis context, has twoimportantly different meanings, "surprising" has three.
An idea may be surprising because it’sunfamiliar,orevenunlikely -- likea100-to-1 outsider
winning the Derby.This sort of surprise goes against statistics.
The second sort of surprise is more interesting. An unexpected idea may "ﬁt" into a style of
thinking that you already had -- but you’re surprised because you hadn’trealized that this
particular idea was part of it. Maybe you’re evenintrigued to ﬁnd that an idea of this general type
ﬁts into the familiar style.
And the third sort of surprise is more interesting still: this is the astonishment you feel on
encountering an apparently impossible idea. It just couldn’t have entered anyone’shead, you feel
-- and yet it did. It may evenengender other ideas which, yesterday,you’dhav e thought equally
impossible. What on earth can be going on?
The Three Ways of Creativity
"What is going on" isn’tmagic -- and it’sdifferent in each type of case. Forcreativity can
happen in three main ways, which correspond to the three sorts of surprise.
The ﬁrst involves making unfamiliar combinations of familiar ideas. Examples include poetic
imagery,collage in painting or textile art, and analogies. These newcombinations can be
generated either deliberately or,often, unconsciously.Think of a physicist comparing an atom to
the solar system, for instance, or a journalist comparing a politician with a decidedly non-cuddly
animal. Or call to mind some examples of creative associations in poetry or visual art.
In all these cases, making -- and also appreciating -- the novelcombination requires a rich
store of knowledge in the person’smind, and manydifferent ways of moving around within it.
The journalist or newspaper-reader needs a host of concepts about both politics and animal
behaviour,and some "personal" knowledge about the individual politician in question.
Cartoonists who depict Ken Livingstone (the ﬁrst publicly-elected Mayor of London) as a newt
are tapping into manydifferent conceptual streams, including gossip about what he keeps in an
aquarium in his home. The surprise you feel on looking at the cartoon is largely caused by seeing
ahuman ﬁgure with a newt’screst and tail: a combination of ideas that’sevenless probable than
the outsider winning the Derby.
If the novelcombination is to be valued by us, it has to have some point. It may or (more
usually) may not have been caused by some random process -- likeshaking marbles in a bag. But
the ideas/marbles have tohav e some intelligible conceptual pathway between them for the
combination to "makesense." The newt-human makes sense for manyreasons, one of which is
Ken’sfamed predilection for newts. (What are some of the others?) And (to return to the example
from Macbeth) sleep is a healer,asknitting can be. Even if twoideas are put together randomly
in the ﬁrst place, which I suspect happens only rarely,theyare retained/valued only if some such
links can be found.
The other twotypes of creativity are interestingly different from the ﬁrst. Theyinv olvethe
exploration, and in the most surprising cases the transformation, of conceptual spaces in people’s
Exploring Conceptual Spaces
Conceptual spaces are structured styles of thought. They’re normally picked up from one’sown
culture or peer-group, but are occasionally borrowed from other cultures. In either case, they’re
already there: theyaren’toriginated by one individual mind. Theyinclude ways of writing prose
or poetry; styles of sculpture, painting, or music; theories in chemistry or biology; fashions of
couture or choreography, nouvel cuisine and good old meat-and-two-veg ... in short, any
disciplined way of thinking that’sfamiliar to (and valued by) a certain social group.
Within a givenconceptual space, manythoughts are possible, only some of which may have
been actually thought. Some spaces, of course, have a richer potential than others. Noughts-and-
crosses is such a restricted style of game-playing that every possible move has already been made
countless times. But that’snot true of chess, where the number of possible moves, though ﬁnite,
is astronomically large. And if some sub-areas of chemistry have been exhausted (every possible
molecule of that type having been identiﬁed), the space of possible limericks, or sonnets, has not
-- and neverwill be.
Whateverthe size of the space, someone who comes up with a newidea within that thinking-
style is being creative inthe second, exploratory,sense. If the newidea is surprising not just in
itself but as an example of an unexpected general type, so much the better.And if it leads on to
others (still within the same space) whose possibility was previously unsuspected, better still.
Exploratory creativity is valuable because it can enable someone to see possibilities theyhadn’t
glimpsed before. Theymay evenstart to ask just what limits, and just what potential, this style of
We can compare this with driving into the country,with an Ordnance Survey map that you
consult occasionally.You can keep to the motorways, and only look at the thick red lines on your
map. But suppose, for some reason (a police-diversion, or a call of nature), you drive off onto a
smaller road. When you set out, you didn’tevenknowitexisted. But of course, if you unfold the
map you’ll see it marked there. And perhaps you ask yourself "I wonder what’sround that
corner?," and drive round it to ﬁnd out. Maybe you come to a pretty village, or a council estate;
or perhaps you end up in a cul-de-sac, or back on the motorway you came offinthe ﬁrst place.
All these things were always possible (and they’re all represented on the map). But you’dnev e r
noticed them before -- and you wouldn’thav e done so now, ifyou hadn’tgot into an exploratory
frame of mind.
In exploratory creativity,the "countryside" is a style of thinking. Instead of exploring a
structured geographical space, you explore a structured conceptual space, mapped by a particular
style of painting, perhaps, or a speciﬁc area of theoretical chemistry.
All professional artists and scientists do this sort of thing. Even the most mundane street-
artists in Leicester Square produce newportraits, or newcaricatures, every day.Theyare
exploring their space, though not necessarily in an adventurous way.Occasionally,theymay
realize that their sketching-style enables them to do something (convey the set of the head, or the
hint of a smile) better than they’dbeen doing before. Theyadd a newtrick to their repertoire, but
in a real sense it’ssomething that "ﬁts" their established style: the potential was always there.
Transforming the Space
What the street-artist may also do is realize the limitations of their style. Then, theyhav e an
opportunity which the Sunday driverdoes not. Give ortakeafew years, and ignoring earthquake
and ﬂood, the country roads are ﬁxed. Certainly, you can’tchange them. Your Ordnance Survey
map is reliable not only because it’sright, but because it stays right. (Have you bothered to buy a
newbook of road-maps within the last fewyears?) But the maps inside our heads, and favoured
by our communities, can change -- and it’screative thinking which changes them.
Some changes are relatively small and also relatively superﬁcial. (Ask yourself: what’sthe
difference?) The limits of the mental map, or of some particular aspect of it, are slightly pushed,
slightly altered, gently tweaked. Compare the situation in geographical space: suppose everyone
in that pretty village suddenly added a roof-extension to their cottage. It may ruin the prettiness
of the village, but it won’tchange the dimensions of the map. At most, the little "portrait" of the
village (assuming that it’s that sort of map) will have toberedrawn.
The street-artist, then -- or Picasso, in a similar position -- has an opportunity.Inprinciple, he
(or,asalways, she) could do the psychological equivalent of adding roof extensions, or building
anew road (a newtechnique, leading to newpossibilities), or evenre-routing the motorway.
Re-routing the motorway (in "real life" as in the mind) is the most difﬁcult of all. The surprises
that would engender could be so great as to makethe driverlose his bearings. He may wonder if
he’sbeen magically transported to a different county,orevenadifferent country.Maybe he
remembers a frustrating episode on his last trip, when he wanted to do something but his
passenger scornfully said: "In England, motorways are like this: theysimply don’tallowyou to
do that. Youwant to do it? Tough! It’simpossible."
Agiv e nstyle of thinking, no less than a road-system, can render certain thoughts impossible --
which is to say,unthinkable. The difference, as remarked above,isthat thinking-styles can be
changed -- sometimes, in the twinkling of an eye.
Someone skilfully writing a limerick won’tﬁnd iambic pentameters dropping from their pen.
But if you want to write a newsort of limerick, or a non-limerick somehowgrounded in that
familiar style, then maybe blank verse could play a role. The deepest cases of creativity involve
someone’sthinking something which, with respect to the conceptual spaces in their minds, they
couldn’t have thought before. The supposedly impossible idea can come about only if the creator
changes the pre-existing style in some way.Itmust be tweaked, or evenradically transformed, so
that thoughts are nowpossible which previously (within the untransformed space) were literally
inconceivable. -- But howcan that possibly happen?
Machine-Maps of the Mind
To understand howexploratory or transformational creativity can happen, we must knowwhat
conceptual spaces are, and what sorts of mental processes could explore and modify them.
Styles of thinking are studied by literary critics, musicologists, and historians of art, fashion,
and science. And theyare appreciated by us all. But intuitive appreciation, and evenlifelong
scholarship, may not maketheir structure clear.(An architectural historian, for instance, said of
Frank Lloyd Wright’sPrairie Houses that their "principle of unity" is "occult".)
This is the ﬁrst point where computers are relevant. Conceptual spaces, and ways of exploring
and transforming them, can be described by concepts drawn from artiﬁcial intelligence (AI).
AI-concepts enable us to do psychology in a newway,byallowing us to construct (and test)
hypotheses about the structures and processes that may be involved in thought. For instance, the
structure of tonal harmony, orthe "grammar" of Prairie Houses, can be clearly expressed, and
speciﬁc ways of exploring the space can be tried out. Methods for navigating, and changing,
highly-structured spaces can be compared.
Of course, there is always the additional question of whether the suggested structures and
processes are actually implemented in human heads. And that question isn’talways easy to
answer.But the point, here, is that a computational approach givesusaway of coming up with
scientiﬁc hypotheses about the rich subtleties of the human mind.
What of the second link between machines and creativity? Can computers be creative?Orrather,
can theyatleast appear to be creative?
Manypeople would argue that no computer could possibly be genuinely creative, no matter
what its performance was like. Even if it far surpassed the humdrum scientist or street-artist, it
would not be counted as creative.Itmight produce theories as ground-breaking as Einstein’s, or
music as highly valued as McCartney’s"Yesterday" or evenBeethoven’sNinth ... but still, for
these people, it would’nt really be creative.
Several different arguments are commonly used in support of that conclusion. For instance: it’s
the programmer’screativity that’satwork here, not the machine’s. The machine isn’tconscious,
and has no desires, preferences, or values -- so it can’tappreciate or judge what it’sdoing. A
work of art is an expression of human experience and/or a commmunication between human
beings, so machines simply don’tcount.
Perhaps you accept at least one of those reasons for denying creativity to computers? Very
well, I won’targue with you here (but see Chapter 11 of Boden 2004). Let’sassume, for the
purpose of this discussion, that computers can’treally be creative.The important point is that
this doesn’t mean that there’snothing more of interest to say.
All the objections just listed accept, for the sakeofargument, that the imaginary computer’s
performance is indeed very likethat of human beings, whether humdrum or not. What I want to
focus on here is whether it’strue that computers could, in fact, come up with ideas that at least
appear to be creative.
Well, think of combinational creativity ﬁrst. In one sense, this is easy to model on a computer.
Fornothing is simpler than picking out twoideas (twodata-structures) and putting them
alongside each other.This can evenbedone with some subtlety,using the (connectionist)
methods described in Chapter 6. In short: a computer could merrily produce novelcombinations
till Kingdom come.
But would theybeofany interest? Wesaw,above,that combining ideas creatively isn’tlike
shaking marbles in a bag. The marbles have tocome together because there is some intelligible,
though previously unnoticed, link between them which we value because it is interesting --
illuminating, thought-provoking, humorous ... -- in some way.(Think sleep and knitting, again.)
We saw also that combinational creativity typically requires a very rich store of knowledge, of
manydifferent kinds, and the ability to form links of manydifferent types. (Here, think
politicians and newts again.)
And we don’tonly form links, we evaluate them. For instance, we can recognize that a jokeis
"in bad taste." In other words: yes, the links that the joker is suggesting are actually there (so it is
areal joke). But there are other links there also, which connect the ideas with sorrow,
humiliation, or tragedy.The joker should have noticed them, and should have refrained from
reminding us of them.
Foracomputer to makeasubtle combinational joke, nevermind to assess its tastefulness,
would require (1) a data-base with a richness comparable to ours, and (2) methods of link-
making (and link-evaluating) comparable in subtlety with ours. In principle, this isn’timpossible.
After all, the human mind/brain doesn’tdoitbymagic. But don’thold your breath!
The best example of computer-based combinational creativity so far is a program called JAPE,
which makes punning jokes of a general type that’sfamiliar to every eight-year-old (see Chapter
12). But making a one-offjest is usually more demanding. Ask yourself, for instance, what Jane
Austen had to knowinorder to write the opening sentence of Pride and Prejudice: "It is a truth
universally acknowledged that a single man in possession of a good fortune must be in want of a
wife." (And why, exactly,isitfunny?)
Artiﬁcial Explorers and Self-Transforming Machines
What about exploratory creativity? Several programs already exist which can explore a given
space in acceptable ways.
One example is AARON, a drawing-program described in Chapter 7. AARON can generate
thousands of line-drawings in a certain style, pleasing enough to be spontaneously remarked
upon by unsuspecting visitors -- and to be exhibited in galleries worldwide, including the Tate.
(The most recent version of AARON is able to paint its drawings, too: see Chapter 12.)
Another is David Cope’s"Emmy," discussed in Chapter 12. This composes music in many
different styles, reminiscent of speciﬁc human composers such as Bach, Vivaldi, Mozart ... and
Stravinsky. Still others include architectural programs that design Palladian villas or Prairie
Houses (also mentioned in Chapter 12), and programs that can analyse experimental data and
ﬁnd newways of expressing scientiﬁc laws (Chapter 8).
Afew AI-programs can eventransform their conceptual space, by altering their own rules, so
that interesting ideas result. Some of these ideas were already known to human beings, though
not speciﬁcally preﬁgured within the program. (See the discussion of the automatic
mathematician, AM, in Chapter 8.) But others are ﬁrst-time-fresh. "Evolutionary" programs, for
instance, can makerandom changes in their current rules so that newforms of structure result. At
each generation, the "best" structures are selected, and used to breed the next generation.
Tw o examples that evolvecoloured images (some of which, likeAARON’s, are exhibited in
galleries world-wide) are described in Chapter 12. In each case, the selection of the "ﬁttest" at
each generation is done by a human being, who picks out the most aesthetically pleasing
patterns. In short, these are interactive graphics-environments, in which human and computer
can cooperate in generating otherwise unimaginable images. These computer-generated images
often cause the third, deepest, form of surprise -- almost as if a coin being tossed repeatedly were
suddenly to showa wholly unexpected design. In such cases, one can’tsee the relation between
the daughter-image and its parent. The one appears to be a radical transformation of the other,or
ev e nsomething entirely different.
Anyone who has watched TV regularly overthe past fewyears, or who has visited museums
of contemporary art, will already knowthat manynovelgraphic images have been produced by
self-transforming AI-programs of this kind. The problem is not to makethe transformations: that
is relatively easy.What’sdifﬁcult is to state our aesthetic values clearly enough to enable the
program itself to makethe evaluation at each generation. At present, the "natural selection" is
done by a human being (for example, the gallery-visitor).
In more well-regulated domains, however, the value-criteria can often be stated clearly enough
to allowthe evolutionary program to apply them automatically.Anearly example, a program for
locating leaks in oil-pipelines, is mentioned in Chapter 8. Now, scientists are starting to use these
techniques to enhance their own creativity.Biochemical laboratories in universities and
pharmaceutical companies are using evolutionary programs to help design newmolecules for use
in basic research and/or medicine. Even the "brains" and "bodies" of robots can nowbeevolved,
instead of being designed (see Chapter 12).
Values and Creativity
One huge problem here has no special relevance to computers, but bedevils discussion of human
Isaid earlier that "new" has twomeanings, and that "surprising" has three. I didn’tsay how
manymeanings "valuable" has -- and nobody could. Our aesthetic values are difﬁcult to
recognize, more difﬁcult to put into words, and evenmore difﬁcult to state really clearly.(Fora
computer model, of course, theyhav e to be stated really, really clearly.)
Moreover, theychange: who will proudly admit, today,tohaving worn a beehive hairdo or
ﬂared trousers in the 1960s? Theyvary across cultures. And evenwithin a given"culture," they
are often disputed: different sub-cultures or peer groups value different types of dress, jewellery,
or music. And where transformational creativity is concerned, the shock of the newmay be so
great theat evenfellow-artists ﬁnd it difﬁcult to see value in the novelidea.
Even in science, values are often elusive and sometimes changeable. Just what "simpliity" or
"elegance" mean, as applied to scientiﬁc theories, is something that philosophers of science have
long tried -- and failed -- to pin down precisely.And whether a scientiﬁc ﬁnding or hypothesis is
"interesting" depends on the other theories current at the time, and on social questions too (might
it have some medical value, for instance?).
Because creativity by deﬁnition involves not only novelty but value, and because values are
highly variable, it follows that manyarguments about creativity are rooted in disagreements
about value. This applies to human activities no less than to computer performance. So even if we
could identify and program our aesthetic values, so as to enable the computer to inform and
monitor its own activities accordingly,there would still be disagreement about whether the
computer even appeared to be creative.
The answer to our opening question, then, is that there are manyintriguing relations between
creativity and computers. Computers can come up with newideas, and help people to do so. Both
their failures and their successes help us think more clearly about our own creative powers.
Boden, M. A. (2004), The Creative Mind: Myths and Mechanisms (London: Routledge). 2nd