Creativity in science

Creativity in Science

Peter Doomen

K.U. Leuven


Ever since humanity began to think about the foundations of the environment, creativity has played a large but underestimated role. Creativity is thought to be a feature of writers, painters, and all kind of strange fish. A good illustration is the kuhnian viewpoint, i.e. that creativity does not belong to the realm of science, although it may play a role in development of new paradigms. I certainly do not agree with their stating that creativity is something for psychologists, and does not lean itself for scientific investigation (a point which is weakened by the Langley (1987) study reported in the third paragraph). However, as I shall claim, the kuhnian view of scientific development makes the role of creativity comprehensible, so I shall pick the useful parts out of it.

In this paper, I shall try to clarify cognitive underpinnings of creativity in science. I therefore shall adopt a cognitive viewpoint (Sternberg, 1996, p. 377). A first paragraph deals with the role of highly creative scientists in scientific revolutions. The second paragraph describes a cognitive perspective on creativity, thereby also considering problem-solving aspects of their discoveries. The third paragraph is devoted to a case study: Max Planck and the quantum-theoretical law of blackbody radiation.

Creativity and scientific revolutions

According to Kuhn, science develops in five stages (Bechtel, 1988).

1)Immature science; that is science that has not become mature enough to develop real paradigms. In immature science, one is trying to find out the basic structure of the domain under study.

2)Normal science, in which paradigms or theoretical matrices rule. A paradigm is a framework for characterizing the problem domain. In cognitive science, the idea of the mind as an information-processing system constitutes the paradigm.

3)Crisis. When the paradigm has become too weak to cover all problems in the problem domain, it is partially left, causing a crisis. This usually short period leads to a shift of the paradigm. Imagination is the important engine here. A good example is the book Verbal Behavior of B. F. Skinner (1957), which marked the end of the behaviorist period. Although intended to show the contrary, the book convinced many psychologists of the poverty of behaviorism for studying psychological phenomena like human language.

4)Revolution, which constitutes the shift of the paradigm. An example will be the replacement of Wien's law by the Planck Law of radiation (see third paragraph).

5)Resolution. After the shift of the paradigm, the newly adopted theory begins to settle down. A new period of normal science starts over again, and the loop continues.

Although the Kuhnian view of science has been criticized by other philosophers of science, it has the advantage that it offers an accurate framework for famous scientific revolutions, like the one started by Planck in 1899. Especially interesting about this theory is that creativity is considered important, above all in revolutionary periods. Because in such periods, the existence of problems with the old paradigm is clear, large parts of creativity will be problem solving. Problem solving can be seen as the best understood form of creativity, largely because it is easier to measure performance.

A cognitive perspective on creativity

Creativity is both a puzzling and an interesting human property. Puzzling, because it is a broad concept: creative persons range from painters as van Gogh to scientists as Stephen Hawking. And interesting, because it might be one of the most typical human properties, in need of almost all higher-order functions of our mind.

Which factors make a person distinctively creative? Several approaches have been proposed, among which (Sternberg, 1996):

The psychometric approach, that tries to measure creativity, usually with open-ended questions. Although measurement of creativity can have practical applications, the approach does seem somewhat impoverished compared with the other viewpoints.

The cognitive approach, that in turn can be split up in several other viewpoints, including the "nothing special view" and the "insight matters view". Central is the statement that cognitive abilities are important for understanding the creative insight. The first view expresses the idea that in fact insight in creative outbursts is only an enlargement of everyday problem solving, while the second states that "insight" is fundamentally different from other processes.

Personality and motivational approaches take into account that personal style, philosophy and desire to solve a problem are important. Being open-minded and nonconventional indeed are folk-psychological concepts that are related with the layperson's notion of creativity.

Social and historical approaches refuse to study creative persons outside their social and historical context. Interesting about this approach is the possibility to explain simultaneous discoveries in science, like the Church-Turing These.

Integrative approaches, as their name indicates, integrate all former viewpoints. They state that multiple individual and environmental factors must converge to yield a creative individual. Although this is a very compelling viewpoint, I shall not use it in depth for the case study developed in the next paragraph, for I am fundamentally concerned with the cognitive part of creative labor.

A case study: Max Planck

One of the most brilliant discoveries of the physical sciences, comparable to the laws of Newton and the evolution theory of Darwin, was the derivation of the quantum-theoretical law of blackbody radiation by Max Planck in 1899, which was the beginning of the quantum era. In order to understand the nature of the problem and the cognitive processes that might have occurred in Planck's mind, I first describe the historical context and the matter under study itself.

In the nineteenth century, physicists paid considerable attention to the so-called "blackbody problem", which is too long a story to tell it here (see Langley et al., 1987, for more details). The aim was to find a formula that fitted all data of a certain experiment. A german physicist, Wilhelm Wien, developed a formula that sufficed. Planck's goal now was to derive his equation from the classical laws of physics. He had worked several years on the problem, and succeeded in 1899. By that time, new findings clearly violated the law of Wien, and Planck immediately succeeded in deriving another law that fitted these data—and of course the former ones. Now he again tried to derive his law from the classical laws. Here, a problem arose. Whatever he tried, it seemed impossible to construct a proof. Now he took a crucial step, deciding to set aside an assumption of the old theories, and reformulating his problem in terms of the new axioms. The derivation following this step was quite straightforward; skilled mathematicians are able to do it in less than five minutes. But the result was a complete new theory: quantum physics.

Several important facts can be seen here. First, there is the expertise of Planck in the field: he had spent more than five years with the problem. Second, Planck was able to put aside the entrenchment his colleagues suffered from. So he decided to drop an assumption of Maxwell's theories, namely that light is continuous. Third, an incubation period clearly occurred: only after having left the problem for a few days, he found the correct way to solve the problem. Fourth, Planck used a heuristic: to combine two approximate equations into one with interpolation. Fifth, his desire to solve the problem made him not go to sleep before the crucial idea was worked out. And sixth, his skill in mathematics helped him to work out the solution completely, for nobody would have believed that this way of representation might have succeeded somehow.

So from a cognitive point of view, Planck's creativity can be analyzed as a combination of setting aside hindrances to the problem solving (especially entrenchment) and using good aids for problem solving (incubation and strong heuristics). Nil nove sub sole, one could claim, we do it—although in a more modest form—almost every day, for example in preparing a meal or planning a route. That's indeed what some researchers claim: according to them, creativity is nothing but extended every-day problem solving. Other stress that in real creative endeavors, something called "insight" is involved. A sudden mental restructuring of the problem causes the emergence of the solution in the mind of the creative person, while others do not have this experience—hence do not find a solution. An important property of this insight is its sudden nature: you cannot force it by thinking harder, or whatever. The very fact that Planck was able to put aside an axiom underlying the old theory can be seen as a sudden insight in the problem he was working on. What followed was only working out his idea: the real creative outburst was over.

Langley et al. (1987) report a family of simulations, called Bacon.x, which is able to repeat the discovery of Planck. The Bacon-approach is data-driven, in that it does not rely on theoretical considerations for finding relations between data (this gap is filled by the successor called "Glauber"). According to my own view, this study does not prove that human scientists work in—even approximately—the same way when they try to construct models for empirical data. The only thing that is interesting about simulation studies like these is that they show that the hard problem is not the application of mathematical tools (for even a relatively stupid computer program performs far much better than human mathematicians) but merely the conceptualization and the formulation of the problem. For most problems holds that they are almost solved when stated correctly. So a nothing-special view of creativity, however appealing from a scientific point of view, seems too weak to explain why some people are able to solve problems in a creative way, while others don't. One question, concerning the nature of "insight", remains.


From what I have described, it is clear that cognitive factors of scientific creativity are embedded in a social and historical context. A phenomenon like entrenchment, for example, can only be understood when one takes into consideration the role of paradigms, which in turn are determined by the scientific community. Creativity, in a sense, is escaping from the paradigm when necessary, but at the same time keeping relevant problem solving machineries from this old paradigm. This establishes also one of the major critics on the kuhnian philosophy, namely that new paradigms do not always differ from the replaced ones in a radical way.

The Langley (1987) study, furthermore, shows that at least certain aspects of the creative cycle are not as mysterious as thought by Kuhn and his intellectual heirs. The ability to change axioms and use heuristics in order to derive laws from empirical data can be modeled by Bacon-like programs. A major disadvantage, however, remains the fact that not all steps in the creativity cycle are apt to formalization. Indeed, one can claim that the essence of the discovery of Planck lies in the reformulation of the problem, namely, the very idea that an agreed-upon axiom may be due to change. Besides, this approach covers only creativity that leans towards problem solving. Not all creative efforts do, however. It might be difficult to let a computer write essays, or poems in the same way.

I conclude that, although creativity and problem solving are sometimes quite alike, it might be useful to draw a distinction between "brute force problem solving" and "problem solving requiring insight". The latter involves expertise, commitment and intelligence.


Bechtel, W. (1988). Philosophy of science: an overview for the cognitive science. New Jersey: Lawrence Erlbaum Associates.

Langley, P., Simon, H.A., Bradshaw, G.L., & Zytkow, J. (1987). Scientific discovery: computational approaches of the creative process. Cambridge: MIT.

Sternberg, R.J. (1996). Cognitive psychology. Forth Worth: Harcourt Brace.

Tip: the left pane works like windows explorer!
Please send your comments to Peter Doomen.
This document was updated 27/05/00.