Powerpoint with speaker notes

CHAPTER 1Two Case Studies in Creativity Creative thinking brings about new things—innovations—ranging from solutions to simple puzzles and riddles to ideas and inventions that have radically altered our world. Creative people are those who produce such innovations, and the creative process consists of the psychological processes involved in bringing about innovations. Figures 1.1A and 1.1B give examples of some of the more impressive products of creative thinking. In Figure 1.1C are some simple exercises that might result in creative thinking on your part. If you had never seen those puzzles and riddles before, and if you solved one or more of them, then you were thinking creatively when you did so—you produced something new. In this book, we will consider the full range of creativity, ranging from solving simple puzzles to producing the seminal innovations shown in Figures 1.1A and 1.1B. We will examine a wide range of recent research on creativity, as well as theories that have been developed to explain the processes involved when people produce innovations. There are many reasons why creativity is a critically important topic for psychologists to understand. First of all, our world has been shaped by the products of creative thinkers. All of our modern conveniences—the telephone and other modes of communication, the automobile, the airplane, computers, and so forth—have been brought about through the creative work of inventors and scientists. Our healthy existences and our ever-longer lives are the result of scientific and medical advances, which are the result of creative thinking on the part of scientists in many domains. Much of the richness of our lives—art, music, drama, literature, poetry—is the result of artistic creativity. Society values greatly the products of creative thinking; we bestow honors, such as Nobel Prizes, on those who have produced such things, and the stories of their lives and accomplishments fill our history books and encyclopedias. By understanding how creative products are brought about, we may be able to increase the likelihood that innovations will occur, thereby making life better for us all. Figure 1.1 Examples of creative thinking (1937): A, DNA: The double helix; B, Picasso’s Guernica; C, Examples of problems In addition, creative thinking is also big business. Our largest and most prestigious corporations, as well as the largest government agencies, are constantly searching for ways to be more innovative, and they pay handsome fees to consultants who will help them achieve new levels of innovation from their employees. Institutions of higher education also take interest in teaching creative thinking. Many university business schools offer courses that are designed to provide business leaders—both those of the future and present-day ones who return for a refresher—with skills that will enable them to solve on-the-job problems. At the grassroots level, one constantly reads accounts of debates concerning the best way to structure our educational system so that children come out as young adults who are able to think creatively. It is therefore important that we have some idea of how creativity comes about, so that we can make decisions concerning how individuals might be helped in dealing with situations that demand creativity. Beliefs about Creativity There are two difficulties in discussing research on creativity. Some people, even people with very deep knowledge of psychological phenomena, come to the subject of creativity with the belief that the topic is so mystical and/or subjective that it could never be captured by psychological methods (Sternberg & Lubart, 1996). In this view, we cannot even define what terms like creativity and creative mean, so as a consequence we cannot even discuss them coherently, much less study them using scientific methods. I have sometimes been asked by other cognitive psychologists—that is, people whose professional lives are involved in bringing difficult-to-study psychological phenomena under scientific scrutiny—how one could ever study creative thinking. They cannot see how one can bring creativity under scientific investigation. One purpose of this book is to demonstrate how something as seemingly difficult to pin down as creativity can be defined and brought under scientific study. Other people, from inside and outside psychology, come to the discussion of creativity with the belief that, even if we can define creativity and begin to study it, there is no purpose in doing so, because creativity comes about as the result of almost supernatural powers. In this view, the people who bring about things like those in Figures 1.1A and 1.1B are basically different from ordinary people: They are endowed with gifts that the rest of us do not have. Learning about what they do and how they do it, even if it were possible to do so, might be of some interest in its own right, but it would not tell us much that would be useful. The differences between the creative greats and ordinary people are in this view assumed to be of two sorts. On the one hand, the greats do not think as you and I do, and the differences between “real” creativity and the activities that you and I carry out are so great as to be unbridgeable. The relatively simple problems presented in Figure 1.1C may require some creativity for solution, but those problems are so different from the situations in which great artists, inventors, and scientists work that entirely different cognitive processes must be involved. So the processes involved when you and I solve such problems would not tell us much about “real” creativity. Second, there are assumed to be critical differences in personality structure between creative and ordinary individuals, and those differences are assumed to play a role in making some people creative. Most psychologists who have developed theories on creative thinking and creative persons take a different perspective on these issues. Although many psychologists believe that creative thinking depends on specific thought processes, they also believe that those processes can be carried out to some degree by all of us. Those who produce great creative advances might be better creative thinkers, but the same thought processes are available to or present in all of us. Similarly, if there is a specific set of personality characteristics that are related to creative achievement, those characteristics are assumed to be present to some degree in many if not all of us; they are simply present to a higher degree in those who produce great creative achievement. According to this perspective, then, creative capacity may to some degree be present in all of us (e.g., Amabile, 1996; Csikszentmihalyi, 1996; Eysenck, 1993; Guilford, 1950; Sternberg & Lubart, 1995). There is also a minority view in psychology (e.g., Perkins, 1981; Newell, Shaw, & Simon, 1962; Weisberg, 1980, 1986, 2003), to which I subscribe, that proposes that the thought processes underlying the production of innovations are the same thought processes that underlie our ordinary activities. From this perspective, the term creative thinking is misleading at least and perhaps a misnomer, because one thinks creatively by using ordinary thinking; one just uses that ordinary thinking to bring about innovations (see also Klahr & Simon, 1999). This does not mean that there is no such thing as creativity, however. There is no doubt that scientists, artists, and inventors, for example, bring forth innovations. It is just that those innovations are based on the ordinary thought processes that we all carry out. One task of this book is to review a representative sample of the various theories of creativity proposed by psychologists and to examine their structure, the predictions that are derived from them, and the evidence for and against them. A further task of this book will be to show that there is a relatively close relationship between creative thinking and other forms of cognition, such as problem solving, reasoning, and the use of memory. That is, the view motivating the presentation in this book is that creative thinking is not different from ordinary thinking—the thinking that we use in carrying out our day-to-day activities. I will show also that the differences in personality and other psychological characteristics between creative individuals and ordinary people may not be very large, and, furthermore, those differences may not be crucial in making creative people creative. Two Case Studies in Creativity In this first chapter, I will discuss two examples of creative thinking at its highest: Watson and Crick’s discovery of the double-helix structure of DNA, the genetic material (Figure 1.1A), and Pablo Picasso’s creation of Guernica, his great antiwar painting (Figure 1.1B). Those two case studies will provide us with “data” of a sort we will have occasion to refer to many times as we consider theorizing concerning creative thinking. At various points in this book, we will discuss the Beatles, Edison, Darwin, the Wright brothers, and Mozart, among other creative thinkers, and the case studies presented in this chapter will provide an introduction to this method. The data from case studies such as those presented here, in conjunction with other results, such as those from laboratory studies of creativity, will allow us to bring an educated perspective to the sometimes conflicting claims made by theories of creativity. The two case studies to be discussed—one from science and one from the arts—are relevant to the question of what differences may exist between the creative processes in those two domains. At first glance, it seems that we are talking about two different things when we talk about creative thinking in the arts versus the sciences. We use different terms to describe the process in the two domains: We talk about artists creating their works (Picasso created Guernica), but we talk about discoveries in science (Watson and Crick discovered the double-helix structure of DNA). There seem to be basic differences in our beliefs concerning the relation between the person and the product in the arts versus the sciences. It is obvious that, if there had never been Picasso, then there would be no Guernica. Similarly, no Beethoven, no Beethoven’s Fifth Symphony. Artistic creativity seems to be an inherently subjective process, as the artist produces something that would not have existed save for the effort of that person. DNA, on the other hand, exists independently of Watson and Crick. If there had been no Watson and Crick, DNA would still have been there, waiting to be discovered, and at some point it would have been discovered. Scientific discovery, in this interpretation, is an objective process: Objects, events, and facts available to all of us are what scientists discover. As we work through the two case studies, I will try to make note of aspects of each that point to similarities, rather than cut-and-dried differences, between creative thinking in science and the arts. Artistic creativity is not as subjective, nor is scientific creativity as objective, as one might think. Creativity in Science: Discovery of the Double Helix In 1953, Watson and Crick published the double-helix model of the structure of DNA, which has had revolutionary effects on our understanding and control over genetic processes. As one example of the impact of Watson and Crick’s work, there has in recent years been much controversy over the possibility that scientists have succeeded or will soon succeed in cloning human beings. This possibility is but one of the remarkable developments that can be traced directly back to the discovery of the double helix. Geneticists, biologists, and other scientists, including Watson and Crick’s teachers, had for more than 50 years been pursuing the question of the composition and structure of the genetic material (Olby, 1994). Watson and Crick succeeded in formulating a model of the structure of DNA after approximately one and a half years of work, and several misdirected attempts. Other research groups were at that time also working on the structure of DNA, and Watson and Crick were not the first to publish a possible structure, but theirs was ultimately judged to be correct (Judson, 1979; Olby, 1994). DNA was a discovery of wide sweep, which involved a large number of contributors. Examining this discovery will provide information concerning how scientists become focused on the questions that they study. What, if anything, does the creative individual know that leads him or her to the important questions, the answers of which will change our world? Studying the discovery of DNA also will allow us to address the critically important question of how different scientists, while studying the same phenomenon, wind up taking different approaches, so that one is successful while the other is not. That is, we will begin to gather information on what, if anything, separates the individual who produces the important scientific discovery from the one who does not. Historical Background DNA was discovered in the middle of the nineteenth century, and by the early twentieth century it had been shown to be present in all cells (Stent, 1980, p. xiii). DNA is made up of a number of different components: a phosphate group, constructed around phosphorus; a sugar; and four different nitrogen-rich bases, adenine, cytosine, guanine, and thymine, abbreviated as A, C, G, and T. (See Figure 1.2.) One phosphate, one sugar, and one base form what is called a nucleotide, the basic unit out of which DNA is constructed. There are four different nucleotides, differing only in their bases. Thousands of nucleotides, strung together, form the complete DNA molecule. So the basic structure of DNA could be described as a polynucleotide, built out of a set of building blocks that repeat again and again. It was not until the late 1940s that researchers began to agree that DNA was the genetic material. Although DNA is found almost exclusively in the chromosomes, which are the sites of the genetic material, there is more protein than DNA in chromosomes, which led to the belief that protein might be the critical material. In addition, it was also thought originally that DNA was a relatively simple molecule, too simple to carry out the tasks required of genetic material (Olby, 1994). It was initially believed that the DNA molecule was simply a tetranucleotide, that is, that the complete molecule consisted of one of each of the four nucleotides, and nothing else. It was soon shown that the molecular weight of DNA was much larger than only four nucleotides, but researchers then assumed that the large molecule simply consisted of the four nucleotides repeating monotonously in the same sequence. In both those analyses, DNA was relatively simple in structure. The function of the genes is to direct synthesis of proteins, which are complex molecules. It seemed to follow from this that the genes would have to be complex as well. Therefore, DNA with its simple tetranucleotide structure could not serve that purpose. It was assumed by some that DNA was present in the nucleus only to serve as a “stretcher,” so that the protein genes could be straightened out to carry out their functions. Figure 1.2 DNA information: A, Nucleotides; B, Chemical structures of the four DNA bases as they were often drawn about 1950. In the 1940s, several different sorts of evidence pointed to DNA as the genetic material. In 1928, Griffith had shown that injection of purified material from virulent pneumococcus bacteria (that is, bacteria that caused illness—in this case, pneumonia) into heat-killed bacteria that were benign (i.e., that no longer produced pneumonia) could transform those benign bacteria into virulent ones (Olby, 1994). Most important, this transformation could also be passed down to subsequent generations, which indicated that the genetic material of those benign bacteria had been altered. The critical question then centered on the chemical composition of the extracted material, or “transforming substance,” and in 1944, Avery and colleagues identified it as DNA. Furthermore, since the transformation could be passed down genetically, identifying DNA as the transforming substance indicated that DNA might be the genetic material as well. Also during this decade, a study by Hershey and Chase examined the mechanisms whereby viruses attacked and killed bacteria, in order to gather information about the composition of the genetic material (Olby, 1994). When a virus attacks any organism, including a bacterium, it takes over the reproductive mechanism of the host organism’s cells and uses them to reproduce itself. A virus is essentially genetic material encased in a shell. Hershey and Chase used radioactive phosphorus (which is incorporated into DNA) and radioactive sulfur (which becomes part of protein) in order to produce strains of viruses with different radioactive “signatures.” They then traced the fate of those radioactive chemicals after the marked viruses had infected bacteria. In a methodological innovation that became legendary, the researchers used a kitchen blender to separate the infected host bacteria from the shells of the viruses that were attached to them. The results indicated that when viruses attack bacteria, the viral DNA is introduced into the host bacteria, while the shell of the virus, which is made up of protein, stays outside the host. The protein shell seemed to serve as a kind of hypodermic that injected the viral DNA into the host. This result provided strong support for the idea that DNA was the material carrying the genetic information from the virus to the bacteria. Finally, in a series of chemical studies of DNA, Chargaff showed that the tetranucleotide hypothesis of the structure of DNA was incorrect (Olby, 1994). He analyzed the relative proportions of the various bases in DNA from different organisms. The results, shown in Table 1.1, contradicted the tetranucleotide hypothesis in two ways. First, within each species, the proportions of the various bases were not equal, and second, the ratios of the various bases differed in different species. So there was much more variability in DNA than researchers had believed, perhaps enough variability for the DNA molecule to function as the carrier of the genetic code. Another interesting finding reported by Chargaff was that, even though the proportions of the bases differed from species to species, in each species there seemed to be equal proportions of A and T, as well as equal proportions of G and C. This pair of findings, which became known as “Chargaff’s ratios,” turned out to be significant in the construction of the double helix. Table 1.1 Chargaff’s data on the chemical composition of DNA from different species (Chargaff’s ratios) Species Base composition (%) A T G C Human (liver) 30.3 30.3 19.5 19.9 Bacterium (tuberculosis) 15.1 14.6 34.9 35.4 Sea urchin 32.8 32.1 17.7 18.4 As a result of this constellation of findings, by the 1950s many researchers, although not all, had come to believe that DNA rather than protein was probably the genetic material. Watson (1968, p. 31) notes this when he comments on meeting Crick, “Finding someone … who knew that DNA was more important than proteins was real luck.” Once Watson had met Crick, the central question for the two of them concerned the way the DNA molecule was structured. As we can see, it is not always obvious to scientists what the important questions are in a discipline. Some first-class researchers were pursuing the study of the structure of the proteins in the cell nucleus as the basis for understanding the structure of the genetic material. Those individuals obviously had no chance of discovering the structure of DNA. So when Watson says that he was lucky to have found a kindred spirit in Crick, we can understand the significance of that statement, and we can ask where that commonality of interest came from. One sometimes sees it stated (e.g., Getzels & Csikszentmihalyi, 1976; see also chapters in Runco, 1994) that individuals who make creative discoveries have an ability or intuition—a skill sometimes called problem finding—that allows them to find a critically important problem to work on, where other less-creative individuals see nothing of importance. The latter individuals therefore spend time and effort studying problems that may lead nowhere, or at least will lead to less important results. So the question of how Watson and Crick focused on DNA, to which we now turn, is one with broad implications. Watson Gets to Cambridge Watson and Crick’s collaboration began in autumn 1951, when Watson joined the staff of the Cavendish Laboratory at Cambridge University, where Crick was working on his PhD (see Table 1.2). Watson already had a PhD in genetics; Crick had been trained as a physicist before World War II, but he was then working toward a PhD in biology, studying the structure of hemoglobin using X-ray diffraction techniques. Even though Watson and Crick had never met, they had intellectual links. Watson had received his PhD in genetics at Indiana University, working under the direction of Salvador Luria, who, along with Max Delbrück and Alfred Hershey (of the Hershey-Chase kitchen-blender experiment discussed earlier), was one of the founders of the “phage group” (see Figure 1.3). This was a group of scientists who were interested in studying bacteriophages, viruses that devour bacteria, in order to understand the genetic mechanisms in all organisms. Phage comes from the Greek for eat; the kitchen-blender study of Hershey and Chase was a study of the mechanisms of reproduction of bacteriophages. Delbrück was a physicist who had moved into biology in search of new research areas (Olby, 1994). He had also convinced other physicists of the importance of biological questions, and a number of other physicists followed him into biology after World War II. Table 1.2 DNA Timeline Date Research team and activity Watson and Crick Wilkins/Franklin/Pauling 1951 Spring Watson attends conference; sees Wilkins’s X-ray photo. Wilkins presents X-ray photo at conference. July Wilkins visits Cambridge, tells Crick DNA was probably helix. Fall Watson joins Cavendish. Mid-September Franklin: X-ray pictures of DNA; fibers made wetter stretched and yielded a new pattern (B form); Wilkins and Stokes: photos showed “helical features.” October 31 Cochran and Crick: helical X-ray pattern theory. November 9–11 Wilkins visits Cambridge; Wilkins: DNA helical. November Franklin’s notes prior to colloquium: structure helical in both states; presents evidence. November 21 Watson attends Franklin’s colloquium, takes no notes, misses much of what she says. November 22 Watson reports to Crick what he remembers from Franklin’s colloquium. Mistakenly recalls amount of water (in B form). Crick: only a few structures are compatible with Crick/Cochran theory of helices. November 26–28 Begin building chain-inside models. 3 chains fit density data; “hole” for water; held together by Mg+ ions. November 28 King’s group comes to see model; points out problems. Franklin: backbones outside; too little water. 1952 Early in year Wilkins letter to Crick: phosphates outside. April 10–19 Franklin: new A-form pictures—it is not a helix. May 1 Franklin tells Watson that DNA is not helical. May 1–6 Franklin: new B-form X-rays (49 & 53); clearly helix. Spring Watson learning crystallography. TMV photo: helix. Wilkins, convinced by Franklin that A form not helical, decides B is not either; stops work in frustration. Late spring Crick tells Franklin A-form photo might be misleading. May 24–27 Chargaff visits, explains results. July 18 Franklin announces “death” of DNA helix. Summer Franklin, with Gosling, begins Patterson synthesis of A form (analytical approach). November 26 Pauling, using density measurements, calculates number of chains to be 3; result surprises him. End Nov. Franklin writes up work, describes unit cell of molecule. December 15 Committee (including Perutz) visits King’s. December 25 Pauling invites colleagues to see 3-strand model of DNA. 1953 January 2 Pauling submits note on DNA to Nature. January 19 Franklin models figure-8 structure. January 28 Pauling’s manuscript arrives; model seems incorrect. January 30 Watson goes to King’s with Pauling’s manuscript, sees Franklin’s B-form photo (51), decides that 2-strand models are not ruled out by density data. February 2 Franklin: “Objections to figure-8 structure” in notebook February 4 Watson builds models; 2 fruitless days on chains inside. February 5 Watson tries 2-chain-outside model; easy without bases. February 10 Crick sees Franklin’s report, deduces anti-parallel chains (based on his thesis), led to 2 chains. Watson & Crick build backbones with 36° rotation; Watson provides deductive evidence for 2 chains. Franklin working on helix for B form. Feb. 16(?)–19 Watson reads on bases; DNA held together by H bonds; builds “like-with-like” model. February 20 Like-like torn to shreds by Donohue; wrong tautomers. February 23 Franklin: helix of B form to probable helix of A February 28 Watson discovers base pairings by manipulating models on desktop. Figure 1.3 Intellectual links between Watson, Crick, and Williams In 1944, Erwin Schroedinger, a physicist and one of the founders of quantum mechanics, published a book called What Is Life? in which he discussed how the then-unknown genetic material might be structured and how it might transmit information and direct the reproduction and other activities of cells. He proposed that the genetic material might be constructed out of small units that repeated over and over in various combinations, with the various combinations serving as letters in a kind of alphabet used to communicate information from the gene to the mechanisms in the cell. Schroedinger was familiar with and influenced by Delbrück’s ideas (Stent & Calendar, 1978, p. 26), and his book can be looked upon as a popularization of those ideas. Schroedinger’s book was read by many physicists who then became interested in biological questions in general and genetics in particular, and it was also read by biologists. One such physicist was Crick, and another was Maurice Wilkins, a friend of Crick’s who was studying the structure of DNA at King’s College, in London, and who, as we will see, played a significant role in the discovery of DNA. Luria and Watson also read Schroedinger’s book. These links are shown in Figure 1.3. Thus, we can understand in a straightforward manner Watson and Crick’s common interest in the structure of DNA: It came directly out of their common intellectual heritage. In this case, and in other cases to be discussed later, the problem that turned out to be important and fruitful was almost thrust upon Watson and Crick—and also upon other researchers who played important roles in the story as it unfolded—by the intellectual milieu in which they were raised. At Luria’s suggestion, Watson went to Europe in the fall of 1950 to study the chemistry of the nucleic acids, because Luria felt that acquiring that knowledge would help Watson gain an understanding of how genes function. Watson did not find the work interesting, however, and he was looking for a more stimulating environment in which to work, especially a place that might provide an opportunity to work directly on the question of the structure of the genetic material. In the spring of 1951, Watson attended a conference in Naples, at which Wilkins presented a paper. During his talk, Wilkins projected a slide of an X-ray photograph of DNA (see Figure 1.4), which completely captivated Watson (Watson, 1968). The fact that one could photograph DNA using X-rays meant that one could make a crystal out of it, which in turn meant that DNA must have a regular structure, which might be analyzable without an impossible amount of work. Watson then decided that he would work someplace where it would be possible to carry out X-ray analysis of DNA. There were only a few places where one could carry out such work; one was the Cavendish Laboratory at Cambridge University, which had been world-famous for its X-ray work since early in the twentieth century. Watson was able to arrange an appointment there, and he joined the staff in the fall of 1951. Watson and Crick’s Collaboration Soon after Watson’s arrival at the Cavendish, he and Crick made two early decisions about DNA that were very important in setting them on the path to success. First, they decided to try to build a model of the structure, as shown in Figure 1.1A. Deciding to build a model led to the question of the shape of the molecule. Was DNA a long chain of nucleotides, one attached to the next? Was it a closed ring, with one nucleotide attached to the next until one came back around to the point where one began? Was it shaped in some other way? One of those choices had to be made; so, in order to initiate their model-building work, Watson and Crick agreed to begin with the working assumption that DNA might be in the shape of a helix. This helical assumption was, of course, correct in general terms, and it, along with the model-building orientation, put Watson and Crick solidly on the path to success. Other researchers who were also working at that time on determining the structure of DNA, including Wilkins, had not decided to build models and/or had not made the assumption that DNA was helical, and they were therefore slower in finding the structure (Judson, 1979). Figure 1.4 DNA X-ray photo We are thus faced with another question of critical importance in the understanding of the creative process of Watson and Crick: Where did they get those two critical ideas—that they should build models and that DNA might be helical? Did they have some magical intuition, some creative sixth sense, that led them along the correct path, where others did not know to tread? It seems not. Both of those critical assumptions were based relatively directly on the work of Linus Pauling, a world-famous chemist who had recently solved the problem of determining the structure of the protein alpha-keratin (Olby, 1994; Watson, 1968), which forms fingernails and hair, among other things. Pauling had proposed that alpha-keratin was helical in shape, and he had built a model of the structure to show how all the atoms fit together. He had also published an unheard-of seven papers in a single issue of a professional journal, in which he and his associate presented the alpha-helix and evidence to support it. Pauling’s success had stirred the scientific community, and especially the Cavendish lab, where similar techniques were being used to investigate proteins. It was felt by some, including some at the Cavendish, that Pauling’s success was at the expense of and an embarrassment to the Cambridge group. However, the fact that Pauling could be seen as a rival of the Cavendish group did not stand in the way of Watson and Crick’s seeing the potential usefulness of his research methods and ideas in the analysis of DNA. Like DNA, alpha-keratin is a large organic molecule—a macromolecule. In addition, alpha-keratin is similar to DNA in one critical way: Proteins are constructed out of large numbers of repeating units, or peptides, which are linked together to comprise the large protein macromolecule. Proteins thus are polypeptides, a structure similar to the polynucleotide structure of DNA. The reason Watson and Crick chose Pauling’s work as the basis for their own is easy even for us non-molecular-biologists to understand: The domains are closely linked. We have here a clear example of what can be called continuity in creative thinking: Watson and Crick built their work on the past. That is, the new work was continuous with the past. Continuity is also a component of our ordinary thought processes, of course, because in our ordinary thinking activities we are always using what we know as the basis for decision making and behaving. Wilkins also contributed to Watson and Crick’s adoption of the assumption that DNA was helical. Around 1950, Wilkins was carrying out what was probably the most advanced work on DNA in the world (Olby, 1994). He had been studying the properties of DNA in response to light, when he accidentally produced long fibers of DNA. When his assistant exposed those fibers to X-rays, the results were the best diffraction patterns that had been seen, one of which was the photograph that had excited Watson at the Naples conference. From the X-rays, one could deduce the diameter of the molecule and the distance between consecutive bases. However, one could not determine any more specific information about the shape of the molecule or how it was constructed. A researcher skilled in interpreting such X-rays could also determine that there was an underlying pattern in the structure of the molecule, which repeated as one went along it. Knowing about the double helix, we can understand that the repetition of the pattern comes about because the helix cycles and comes around to the same place as one travels along the molecule. As an analogy, when you enter a spiral staircase and start to ascend, you will reach a point as you go around at which you will be standing directly above the first step of the staircase, where you began. That is the point at which the structure begins to repeat. Of course, at that time, no one knew why there was a repeated pattern; all that could be seen from the X-ray photograph was that a repetition of some sort occurred. In the summer of 1951, before Watson arrived at the Cavendish, Wilkins had given a talk at Cambridge during which he discussed the possibility that DNA was helical (Judson, 1979; Olby, 1994). At the time, judging by measurements of its density, as well as other data (some of which turned out to be incorrect), he felt that it might be a single strand. In the fall of 1951, after Watson joined the Cavendish staff, he, Crick, and Wilkins met and discussed DNA; they agreed that the molecule was probably helical. By the time of this meeting, Wilkins was leaning toward a theory, based on new data, that there were three strands. It is important to note here that seeing a helical pattern in the X-ray in Figure 1.4 is not the same as seeing a smile on the face of your friend. That is, the diffraction pattern produced by exposing the DNA crystal to X-rays does not look anything like a helix. There is no visible evidence that can be directly seen as being from a helix. One must understand X-ray crystallography in order to see a helical pattern in an X-ray photograph. One must first make certain assumptions about how the X-ray beam will be broken up by a crystallized molecule of a given shape and structure. Then one can make predictions about what the diffraction pattern will look like, although, as just noted, the pattern will look nothing like a helix. This is a crucially important interpretive skill, and Watson and Crick were in a unique position to develop that skill. Soon after Watson’s arrival at Cambridge, Crick, in collaboration with William Cochran, another member of the Cavendish staff, had carried out theoretical work concerning the mathematics of the interpretation of helical X-ray diffraction patterns, which proved to be critically important in enabling Watson and Crick and others to interpret and make sense of X-ray data (Judson, 1979; Olby, 1994; Watson, 1968). This point is relevant in a broader way for our understanding of scientific creativity: More than simple observation is involved in scientific research. Scientists often draw conclusions from very indirect evidence, so their knowledge and comprehension are critical to their success. This is a step away from the notion of science as the simple discovery and study of objective facts. One could say that the helical shape of the DNA molecule was not an objective fact, in the sense that it was not sitting there to be observed. One might go even further and say that it was a “created fact.” On the Origins of New Ideas So, if one asks where new ideas come from, in this case the answer is that the new ideas used by Watson and Crick—the idea of building a model of DNA and the idea that the structure might be a helix—came about first of all through the adoption and extension of already-existing ideas that had been developed by someone else (Pauling) in dealing with a similar problem in a closely related area. This view was also supported by Wilkins, a researcher respected by both Crick and Watson. We can also see continuity in Wilkins’s thinking about DNA: He used experimental techniques on DNA—the response of the fibers to light—that had been used by earlier researchers in the study of other large organic molecules. It is also interesting to note that when Watson and Crick adopted the working assumption that DNA was helical, they changed the structure of their problem. That is, now they did not have to sift without focus through data and ideas in order to find something that might give them direction. Rather, they were now examining all available information from the perspective of the assumption that DNA was helical, which means that concepts and ideas were directing their work. The situation is similar to working on a jigsaw puzzle with a picture of the completed puzzle as opposed to working without one. The latter is the position that other researchers were in at about that time; that is, they had many pieces of data, and the task was to determine, in a necessarily piecemeal manner, how they fit together. Once Watson and Crick had adopted the helix assumption, they could look at each piece of data and ask, “What, if anything, does this tell us about the structure of the helix?” The relative difficulties of the two sorts of problems seem obvious. In addition, adopting the helical assumption led Watson and Crick to raise questions about the available data. That is, they questioned whether some pieces of data were accurate, because those data conflicted with the helical idea. Other researchers, who lacked the helical assumption, were forced to treat all pieces of data as equal—as we shall see—and that sometimes led them astray. Adopting Pauling’s method and his conclusion did not settle all the issues facing Watson and Crick, however. Before they could start to build a model, they had to make several further decisions. Experimental evidence, based primarily on X-ray pictures of DNA taken by Wilkins, was consistent with the idea that DNA might be a helix, but that evidence did not specify how big it was. The X-ray evidence could be used to calculate the diameter of the molecule, but more details than that were impossible to ascertain. Therefore, Watson and Crick did not know exactly how many strands or backbones the helix contained. As we now know, DNA contains two strands (it is a double helix), but when Watson and Crick started working, evidence indicated only that the molecule was thicker than a single strand. There might have been two, three, or four strands, for example, as shown in Figure 1.5A. The number of backbones in the model that they planned to build was the first decision to be made. The second concerned where to put the bases, the four different compounds (A, C, T, and G) that we now know form the rungs of the spiral staircase of DNA, and which carry the actual genetic information (see Figure 1.5B). The specific sequence of bases determines which proteins are constructed by the cell, which is how the specific genetic information gets translated into the physical structure of the organism. When Watson and Crick started their work, the location of the bases was not known; they could have been inside the helix—that is, between the backbones—or protruding from the outside, with the backbones in the center, as shown in Figure 1.5B. Another question that could not be answered at the time concerned the specific angle or pitch of the spiral of the helix. Once can visualize a helix by imagining a spring, as shown in Figure 1.5C. The spring can be either tight, so that the spiral is almost flat in cross-section, or stretched open, in which case the angle of the spiral formed by the backbones is steep. This angle is called the pitch of the helix, and it was unknown at the time Watson and Crick began their work. Finally, the specific way the backbones were structured was not known; it was assumed that they were structured as in Figure 1.2A, with one nucleotide linked to the next, but the details were not known. So, before Watson and Crick could carry out specific model building, they needed several additional pieces of information. Franklin’s Colloquium On November 21, 1951, Watson attended a colloquium, or professional talk, given at King’s College by Rosalind Franklin, who was also carrying out research on the structure of DNA (Judson, 1979; Olby, 1994; Watson, 1968). Franklin was a knowledgeable crystallographer, but her experience had previously been limited to studying the structure of coal. The study of DNA was her first exposure to biological molecules. Franklin and her assistant had recently produced X-ray photographs of DNA after it was exposed to humidity. This was a task that Wilkins had been trying to carry out earlier, but it was Franklin, with her deeper experience with X-ray diffraction techniques, who was successful at it. Those photographs of what was called the wet or B form of DNA were especially informative to the knowledgeable researcher, producing an exposure pattern that was strongly supportive of a helical structure (see the X-shaped pattern in Figure 1.6, a further example of the indirect nature of scientific “facts”). We shall shortly see more evidence of Franklin’s success. At this colloquium, Franklin discussed her most recent work on DNA. Franklin had not been at King’s for very long, and relations between her and Wilkins were not good. When Franklin came to King’s, she thought her assignment was to carry out research on the structure of DNA. However, Wilkins believed that that was his task, and therefore he felt that she was an interloper who was trying to push herself into his research area. This resulted in some hard feelings between the two and some resentment toward her from Watson and Crick. Figure 1.5 DNA: A, Multiple-strand helices; B, Possible positions of bases; C, Pitch of helix Figure 1.6 Franklin’s B-form X-ray In her talk, Franklin presented recent results of her research on DNA. There is a fascinating subtext concerning this presentation. In her notes for the talk, which have been preserved, Franklin discussed how her X-ray data and other results constrained the structure of DNA, which she described as a large helix. She also noted how much water DNA absorbed on exposure to humidity. In relating the content of Franklin’s talk to Crick, Watson told him nothing about Franklin’s ideas concerning the helical structure of DNA (Judson, 1979; Olby, 1994; Watson, 1968). This is somewhat surprising, since nothing had been said earlier about Franklin’s believing that DNA was a helix, so one would think that Watson would have been struck by her describing the molecule as a helix. Olby, in discussing Watson’s omission, speculates that he might not have perceived such comments by Franklin as adding anything new to what he and Crick were already thinking about and what they had also discussed with Wilkins. In addition, Watson usually did not take notes when he attended talks, relying instead on his memory, as he did at Franklin’s talk, and this might have affected what he could report. Also, Watson was new to the type of analysis being carried out by Franklin and Wilkins at King’s (and also at the Cavendish), and he later said that much of what Franklin reported went over his head. I have another interpretation, based on no additional information, of Watson’s failure to report to Crick what Franklin wrote in her notes about the helical structure of DNA: Perhaps she decided not to make those ideas public at the talk. (It should be noted that Olby briefly considers this possibility and rejects it; see Olby, 1994, p. 351.) There are several reasons why I believe that this speculation is not completely without base. Wilkins did not recall Franklin discussing helices (Olby, 1994, p. 351), although he was not certain about it. It seems to me very unlikely that Watson would have either ignored or forgotten a discussion by Franklin of the structure of a DNA helix. Since she had never made such claims in public before, they would surely have captured Watson’s attention, especially since he specifically went to King’s to hear her talk. Since at that time Franklin was not on good terms with Wilkins and other members of the staff at King’s, she might have felt that her ideas would not meet with a sympathetic hearing or that they might help others advance their work, both of which she might have wanted to avoid. In any case, what Watson seems to have taken away from the King’s talks was some information about the density of DNA, as well as some information concerning how it behaved when exposed to moisture. Since at that time he was not very sophisticated in interpreting X-ray-diffraction photographs, he did not get much useful information out of Franklin’s B photo. Consideration of Franklin’s colloquium adds another fascinating plot line to the story of the discovery of double helix. In the fall of 1951, Franklin seemed to have moved far down the path to the double helix, perhaps even farther than Watson and Crick. And yet, she did not formulate the structure; Watson and Crick did. Why was she not successful? We will discuss this question once we work our way through Watson and Crick’s story, because it provides important information concerning the factors that separate the “greats” from the “near greats.” The Triple Helix When Watson reported his recollections of Franklin’s talk to Crick, the latter concluded that only a small number of helical structures would fit both the data and Crick and Cochran’s theory concerning how the X-ray pattern should look: structures of two, three, or four strands. Here is another example of the indirect nature of some scientific facts. Drawing on what they knew from experimental work, including Franklin’s, and from discussions with other investigators, most importantly Wilkins, in November 1951 Watson and Crick first built a three-strand model of DNA—a triple helix—with the bases on the outside (see Figure 1.5). That is, in the two critical decisions they had to make—number of backbones and location of the bases—they made incorrect choices. However, each of those choices was reasonable at the time (Judson, 1979; Olby, 1994; Watson, 1968). One piece of evidence to support the idea of three strands, for example, was the calculated density of DNA. From the X-ray photos, the overall dimensions of the molecule could be determined. The weight could also be determined, so one could divide the weight by the volume calculated from the dimensions to determine the density of DNA. If one knows the density of the molecule and if one makes some assumptions about the structure, one can draw conclusions concerning the number of strands it must contain. The evidence concerning density turned out to be incorrect, which meant that the conclusion concerning number of strands was also incorrect, but Watson and Crick could not know that when they built the triple helix. Thus, Watson and Crick used calculation and deductive logical reasoning to develop their new triple-helix structure. This is another example of creative thinking using components we all use in our ordinary thinking. In the case of the triple helix, at least, we do not need to call on any exotic thought processes to understand what Watson and Crick produced. Placing the bases on the outside of the helix was done for different sorts of reasons, but again ones that followed from what Watson and Crick knew at the time. One reason Watson and Crick put the bases on the outside of the helix, rather than between the backbones, was that they could not figure out how to fit the bases inside the rigid backbones (Watson, 1968). The bases are of different sizes (see Figure 1.2B), so in order to fit base pairs inside the helix, the backbones would have to bend back and forth, and this would not make for a rigid structure, which they knew that DNA was. Putting the bases outside eliminated this problem. There was another hurdle to putting the bases inside. Molecules sometimes appear in nature in more than one form, called tautomeric forms. Crick believed that at least some of the bases might appear in more than one form, which meant that the number of possible base combinations was even larger than they had originally believed, which made even more remote the possibility of constructing a uniformly shaped molecule with the bases inside. Again, putting the bases outside eliminated this difficulty. Here too we have several examples of creative thinking based on logic. Placing the backbones at the center of the triple helix raised a problem, however, because the phosphates would be negatively charged and would repel each other, meaning that the molecule would blow apart. In order to hold together the three strands of the helix, Watson assumed that magnesium ions served to link them. The positively charged magnesium ions would serve to hold together the chains of the helix. This idea was based on an analogy to other molecules, where magnesium played such a role (Judson, 1979; Olby, 1994). There were two potential problems with this assumption. There was no evidence of magnesium in DNA, and there was no evidence of ions in DNA with the +2 charge of magnesium ions. The triple helix, then, was a creative response to the information available to Watson and Crick at that time, and it demonstrates several aspects of the creative process. It is also interesting to note that at about this time Bruce Fraser, a physics student at King’s, built a triple-helix backbones-inside model of DNA, similar to that of Watson and Crick (Judson, 1979; Olby, 1994). Franklin’s response to Watson and Crick’s model was that any modeling was premature, because not enough data were available. Triple Helix to Double Helix The same week that the triple-helix model was constructed by Watson and Crick, the King’s group was invited to see it. The meeting was a disaster for Watson and Crick, because they were informed of many problems with their new model (Judson, 1979; Olby, 1994; Watson, 1968). First of all, the magnesium ions postulated by Watson to hold the backbones together would probably, according to the King’s group, be surrounded by water due to the ions’ charge, and therefore the magnesium could not hold the strands together. Second, it turned out that Watson had drastically misrecalled the report by Franklin of the amount of water in the molecule after exposure to moisture, and the correct amounts changed their calculations and made parts of the structure unlikely. Finally, at this meeting Franklin presented arguments for the backbones’ being on the outside of the structure. Her evidence came from information from her X-ray photos, among other sources. As a result of this meeting and the embarrassment it brought to the Cavendish lab, Watson and Crick were told by Dr. Lawrence Bragg, the lab’s director, not to work on DNA and to leave that area to the King’s group. So, as we will see frequently in studying creative work, Watson and Crick’s first attempt was a failure. It took more than a year, from late fall 1951 to early 1953, before they formulated the correct model of DNA. During that time, Watson and Crick kept thinking about DNA, although more privately, and Watson learned how to carry out crystallography and to read X-ray photos, which would be critical to their success. Most important, he took X-ray photos of tobacco mosaic virus (TMV), a virus that attacks tobacco plants, and produced evidence that TMV was helical in structure. In this way Watson acquired expertise in reading helical patterns in X-rays. Watson and Crick also acquired several important pieces of information, some of which directed them away from the triple helix, and others of which pointed to the double helix. One example of information that led away from the triple helix came about because Watson was not convinced by the King’s group that using magnesium ions to hold the structure together was a problem. He tested a sample of DNA for the presence of magnesium and found none, which ended that possibility. Watson and Crick also found that three-strand models could not be made to fit together without violating some basic laws of chemistry. Unless they had missed something, this led to another dead end, which also indicated that a three-strand model was probably wrong (Judson, 1979; Olby, 1994; Watson, 1968). At the end of January 1953, Watson and Crick saw a copy of a paper published by Pauling and his associate Robert Corey that contained a proposed structure for DNA. Watson and Crick read the paper with trepidation, because they assumed that Pauling, with his record of past successes in similar areas, had solved the problem (Judson, 1979; Olby, 1994; Watson, 1968). To their surprise, they saw that Pauling’s model was a triple helix, at least superficially like their own earlier model. Watson and Crick also saw relatively quickly that its chemical structure was incorrect, and Watson got confirmation of this conclusion from organic chemists at Cambridge. The fact that Pauling was incorrect gave Watson and Crick a little time to work, but not much, because they assumed that as soon as Pauling learned of the problems with his model, which would not take very long, he would correct it. On Friday, January 30, 1953, more than a year after being forbidden to work on DNA, Watson visited King’s with a copy of Pauling’s paper. He met with Franklin and then with Wilkins. The meeting with Franklin was unpleasant, according to Watson (Judson, 1979; Olby, 1994; Watson, 1968), because she asserted that there was no evidence that allowed anyone to conclude that DNA was helical. As we know, she had at least been speculating about such a possibility over a year before, but, as we shall see shortly, things had changed. In addition, she had shown copies of her X-ray photos to Pauling’s associate Corey (Olby, 1994, p. 396), so she might have known what they were thinking about. By this time, relations between Franklin and Wilkins were very poor, and she also felt negatively toward Watson and Crick, which may partially explain her disparaging response. Wilkins showed Watson some new X-ray photos, including a new photo of the B form, taken by Franklin in May 1952 (Franklin’s photo 51). Photo 51 was not only new to Watson but also extremely precise and informative, and he examined it with excitement. There was no doubt in his mind that it was from a helix-shaped molecule. Also, with his newly acquired expertise at reading and interpreting (and, therefore, remembering) X-ray photos, he was able to get much additional information out of the photo, and he made a sketch of it from memory later that evening. Wilkins also told Watson that Franklin had found that exposing the DNA fibers to moisture (going from A to B) resulted in a 20 percent increase in their length. One of the main sources of evidence for the existence of three strands was the amount of water the DNA molecules could absorb, but there were questions about the accuracy of those results, which meant that the case for three strands was not that strong. The new information, along with information from the new B photo, enabled Watson to draw the conclusion that the molecule contained two strands, not three. There has arisen an interesting question concerning the propriety of this conversation between Watson and Wilkins, which provided information that obviously helped greatly in the development of the double helix. The question is whether Franklin’s photo and results were private information, which Wilkins should not have shared. The poor relations that Franklin had at that time with Wilkins and with Watson and Crick raise the question of how much Wilkins should have told Watson (see Sayre 1975, and Elkin, 2003, for further discussion). The next day, Saturday, January 31, 1953, Watson convinced Bragg, the director of the Cavendish, that Pauling was getting close to determining the structure of DNA and that no one at King’s was doing much in the way of modeling the structure (Judson, 1979; Olby, 1994; Watson, 1968). Watson also presented what he and Crick then knew about the structure, and Bragg gave them permission to return to building models of DNA. When Watson and Crick met and went over the new information obtained by Watson the day before, Watson was ready to concentrate solely on two-strand models, but Crick felt that they should not completely abandon three-strand ones. Watson, still concerned about what to do with the bases, then spent two days unsuccessfully trying to construct two-strand models with the backbones inside. Again, he could not get the backbones to fit together without violating laws of chemistry. He then finally gave up on the backbone-inside structures, which raised two problems: How could the bases be made to fit inside, and how were the two strands connected? In the second week of February, Watson and Crick obtained another critical piece of information. As noted earlier in the analogy of the spiral staircase, a helix has a repetitive pattern as one moves along it. This repeating component is called the unit cell of the structure. Max Perutz, a senior researcher at the Cavendish, gave Watson and Crick a report that the researchers at King’s had prepared for an outside committee, of which Perutz was a member, that was charged with evaluating the progress of the research being carried out there. In this report, which was a public document, Franklin discussed the shape of the unit cell of the molecule, which enabled Crick to deduce that the backbone chains were antiparallel—that is, they ran in opposite directions—and also enabled him to determine the pitch of the helix. Crick was able to make those critical deductions at least in part because, in his work on hemoglobin (under Perutz’s direction), he had been working with a unit cell of the same structure as that reported by Franklin for DNA (Judson, 1979). Thus, he was immediately able to see the consequences of that structure. Those were almost the last pieces of information that Watson and Crick needed. Constructing the Double Helix By mid-February 1953, Watson and Crick had built part of a model that had one spiraling backbone but no bases inside. Since the backbones were antiparallel, one chain was all that was necessary, since the structure of the other chain could be determined from it. The development of this model was based on the contributions of many people (Judson, 1979; Olby, 1994; Watson, 1968). Watson was the strongest advocate for the two-strand structure. The outside backbones were based on Franklin’s work. Crick contributed the antiparallel structure of the backbones and the pitch of the helix. Wilkins had contributed to the general orientation of the work. Watson and Crick then began to consider specific possible arrangements of the bases inside the backbones, and Watson first tried pairs of the same bases—that is, A with A, C with C, and so forth—as the rungs of the staircase, a scheme that was called like-with-like or like-like. He found some evidence in the literature, including some work done at the Cavendish, that like-with-like pairing might occur between bases. In addition, several of Watson’s professors in graduate school had discussed like-with-like pairings in other contexts. So, even though the different sizes still raised problems, Watson thought the possibility was worth looking into. However, the like-with-like structure was torn to shreds by Jerry Donohue, a postdoctoral fellow at the Cavendish who had received his PhD in Pauling’s laboratory and had published with Pauling. Donohue shared the office with Watson and Crick (and also, in a further twist, with Pauling’s son Peter, who was also a researcher at the Cavendish). Donohue informed Watson that the forms of some of the bases that he was using in his model, which he had gotten from a reference book, were the wrong tautomeric forms and would never be found in nature. That meant, if Donohue was correct, that Watson’s tentative structure was impossible. Since Donohue came with the highest-quality credentials, there was no question of the accuracy of his observation. In addition, Crick noted that Watson’s like-with-like plan went against the shape of the crystal found by Franklin and also did not fit Chargaff’s findings concerning base ratios (see Table 1.1). So that like-with-like model was rejected. Watson then set out to try to find different pairings that would fit. Crick has recalled their agreeing at this point to try to pair the bases following Chargaff’s ratios (that is, pairing A with T and C with G), but Watson has indicated that he does not remember there being that specific a plan (Olby, 1994, pp. 411–412). On a Saturday morning in February 1953, Watson used cardboard models of the bases, in the tautomeric forms suggested by Donohue, to try to determine how they might fit together in the center of the helix. He moved them about on his desktop like pieces in a jigsaw puzzle. Here is his account of what happened (Watson, 1968, pp. 123–125). When I got to our still-empty office … , I quickly cleared away the papers from my desk top [sic] so that I would have a large, flat surface on which to form pairs of bases…. Though I initially went back to my like-like prejudices, I saw all too well that they led nowhere…. [I] began shifting the bases in and out of various other pairing possibilities. Suddenly I became aware that a[n A–T] pair held together by two hydrogen bonds was identical in shape to a [G–C] pair held together by at least two hydrogen bonds. All the hydrogen bonds seemed to form naturally; no fudging was required to make the two types of base pairs identical in shape…. Upon his arrival Francis did not get more than halfway through the door before I let loose that the answer to everything was in our hands. Those pairings (A–T and C–G) turned out to be the same overall size, so the problem of the different-sized rungs for the spiral staircase was eliminated. Thus, the final piece of the model came about by moving cardboard pieces around on a desktop in what one might call trial and error. It is important to note, however, that this final step by trial and error was possible only because the whole structure was almost complete. I have sometimes seen Watson and Crick’s discovery of the double helix labeled as the result of trial and error (e.g., Simonton, 1999), but that overlooks two facts that are important to keep in mind. First, only the very last step—fitting the bases together—occurred through what could be called trial and error. Second, that trial-and-error step was possible because the backbones had already been constructed and the locations of the bases had already been decided upon, and neither of those steps came about through trial and error. In addition, calling this last step trial and error may lead one to think of Watson as simply blindly putting the bases together in every possible combination, but that is not what happened. The various possible configurations that the pairs of bases could form with each other were limited by the chemical structures of the bases: Only certain spatial configurations were likely, based on the necessity that the bases be held together by hydrogen bonds. The situation faced by Watson before this last discovery was analogous to having a jigsaw puzzle completed except for one or two pieces. One might try to fit those last pieces in by trial and error, but that is only possible because almost all the work has already been done. In addition, even if one used trial and error, one would not put the pieces in blindly, with one’s eyes closed; one would place them in the orientations in which they would most likely fit. So Watson was not blindly pushing models of bases around without thought; there was a plan behind even those last steps. Conclusions: Watson and Crick’s Discovery of the Double Helix Watson and Crick’s model of DNA was one of the great discoveries of twentieth-century science and has had great effects on our lives in many ways. However, from the perspective of cognitive processes, nothing extraordinary was involved in bringing it about. This case study demonstrates that one must keep separate the importance of a product, which may be extraordinary, and the thought processes that brought it about, which may be very ordinary. As we have just seen, Watson and Crick took what was known and used that as the basis for building a possible model of DNA. When their first model proved incorrect, they went back to work, and over more than a year acquired new information and expertise that allowed them to ultimately produce the correct structure. The Double Helix: Why Not Franklin, Wilkins, or Pauling? Examination of the story of Watson and Crick’s discovery of the double helix leaves us with a set of related questions: Why didn’t Franklin or Pauling discover the double helix? And why were Watson and Crick the ones who did? The question of what separates the successful from the unsuccessful thinkers has broad implications for understanding the creative process. Franklin’s situation is particularly interesting, since, as we have seen, at the time of her colloquium—that is, in November 1951, more than a year before Watson and Crick finalized their model—Franklin believed that DNA was a helix, and she had discussed relatively specific aspects of the molecule. She was ahead of Watson and Crick at that time, so why did she not carry her thinking through to the double helix? Similarly, we know that Watson and Crick based their work on that of Pauling, which might lead one to the expectation that Pauling had a leg up in theorizing about DNA. Since his general orientation and some of the specifics of his method had served Watson and Crick well, one might expect that Pauling, as the originator of that orientation and those methods, would have had an advantage over them. Here too the answer to the question has potentially broad implications. Why Not Franklin or Wilkins? In her notes for her November 1951 colloquium, Franklin had laid out evidence for a spiral structure for DNA. She cited many sorts of evidence, ranging from a probable lack of stability in a straight untwisted chain to the pattern of spots on her X-ray photos. So she seems to have been far along the road to the double helix. Why did she not take those last few steps, which Watson and Crick did, using some of her data? In April 1952, Franklin took some photos of the A (dry) form of DNA, and one of those photos, of very good quality, led her to the belief that the A form of DNA was not a helix. That photo had an asymmetrical spot in it that Franklin believed could not have occurred if the structure were a helix. Based on an analogy with earlier work on alpha-keratin by William Astbury, a well-known researcher of the previous generation, she concluded that the change between the A and B forms of DNA resulted in basic changes in the structure of the molecule, so the A form might be very different from the B form (Judson, 1979). On July 18, 1952, she wrote as a joke a black-bordered note announcing the death of the DNA helix and stating that Wilkins would speak at the funeral. She spent much of the next year trying to work out the structure of the A form based on that photo, using very complicated and tedious methods of traditional crystallography, and it was not until early in 1953 that she became convinced that that photo had been misleading her. Franklin had also convinced Wilkins that the A form of DNA was not a helix, and he then carried that conclusion one step further and decided that the B form might not be helical either. He then stopped work on DNA for several months in frustration. One reason for Franklin’s concentration on the methods of crystallography rather than Pauling’s model building was that she was concerned that there might be more than one model structure that could fit the data available from studies of DNA. That is, if one built a model that was consistent with those data, how could one be sure that there were not other models, perhaps of an entirely different structure, that would also fit the data? This sense of caution, combined with her inconsistent findings, led her to methods that she felt confident in using. In a professional meeting in late spring of 1952, Franklin told Crick about the evidence that she considered compelling against a helix structure for the A form. Crick told her that a helical molecule might still produce the irregularities in structure that she had found. Thus, he was not sure that her data supported the antihelical conclusion that she was drawing, and therefore he was not convinced that he should change his orientation. In his maintaining his helical orientation in the face of Franklin’s potentially conflicting data, Crick was behaving in a manner similar to the way Pauling had worked in analyzing the alpha-helix (Judson, 1979). In that case too there had been a datum that indicated that the structure was not helical, but Pauling had ignored that information, assuming that it was not critically important to the analysis of the structure. It turned out that he was correct, and that datum was not inconsistent with a helix, which made Crick (as well as Watson, who also discussed that issue with Franklin) a bit skeptical when faced with data that stood out from an otherwise consistent pattern. Franklin, in contrast, was at that point convinced by her results, perhaps because they were her results. In addition, in early 1953, when Franklin had progressed far enough in her crystallographic analysis that she felt comfortable beginning to model, the results initially led her to an incorrect structure, a figure-eight configuration. So Franklin was unsuccessful at least in part because she had data that led her away from the correct structure, data she took very seriously. Wilkins was also influenced by Franklin’s results. It is also notable that just about the time that Watson and Crick completed their model, Franklin had almost completed a paper in which she described a double-helix structure of DNA (Elkin, 2003), so at the end she was very close to the correct structure, and probably would have produced it had not Watson and Crick published their model when they did. Why Not Pauling? Pauling’s lack of success was obviously not owing to reluctance to theorize in a bold manner or disbelief that DNA was helical. Rather, his failure was probably due to a lack of access to the most accurate information concerning the molecule (Judson, 1979; Olby, 1994). For example, he was basing much of his model building on X-rays of DNA taken by Astbury in the 1940s, which were problematic in two ways. First, they did not provide precise information about the structure: They were blurry, so that specific parameters were difficult to determine. Second, and not unrelated, it turned out that Astbury’s X-rays combined the A and B forms of DNA, although at that time it was not known that two forms existed. Thus, part of the reason that Astbury’s X-rays were hard to read was that they combined the two forms. The density measurements that Pauling was using, which led him to conclude that three strands were involved, also were based on Astbury’s work, and we have seen that those measurements were incorrect. If Pauling had had access to Franklin’s B-form photo 51, he might have made more progress. Why Watson and Crick? The development of the double helix is outlined in Figure 1.7, which points out how the various components of the final model came about. In the figure, a number of the critical components of the double helix are listed in the right-hand column, and the development of that component is shown from left to right. Examination of the components indicates that Watson and Crick were successful for several reasons. First, they built their theorizing on the work—both the theory and the empirical findings—of others, and that work was relevant to DNA. In addition, Watson and Crick both brought unique expertise to the enterprise, so that each was able to make a contribution to the final product that perhaps no one else in the world could have made at that time. Examples are Crick’s use of the unit-cell information to deduce the antiparallel chains, based on his particular expertise in that domain, and Watson’s use of the change in length in the A⇔B transformation to deduce that two chains were involved, again based on his particular expertise. Thus, one can conclude that Watson and Crick were the first to specify the structure of DNA because they were uniquely capable of putting together the necessary pieces of information. They were also in an environment that provided several critical corrections to possible mistakes (e.g., Donohue and the tautomeric forms of the bases). Artistic Creativity: Development of Picasso’s Guernica The cognitive processes underlying the discovery of the structure of DNA seem to be comprehensible in a straightforward way. However, one might wonder if one can carry out an analysis of creative thinking in the arts in a comparable manner. Scientific investigation seems to have at its base a logical or analytic method of working, which might make it relatively easy to study in a systematic manner. Actually, careful analysis indicates that the discovery of the double helix of DNA did not proceed in a strictly logical manner. When Watson and Crick began, they assumed that the general model-building method of Pauling might be useful, and they also assumed that Pauling’s discovery of the helical structure of one organic molecule might be relevant to DNA. When one makes the assumption that a piece of information or method of working that has been useful in one situation will be relevant to another situation, that is not a strictly logical process. That is, it does not follow logically that, since alpha-keratin is a helix, DNA is also a helix. This might appear to be a reasonable hypothesis, at least to some people, but it is not logically necessary that it be true. Thus, creative thinking in science is perhaps less strictly structured and logical than we might have thought. Figure 1.7 Timeline: History and development of components of DNA Even if we agree that creative thinking in science is not strictly logical, however, the thought processes underlying artistic creativity still might be too quick, too fragile, perhaps too emotionally laden, too intuitive or illogical, to ever be captured by analysis of the sort just performed. However, I have carried out a case study of Picasso’s creation of his great painting Guernica (Weisberg, 2004; see Figure 1.1B), which indicates that it may be possible to analyze and capture the thought process underlying artistic creativity. Guernica, a landmark of twentieth-century art, was painted in response to the bombing during the Spanish Civil War of the Basque town of Guernica, in northern Spain (Chipp, 1988, Chap. 3). The bombing was carried out on April 26, 1937, by the German air force, allied with Generalissimo Francisco Franco, who became the Spanish dictator after his forces won the Civil War. There seems to have been little strategic value to the town, and when news of the bombing reached the world over the next few days, the destruction of the town and killing of innocent people horrified the world. Early in 1937, Picasso had been asked by the Spanish government (i.e., those in opposition to Franco), who were losing the Civil War, to provide a painting for the government’s pavilion at an international exhibition (a world’s fair) to be held in Paris in June 1937. By the end of April, Picasso had already begun a painting, but one that had nothing whatever to do with the political situation in Spain: It depicted an artist and a model in the artist’s studio, a subject that Picasso used again and again throughout his career (see Figure 1.8). When reports of the bombing reached Paris, Picasso changed his plans and produced a painting that has been universally hailed as one of the great antiwar documents of our time. Picasso’s painting, begun on May 1 in response to news reports of the bombing, was completed in about six weeks, in a burst of creative activity. Guernica quickly became an international symbol of resistance to fascism. In his will, Picasso stipulated that the painting was to be loaned to the Museum of Modern Art in New York City and to be sent to Spain when democracy was restored. That occurred in 1981, after Franco’s death, and the painting then returned to Spain like a hero and was put on display in Madrid. Figure 1.8 Picasso, The Studio (18 April 1937) When one examines Guernica as an antiwar statement, there is something missing: There is no war in the painting (Chipp, 1988). There are no soldiers, no planes dropping bombs, no rifles or cannons, no tanks. There are several animals, four women, a baby, and a broken statue. In the left-hand portion of the painting, a bull stands over a mother whose head is thrown back in an open-mouthed scream; she holds a dead baby, whose head lolls backward. Below them, a broken statue of an ancient warrior (the only human male presence and the only kind of soldier in the painting) holds a broken sword and a flower. Next to the bull, a bird flies up toward a light. In the center of the painting, a horse, stabbed by a lance (again, a hint of warfare), raises its head in a scream of agony. In the upper center, a woman leans out of the window of a burning building, holding a light to illuminate the scene. Beneath the light-bearing woman, another woman with bared breasts hurriedly enters the scene from the right. At the far right, a woman on fire falls from a burning building. Picasso uses those characters to make us feel that something terrible has happened, but he does not present the event directly. Another striking aspect of the painting is that it has no color; it is monochromatic—painted in black, white, and shades of gray. This physical darkness in the painting serves to present a psychological mood of darkness, highlighted by a few bright objects, into which we are drawn. The painting is also massive in size, measuring almost 12 feet by 26 feet, which adds to the strong effect that it has on viewers. When one tries to analyze how new ideas in art come about, one must go “underneath” the paintings that hang in museums or are pictured in art books, because a finished painting tells us little about its birth. However, creative works do not come out of nothing: Especially for large-scale creative works—for example, scientific theories, symphonies, novels, or large paintings—there are several potential sources of information that can help us understand how the work developed. First, creative thinkers, including painters, often carry out preliminary work, thinking about what they might do before they commit to doing anything; artists often carry out this preliminary thinking by producing sketches of various sorts. Those sketches, if available, can tell us where the artist began and how the painting reached final form. Obviously, not everything an artist thinks about is put down in sketches, but, at the very least, sketches can give us an estimate of the relationship between the artist’s early ideas and the final product. Picasso produced many preliminary sketches for all his major works, including Guernica. Even more valuable from our perspective, he dated and numbered the preliminary works for Guernica—a total of 45 works—because he thought that others might be interested in how he progressed. In 1937, Picasso was perhaps the most famous artist in the world, and he was correct concerning others’ interest in his methods. Today, some 70 years later, people still examine the sketches for Guernica (as we are doing here). The Preliminary Sketches When news of the bombing reached Paris, Picasso dropped work on the studio painting and began work on Guernica, producing his first preliminary sketches on May 1; the last sketch is dated June 4. He began work on the painting itself on approximately May 11, and the completed work was put on display early in June. There are two different types of preliminary sketches for Guernica. Seven composition studies present overviews of the whole painting; the remaining sketches, character studies, examine characters individually or in small groups. Samples of preliminary sketches are shown in Figure 1.9. The preliminary sketches provide us the opportunity to get inside Picasso’s creative process. Picasso worked on the sketches for Guernica over a period of a little more than a month; for ease of exposition, that overall period can be summarized into three periods of work: the first two days (May 1–2); an additional six days, commencing about a week later (May 8–13); and a final two weeks of work, which began about a week later (May 20–June 4). As can be seen in Table 1.3A, the first two days resulted in composition studies and studies of the horse, the central character physically and arguably the central character psychologically in the painting. In the second period, the composition studies are fewer, and other characters appear. In the last period, there are no composition studies, and peripheral characters (e.g., the falling person) are seen for the first time. This pattern can be made clearer by combining categories of sketches, as shown in Tables 1.3B and 1.3C. These results indicate that Picasso spent the bulk of his early time working on the overall structure of the painting and on the main character, and then moved on to other aspects of the painting. Thus, analysis of the temporal pattern in the whole set of sketches indicates that Picasso was systematic in working out the structure of Guernica, and he had it more or less completed before he went on to the specifics of the painting. Figure 1.9 Examples of preliminary sketches: A–B C, Composition study; D, Picasso’s falling man E, Picasso’s mother with dead child Table 1.3 Summary of Picasso’s preliminary sketches for Guernica A. All preliminary works tabulated by three periods of work Period Composition Horse Bull Mother and child Woman Hand Falling person Man Total 1 (May 1–2) 6 5 0 0 0 0 0 0 11 2 (May 8–13) 2 4 2 5 1 1 0 0 15 3 (May 20–June 4) 0 2 2 2 8 1 3 1 19 B. Composition sketches versus all others Period Composition All others Total 1 (May 1–2) 6 5 11 2 (May 8–13) 2 13 15 3 (May 20–June 4) 0 19 19 Total 8 37 45 C. Composition sketches + horse + bull versus all others Period Composition + horse + bull All others Total 1 (May 1–2) 11 0 11 2 (May 8–13) 8 7 15 3 (May 20–June 4) 4 15 19 Deciding on an Idea: Analysis of the Composition Studies We have seen that Picasso was systematic in working out the overall structure of the painting; we can now investigate the specifics of how Picasso decided on that structure. By examining the contents of the various composition sketches, we can see if he experimented with several radically different possible structures for the painting, say, or if he had one basic structure in mind when he began to work. As can be seen in Table 1.4, the structure of the painting is apparent in the composition studies produced on the very first day of work. In seven of the eight composition studies, including the very first one, the light-bearing woman is in the center, overlooking the horse. In addition, each of the central characters (horse, bull, light-bearing woman) is present in almost all of the composition sketches, with other characters appearing less frequently. This pattern supports the view that Picasso had at least the skeleton or kernel of Guernica in mind when he began to work: Guernica is the result of Picasso’s working out this kernel idea. We see here further evidence that Picasso’s thought process was relatively constrained from the beginning. Indeed, if the sketches can be taken at face value as the record of Picasso’s thought processes concerning the painting, then from the very beginning he considered only one idea. Antecedents to the Structure of Guernica? Analysis of Guernica as arising from a kernel idea that was available to Picasso from the beginning of his work immediately raises another question: Whence did the kernel idea arise? Chipp (1988), in his extensive discussion of Guernica, was struck by the quick gestation of the painting. When Picasso painted Guernica, he was in his mid-50s and had been an artist for most of that time. Therefore, he had available a history of his own on which (literally and figuratively) to draw; that history seems to have played a significant role in the creation of Guernica, which is closely related to many of Picasso’s works from the 1930s. One striking example of a work that presages Guernica is Minotauromachy, an etching made by Picasso in 1935 (see Figure 1.10). In this composition, a dead woman in a matador’s costume, holding a sword in one hand, is draped over the back of a rearing horse. A minotaur (the half-man half-bull of mythology) raises a hand in front of his eyes to shield them from the light of a candle held by a young woman who is observing the scene. Two other women observe the scene from a window above, where two birds also stand. On the far left, a man is climbing a ladder. Table 1.4 Guernica composition studies: Presence of various characters Figure 1.10 Minotauromachy (Etching; 1935) Table 1.5 summarizes a number of correspondences between Guernica and Minotauromachy, which indicate that Minotauromachy contains the same kernel idea and may have served as a source for Guernica. In order to demonstrate that those correspondences in Table 1.5 reflect more than chance, however, one needs a “control” painting to compare with Guernica. As indicated earlier, just before the bombing of Guernica, Picasso was working on a painting of an artist’s studio. He never progressed beyond sketches for that work; one of those sketches was shown in Figure 1.8. If one compares that work with Guernica, as shown in Table 1.6, one finds very little overlap in subject matter, especially as regards major characters and structure. This comparison among Guernica, Minotauromachy, and the Artist’s Studio is presented graphically in Figure 1.11. Furthermore, the strong correspondence between Minotauromachy and Guernica shown in Table 1.6 and in Figure 1.11 is actually an underestimation of the true correspondence. Minotauromachy, an etching, was printed from a drawing made by Picasso on a printing plate, so the scene Picasso drew on the plate was actually reversed from left to right in comparison with the print shown in Figure 1.10. The “vertical person” was drawn on the right, and the bull was on the far left. The light-bearing female also faces in the same direction as the corresponding character in Guernica. Thus, not only does Minotauromachy contain many characters similar to those in Guernica, but the absolute spatial organization of the two works is also similar. This analysis of Picasso’s creative thought process has provided evidence of planning in Picasso’s thought, as well as structure of several sorts, and continuity with the past. Table 1.5 Corresponding elements in Guernica and Minotauromachy Minotauromachy Guernica Bull (Minotaur) Bull Horse—head raised Horse—head raised (stabbed—dying) Dead person Dead person (broken statue) Sword (broken—in statue’s hand) Sword (in Minotaur’s hand) Flowers (in girl’s hand) Flower (in statue’s hand) Two women above observing Woman on ground holding light Woman above observing + holding light Birds (standing in window above) Bird (flying up toward light) Vertical person (man fleeing) Vertical person (burning woman falling) Sailboat Electric light Mother and child Woman running in The Link between Minotauromachy and Guernica Assume for the sake of discussion that when Picasso began to paint Guernica he had Minotauromachy in mind and used it as a model for the new work. That raises the question of why the bombing of Guernica caused Picasso to think of Minotauromachy. There are several links that can be traced between the bombing, Minotauromachy, and Guernica, which can help us understand why Picasso’s thinking may have taken the direction that it did. First, the bombing took place in Spain, Picasso’s native land, and Minotauromachy is a representation of a bullfight, which obviously has deep connections to Spain and to Picasso’s experiences there, because he painted bullfight scenes from his very earliest years (Chipp, 1988). Guernica also contains the skeleton of a bullfight: a bull and horse, a person with a sword (the statue), and spectators overlooking the scene. In addition, the emotionality of the bombing may have provided a further link to the bullfight, which is an event of great emotional significance for a Spaniard. It may also be of potential importance that when Picasso was growing up in Spain, the bull was not the only victim in a typical bullfight: The horse that carried the picador (the lance-carrier, whose task is to drive the lance into the shoulders of the charging bull) was not protected by padding and was often an innocent victim of the bull’s charge. Based on this reasoning, the horse in the center of Guernica, whose head is raised in a scream of agony, can be seen as a representation of an innocent victim, and one can understand how that symbolization might have arisen in Picasso’s mind. Table 1.6 Comparison of elements in Artist’s Studio and Guernica Artist’s Studio Guernica Bull Horse—head raised (stabbed—dying) Broken statue Sword Flower (in statue’s hand) Woman above observing + holding light Bird flying up toward light Burning woman falling Electric light (above and spotlight below) Electric light above Mother and child Woman running in Reclining model Artist Male spectator Easel Window Window Figure 1.11 Summary of correspondences among Picasso’s Guernica, Minotauromachy, and Artist’s Studio Thus, Guernica and Minotauromachy are linked by a web of interrelationships, and it is not hard to understand why the bombing might have stimulated Picasso to think of Minotauromachy, which then played a role in directing his further thinking. It is also notable that the links among the various pieces that came together to produce Guernica are not particularly exotic. They are simply a complex set of experiences unique to Picasso. Structure in Development of Individual Characters We can also use the sketches of individual characters to examine the question of structure in Picasso’s thought process. First, did he tend to concentrate on only one character at any given time? Second, if we look at all the sketches that contained one particular character—say, the horse—do we find that aspects of that character are randomly varied from one sketch to the next, or is Picasso systematic in his explorations of that character? As with the development of the overall structure of the painting, Picasso was systematic in his development of the individual characters. Attention to One Character at a Time In order to determine whether Picasso tended to concentrate on a given character at a given point in time, we can analyze the sequential pattern over all the sketches (see Table 1.3A). As can be seen, the sketches for the horse were concentrated in the first two periods, the mother and child were most frequent in the second period, and the isolated woman and the falling person were most frequent in the third. Thus, Picasso also was systematic in his working on the individual characters over time. Development of Individual Characters If we consider the development of individual characters over the series of sketches, we can also find evidence for structure in Picasso’s thought. For several of the individual characters, one can focus on elements that Picasso varied separately. As one example, in the sketches of the horse, Picasso varied the position of the head: up versus down. Another example can be seen in the sketches of the woman: whether her eyes are dry or tearing. A third is whether the woman is alone or with another individual (usually the baby). We can examine each of those components in order to uncover structure in Picasso’s thinking as he worked on each character. If we include composition studies, the horse was sketched a total of 19 times. The position of the head of the horse in those sketches is summarized in Table 1.7A, and a clear differentiation is seen: In the earlier sketches, the head is predominantly down, which changes in the later sketches. A similar pattern is seen in the sketches containing women. Tables 1.7B–D summarize all 20 sketches, both composition sketches and character studies, in which there was at least one woman participating in the action (the light-bearing woman was ignored, as were any dead women). Once again there is a pattern in the presence of the various elements of the women over the two periods of work in which women appeared in sketches. In the early sketches, the woman usually is holding a dead person, whereas in the later sketches she is usually alone (Table 1.7B). Similarly, in the early sketches, she is screaming but tearless; in the later sketches, tears are almost always present (Table 1.7C). Finally, Table 1.7D summarizes for all the sketches the relationship between the facial expression of the woman and whether she is presented alone or with a dead person. When she is alone, she is almost always weeping; when she is holding the dead person, she sheds no tears. Thus, analysis of the character sketches supports the conclusions drawn from analysis of the composition studies: Picasso was systematic in his working out of the elements of Guernica. Table 1.7 Summaries of presentation of the horse and of women in the sketches A. Position of head of horse summarized over periods 1 and 2 Period Head up Head down 1 (May 1–2) 8 1 2 (May 8–20) 2 7 B. Types of women in periods 2 versus 3 Period Mother and child Solitary woman 2 (May 8–13) 8 1 3 (May 20–June 3) 2 10 C. Expressions of women in periods 2 versus 3 Period Open-mouthed Weeping 2 (May 8–13) 9 0 3 (May 20–June 3) 2 10 D. Relationship between social environment and emotional expressions of women Type of woman Open-mouthed Weeping Mother and child 8 2 Solitary woman 3 8 Antecedents to Characters in Guernica A further question that can be examined here is whether there were antecedents to specific characters in Guernica, comparable to the antecedents of the overall structure discussed earlier, and one can indeed find what seem to be specific connections to characters in Guernica from works of other artists. Two examples will serve to make the point that Picasso’s thought process was structured in various ways by art with which he was familiar and that was relevant to the theme of Guernica. One particularly distinctive character in the sketches is the falling man in sketch 35 (see Figure 1.9D). Picasso included no men among the actors in Guernica (the only male is the broken statue), so the content of this sketch is intriguing. This distinctive individual, with sharply drawn profile, striking facial expression, facial hair, and idiosyncratic placement of eyes, as well as the falling posture with outstretched arms, bears a striking resemblance to the man shown in Figure 1.12A. The latter drawing is from an etching in a series by Francisco de Goya (1746–1828) called Disasters of War, which was created more than 100 years before Guernica. Picasso had great respect for and knowledge of Goya’s works (Chipp, 1988), and it would not be surprising if the events that stimulated Picasso’s painting of Guernica also resulted in his recollection and use of Goya’s work as the basis for his own, especially given the commonality of theme. Picasso changed the man into a woman in the painting, but the falling woman in Guernica bears residue of the Goya etching from which she began: Her profile is similar to that of Goya’s man, and her outstretched hands with exaggeratedly splayed fingers echo those of the Goya figure. Figure 1.12 Details from Goya’s Disasters of War: A, Falling man (No. 41); B, Woman (No. 13) A second example of a correspondence between one of Picasso’s characters and another work from Goya’s Disasters of War is shown in Figure 1.12B. Picasso’s sketch 14 (Figure 19.E) contains a mother and child, and the woman is distinctive in her sharply profiled head thrown back; in her pose, with her outstretched left leg producing a distinctive overall triangular shape; and in the arrangement of her skirt, which is folded between her legs. The woman in Goya’s etching is similar in facial profile and expression, and in her posture, with an outstretched left leg producing an overall triangular shape, and her skirt is folded between her legs. Structure in Picasso’s Creative Process: Conclusions This analysis has provided evidence for what one could call layers of antecedents to Guernica. The overall structure, based on the kernel idea seen in Minotauromachy, is one level of structure. Within that structure or framework, specific characters are based on other antecedents, meaning that one can trace antecedents nested within antecedents in Guernica. If the present analysis is accepted as valid, it means that it has been possible to trace the origins of some microscopic aspects of Guernica: for example, the facial expressions and postures of some characters, as well as the appearances of the characters’ hands. Thus, the present analysis supports the proposal that creative works may be closely linked to previous works, although how often this occurs and how close the links are remain to be answered through the analysis of other works, both in painting and in other domains (for further discussion, see Weisberg, 1993). Structure in Creative Thinking: Conclusions from the Case Studies The most striking point to be drawn from these two case studies is that it seems possible to analyze in a straightforward way the creative thought process—even the creative thought process at the highest levels. In addition, the creative process seems to be highly structured and not very different from the thought processes involved in more mundane activities. There seems to be a large gap between the importance of the products and the ordinariness of the thought processes that brought them about. In both cases, it was not necessary to introduce any thought processes that seem to go beyond ordinary thinking. For example, in developing the double helix, Watson and Crick used information that came from domains closely related to the one in which they were working. No wide-ranging creative leaps occurred in which ideas from totally unrelated areas were brought together in some sort of magical synthesis. Similarly, Picasso did not leap far afield in creating Guernica: He built on his own work from that period of time and incorporated related work by others. The products may in both cases have been extraordinary, but the processes by which they were brought about were not. Concluding that ordinary thought processes underlie creative thinking does not mean that just anyone could have produced either the double helix or Guernica (see also Klahr & Simon, 1999). First, in order to have produced the double helix, one first had to acquire the expertise of Watson and Crick, who, it must be remembered, were two people with two complementary sets of expertise. One had to have their intellectual capacity and to have put in years of hard study. Second, one had to be directed toward constructing models à la Pauling, which may involve ways of thinking and visualization skills that at the very least had to be developed. Concerning Guernica, Picasso had been an artist for some 40 years when he painted it, so one would have to have acquired his unique lifetime’s worth of expertise; in addition to his skill as a painter, he had a voluminous knowledge about art. Furthermore, in Picasso’s work one sees references to classical mythology (the Minotaur) and other bodies of knowledge, meaning that one would have to go beyond painting itself before one could produce works with the broad range of connections that people find in the works of Picasso. Thus, the premise that creative work may not go beyond ordinary thinking processes does not mean that creative work is effortless and that anyone could do it. These case studies have served to present the orientation that I will take toward the analysis of creativity: I will begin with analysis of the fine structure of the creative thought processes and work outward from there. The case studies just presented also provide the beginnings of a database for us to use in evaluating theoretical claims made about the creative process and about creative people. Antecedents for New Ideas and the Question of “Real” Creativity We have seen in these two case studies several examples of a phenomenon that will be seen numerous times in the cases of creative thinking that are discussed in this book: Creative ideas, even those that are radically new, are firmly planted on ideas that came before. There are always antecedents to any creative idea. The reason that it sometimes looks like an idea comes out of nothing is because we observers are ignorant of the knowledge base of the individual producing the new idea. If we knew what he or she knew, then we could see where the new idea came from. In response to this phenomenon (which, it must be emphasized, we will see again and again, in every case in every domain we examine), some people dismiss the specific examples as not being creative, and assume that we must look for “real” creativity elsewhere. That is, they conclude that Watson and Crick were not creative, or that Guernica was not a creative work. Other people conclude that the demonstration of antecedents for ideas means that there is no such thing as creativity: Everyone just rips off ideas from everyone else, and nothing is new. Both those conclusions, however, are incorrect. First of all, both Guernica and the double helix are creative products: They were novel. Simply because a new work is related to—or based on, or developed out of—an earlier work does not mean that the new work is not novel. The difficulty comes if one assumes that a creative work must be completely novel, which may be an impossible criterion. So, if one lowers one’s (perhaps unreasonable) expectations a bit, and looks for something less than complete novelty, then one sees that we are dealing with creative products here. The second aspect of this problem is the unspoken assumption that there will be cases where complete novelty is to be found and that these are what we should be looking at. However, going beyond the two case studies just reviewed, I believe that all creative products are less than completely novel. Although I cannot prove it, it is my belief that one will never find a creative product for which there are no antecedents. I will try to get as close as I can at this time to proving that there are no creative products without antecedents, by examining a wide range of products from a wide range of domains, and showing the antecedents for each of them. This may help make people more willing to entertain the idea that there may be no creative products without antecedents. Are Creative Ideas Just Recycled Old Ideas? The two case studies discussed in this chapter provide support for what could be called the foundation view of the relationship between expertise and creativity: Experience provides the foundation on which the creative process produces innovations. It is sometimes concluded that the foundation view trivializes creative thinking, because accepting the foundation view means that we must also assume that there really is no such thing as creative thinking because no creative products are novel. In response to this objection, I must emphasize that the foundation view does not trivialize creative thinking by declaring that all new products are simply recycled old ideas. Creative thinkers go beyond the past to produce genuinely novel ideas and objects. For example, as we have seen, the double helix model of DNA, although based on Pauling’s alpha-helix, differed from that structure in several critical ways. Likewise, although Guernica was related to and perhaps built on aspects of Picasso’s work and that of others, it was in many ways a novel work. Additional evidence that the creative process requires more than simple recycling of old ideas can be gleaned from another stream of the DNA story (Judson, 1979; Olby, 1994; Watson, 1968), which was briefly mentioned earlier. As we saw, in an ironic twist, at the time Watson and Crick were developing their model, which was based on Pauling’s ideas, Pauling himself formulated an incorrect model, a triple helix, similar to the one initially developed by Watson and Crick. Pauling was not able to correct his model before Watson and Crick published their work on the double helix. If creativity simply involved recycling old ideas, then it is hard to understand how Pauling could have failed, since he was the one who developed the helical notion in the first place (although in another context). Pauling’s failure leads to the conclusion that much more must have been involved in the creation of the double-helix model of DNA than simply taking old ideas and using them again. Further evidence for the active thinking involved in creativity can also be seen in the fact that Watson and Crick also first produced a triple helix, which (as we know) they soon realized was incorrect, for several reasons. One problem with the triple helix was that it was difficult to put the atoms in the model in such a way that they fit together as they were supposed to. After further work, they then rejected it in favor of the double helix. Thus, there was much thinking involved in both creating and rejecting the triple helix; it was not simply an old idea recycled. Revisiting the Question of Artistic Creativity versus Scientific Discovery We can now turn again to the question of differences in the creative process in science versus art. At the beginning of the chapter, it was suggested that one might look upon artistic creativity as an inherently subjective process, since the artist brings objects into existence as he or she carries out the creative process. Once again: no Picasso, no Guernica. On the other hand, scientific creativity is an objective process that deals with objects that exist “out there,” independent of the scientist. The scientist does not bring objects into being through the creative process; he or she discovers objects that exist independent of the scientist and of the creative process. Again: With no Watson and Crick, there would still be DNA. This subjective/objective distinction is not as clear as it seems, however. Crick, for example, has stated that he believes that if he and Watson had not made their discovery others would have done it, probably within a few months of when they did (Olby, 1994). Crick has also asserted his belief that if he and Watson had not published their discovery, which presented the entire structure as a whole, the structure would have been revealed in bits and pieces. In Crick’s view, such a presentation of the structure would have had a less dramatic effect on other scientists, and the appearance of the structure would then have been less influential and important than it was. This is an interesting point, because, if correct, it indicates that the same “facts” can be presented in different ways, which can change the influence of a discovery. This leads one away from the simple notion that discovery deals simply with objective facts waiting for us to find them and present them to others. Carrying this line of reasoning further, if we look more carefully at the way we use the terms discover and create, and at some of the conclusions arising from the case studies just examined, we may conclude that the creative processes in the arts and sciences may be more similar than different. Multiple Forms of Discovery Perhaps the simplest and most direct use of the term discovery occurs when we talk about a person discovering a dollar on the street. It is not unreasonable to say that anyone—even a child—could have made that discovery, and that the dollar was “waiting” to be discovered. However, we also talk about Columbus’s discovering America, and that discovery seems different from discovering a dollar, since not just anyone could have made that discovery. America might have been waiting to be discovered, but it was not visible to Columbus or anyone else in Europe when Columbus made his presentation to King Ferdinand and Queen Isabella of Spain. While any person high up in the rigging of one of Columbus’s ships—perhaps even a child—would have sighted approaching land, in order to get those ships and that person into that location, Columbus had to bring special elements to the enterprise. Most important for the present discussion, the discovery of America depended on Columbus’s having a theory about the shape of the earth. If he had not had that theory, there would have been no reason for his voyage, and America would not have been discovered then. So, although America was waiting to be discovered, in order for that act to occur it had to be embedded in a complex set of beliefs that served to initiate a complex sequence of actions that culminated in the discovery. Discoveries, then, can be of several kinds; and Columbus’s discovery, although designated by the same word as the child’s discovery of a dollar on the street, turns out to have different antecedents and to be part of a complex set of circumstances. Now we can raise the question of whether the discovery of DNA more was like the child’s discovery of the dollar or Columbus’s discovery of America, and it seems obvious that it was more like the latter. DNA is not visible to humans, which means that when we talk about seeing DNA itself or seeing that it is helical—from an X-ray, for example—we are not talking about ordinary seeing. Rather, in order to discover the characteristics of DNA, an act comparable to discovering America, one must have a theory as to how the available evidence is to be interpreted. That is why the Cochran-Crick theory of interpretation of X-rays was important in the discovery process: It allowed investigators to “see” things they would not have seen (Olby, 1994). Similarly, we saw that Watson and Crick’s adoption of Pauling’s helical idea made it possible for them to interpret much information and to ignore other information that became irrelevant. Watson’s learning to interpret X-ray photographs also made it possible for him to see things in Franklin’s B-form X-ray that he would not have seen several months earlier. Crick’s knowledge concerning X-ray crystallography enabled him to interpret the unit cell from Franklin’s analysis of the B form of DNA, and so on. So the process of discovering the double helix required that the investigators bring much to the situation in the way of knowledge and beliefs. The situation seems more subjective than we might have thought, as we have moved away from the notion of discovery involving an objective investigator using objective facts to draw an inevitable conclusion. Let us now look in the same way at the creation side of the creation-discovery distinction; we will see that the dichotomy is not very clear from that side either. Multiple Forms of Creativity We use the term create in several ways, and although they are similar, they should be distinguished. On the one hand, we talk about God creating the heavens and the earth, which is the paradigmatic case of creation. Divine creation is extraordinary in one critical way: God created the heavens and the earth out of nothing—ex nihilo. That would seem to be the epitome of subjectivity in creation. However, human creativity is different from God’s creation of the heavens and the earth. Human creation does not occur in a vacuum; human works are not created ex nihilo (and perhaps cannot be). We have seen that Picasso was influenced by and built his work on that of others, which means that his creative work was less subjective than one might have thought. That is, his work was embedded in the history of art, with influences stretching back thousands of years (e.g., the myth of the Minotaur). Thus, we do not see the solitary creator producing something out of himself independent of what others have done, as the notion of artistic creation might lead one to believe. Just as there is a subjective aspect to scientific creativity, there is an objective aspect to creativity in the arts. Figure 1.13 Artistic/scientific creativity continuum Artistic Creativity and Scientific Creativity: A Continuum It follows from this discussion that artistic creativity and scientific discovery are not two separate categories of activities. Rather, they overlap in various ways. One can represent this relationship as in Figure 1.13, which puts artistic and scientific creativity on a continuum, with some degree of overlap, in the components just discussed. At one end of this continuum is God’s creation, and on the other is the child’s discovery of the dollar. Artistic and scientific creativity are seen as occupying more central positions in this continuum, with overlap arising from the role of antecedents and the creator’s knowledge in both. From this perspective, it is not absurd to say that Watson and Crick created the double helix, although it seems less acceptable to say that Picasso discovered Guernica. (It should be noted, however, that art critics [see, e.g., discussion in Rubin, 1989] have used the term discovery to discuss some of Picasso’s accomplishments.) This discussion provides support for the belief that one can understand creative thinking in general with one set of theoretical constructs. Beyond Case Studies: Outline of the Book We now have a basic background for the study of creativity; examination of two case studies of creativity at its highest has indicated that those examples are amenable to analysis, and that the processes through which innovations come about are comprehensible. This background becomes useful immediately, as we now turn to a broader consideration of psychological studies of creativity. In Chapter 2, I will introduce the study of creativity, examining questions of definition of the relevant concepts and examining the entire range of methods used by psychologists to examine its multiple facets. I have already introduced one theoretical perspective for studying creativity—the “ordinary-thinking view” —which proposes that creative thinking uses the same processes as our ordinary thinking, and that no new thought processes need be postulated in order to understand how creative advances are brought about. That theoretical view is only one of several different perspectives that psychologists have proposed in order to understand creativity. Chapter 2 will present an introduction to the range of theories of creativity, and subsequent chapters will present specifics of the various theoretical perspectives. Chapter 3 begins the detailed presentation of the “ordinary thinking” or cognitive perspective, focusing on the relation between problem solving and creative thinking. Chapter 4 examines another component of the ordinary-thinking view, the idea that experience serves as the crucial component on which problem solving and, by extension, creative thinking operate. In Chapter 5, the now fleshed-out ordinary-thinking view will serve as the basis for the analysis of several additional case studies of creative thinking at the highest levels. Those case studies will be drawn from a wider range of domains than those discussed in this chapter, and several of them will represent radical breakthroughs. As such, those case studies will provide a strong test of the notion that ordinary thinking underlies all instances of creativity. We will then turn to an examination of various aspects of the dominant view in psychology of creativity: the idea that creativity is the result of extraordinary thought processes carried out by extraordinary people. Chapter 6 examines the notion of insight in problem solving, which is the idea that creative ideas sometimes come about through “leaps,” which go beyond what one knows. This view, which has a long history in psychology, challenges the premises that creativity depends on experience and on ordinary problem solving, and so it deserves detailed discussion. Chapters 7 and 8 examine two other examples of the premise that creativity depends on extraordinary thought processes, which is contrary to the view underlying the presentation in this book. Chapter 7 examines the question of genius and madness—that is, the hypothesis that creativity is related to, and may actually be fostered by, psychopathology. Chapter 8 considers the possible role of the unconscious in creativity. In Chapters 9–11, we move considerably beyond the consideration of the creative process to examine the possible broader psychological aspects of creativity. Chapters 9 and 10 present the psychometric perspective on creativity, which emphasizes testing for creative potential and the possible role of personality factors in creativity. Chapter 11 reviews several important recent confluence theories of creativity that have attempted to bring together many different components—cognitive, personality, and environmental—in understanding creativity. Chapter 12, the final chapter, presents an overview of the ground we have covered.

CHAPTER 2The Study of Creativity We now have considered two examples of creative thinking at the highest level, and it seems from those case studies that one can analyze in a reasonable way the thought processes underlying seminal creative advances in both science and art. We can now put those examples into the broader context of the study of creativity. The present chapter provides a general introduction to the psychological study of creativity, to set the stage for the discussion in the rest of the book. Outline of the Chapter We must first define the phenomenon in question. What do we mean when we say that some product is creative, or when we say of someone that he or she is a creative person? The two case studies in Chapter 1 were without doubt creative achievements, which allowed us to put aside the question of how one decides whether a product is creative. In order to go further in our discussion, however, we have to consider how to explicitly define the phenomena of interest. Once we have arrived at a definition, we will examine the kinds of methods used by researchers to study creativity. Those methods range widely, from the case studies with which we are already familiar, to self-reports of individuals of renown concerning how they achieved their creative breakthroughs, to laboratory studies of undergraduates solving problems like those presented in Figure 1.1C. I will then briefly examine several examples of the types of theories developed to explain creativity. Here too we will see a wide range, this time of theories, as psychologists have attempted to understand the varied phenomena encompassed by the study of creativity. Creative Product, Creative Process, and Creative Person: Questions of Definition The critical element in calling some product creative is that it be new; if someone produces something that he or she has produced before, then that product is not creative. There are two aspects of novelty that must be separated: novelty for the person versus novelty for the world (Weisberg, 1986, 1993). It happens relatively frequently that a student will report to me that she (or he) was reading an assignment and got a great idea for an experimental study, only to turn the page and find, much to her disappointment, that the study had already been carried out. Was the student creative in thinking of her study, even though it had been thought of earlier by someone else? Yes, because the idea was novel for her. So long as she was not aware of what had been done before, then she was creative. It follows from this that the creative process or creative thinking comprises the thought processes that bring about products that are novel for an individual. A creative product—or an innovation—is one that is novel for an individual, and a creative person is one who produces such products. Creativity is made up of the factors that enable a person to produce creative products, as in “He has a lot of creativity”; it is also the activity involved in such production, as in “Creativity is hard work.” Creativity encompasses at least the creative process, but it may also depend on personality characteristics and on motivation (Amabile, 1996; Simonton, 1999; Sternberg & Lubart, 1995). Creative Accidents? An important point to be considered in defining creativity concerns the possibility of accidental creativity. Let us say that I am a painter, and one day I accidentally spill paint on a canvas, which leaves a stain on my partially finished work, making it unusable. Let us further assume that I am visited by the director of a museum, who sees my stained canvas, loves it, and purchases it for display in the museum. The painting is then discussed in art books, and other artists use my spilled work as the basis for innovations of their own. My piece of junk has thus become part of the world of art. Was I creative in producing that painting? No; since the novel element, the stain produced by the spilled paint, was an accident, then I get no credit. I am creative only when my novel product is produced intentionally. Similarly, if you ask a person who suffers from schizophrenia to try to solve a problem during a schizophrenic episode, he or she may produce a novel response, such as a stream of free associations that you will not be able to understand and that are not relevant to the problem. Is that a creative product? Again, the answer is no; someone in the grip of a schizophrenic episode would not be able to direct his or her thought processes toward the problem at hand, so no creativity would be possible. There is an interesting twist concerning schizophrenia as it relates to creativity. Sass (2000–2001) discusses cases in which individuals suffering from schizophrenia design complicated “machines” in order to carry out some task in their delusional worlds—say, to protect themselves from having their thoughts read by the CIA—although those machines obviously cannot succeed in carrying out the task. Is a schizophrenic individual exhibiting creativity in designing and building such a machine? In my view, the answer is yes. If the machine is a novel product, and if, once one accepts the premises that the schizophrenic individual is working under, the construction of the machine follows intentionally from those premises, the person is creative. There is, in my view, no contradiction in saying that a person is psychotic and creative. So long as he or she can carry out some thought process in pursuit of a goal, and if that thought process results in a novel product, then the person is creative. The Question of Value in the Definition of Creativity Almost all researchers who study creativity use a definition a bit different from the one I have just given: In addition to a product’s being novel, it is also proposed that the value of a product is important in determining whether it is creative (for examples, see chapters in Shavinina, 2003, and Sternberg, 1999a). In this view, to which I used to subscribe (e.g., Weisberg, 1986), in order for an invention to be called creative, it must carry out the task for which it was designed; a creative scientific theory must help us understand the domain in question; a creative work of art must be appreciated by some audience beyond the artist; and a creative solution to a problem must actually solve the problem. Csikszentmihalyi (e.g., 1988, 1996, 1999) has presented a detailed analysis of how novelty and value are both relevant to creativity. Since that view is now the dominant one in the field, and since my present view is different, it is worth discussing in some detail the issue of value in creativity. Csikszentmihalyi (1988) proposed that three components, shown in Figure 2.1, play a role in making any product creative. First, we have an individual, who is working in some domain, say a researcher in molecular biology in a university, a fashion designer working in a design house, or a painter in a studio. The domain of molecular biology is all the accumulated knowledge concerning that subject matter, everything available in books and journals, that fills our libraries. Similarly, the domain of fashion design is all the accumulated knowledge about fashion, and here too we have libraries of accumulated information. The domain of painting consists of those works of past artists that are displayed in museums and galleries and are discussed in books on art and among artists. Figure 2.1 Csikszentmihalyi’s systems view of creativity Note: This map shows the interrelations of the three systems that jointly determine the occurrence of a creative idea, object, or action. The individual takes some information provided by the culture and transforms it, and if the change is deemed valuable by society, it will be included in the domain, thus providing a new starting point for the next generation of persons. The actions of all three systems are necessary for creativity to occur. Source: Adapted from Csikszentmihalyi (1988). At some point, that individual makes a novel contribution to the domain: The molecular biologist makes some novel research discovery or develops a new theoretical perspective; the designer comes up with new ideas for next season’s collection; the painter creates a new painting or a new style of painting. When an individual produces something new, he or she has to make it public. In other words, the individual must present the discovery to the field, that is, those individuals who work in the area. The field in molecular biology is made up of the scientists who carry out research in that discipline; in fashion, it is the individuals in the fashion industry; in art, it is art critics, museum directors, gallery owners, artists, and the members of the public who attend museum and gallery shows and who purchase art. In science, new ideas are usually presented through publication in a professional journal in the field. The best-known members of the field serve as journal editors and reviewers, and they decide whether the new finding is important enough—valuable enough—to merit being published. Those individuals thus serve as gatekeepers to the domain. Similarly, in fashion, the new ideas are presented to the field in a fashion show; fashion critics, clothes buyers for stores, and individuals who buy high-fashion clothing decide whether the new clothes are worthy of wearing. The world of art also has gatekeepers. Paintings are displayed in museums and art galleries; museum curators and gallery owners decide whether to put on display new works from artists. In each of those areas, if the members of the field decide that an innovation is not of interest, then it may never see the light of day, and it will have no effect on the domain. The molecular biologist’s discovery will not be published, and consequently most people in the field will never learn of it; the designer’s clothes will never be worn; the artist’s paintings will never be seen. On the other hand, if a new scientific finding is deemed worthy of publication, it then becomes part of the domain, where it is studied by other scientists, and the cycle continues. If the designer’s new clothes are deemed worthy, they will be seen in fashion magazines, in stores, and eventually on the streets, and other designers may be stimulated by them. Similarly for the artist, the painting will be seen in a museum or a gallery; it may be discussed in a book or journal article; and it may influence the thinking and work of other painters. In Csikszentmihalyi’s (1988) view, the term creative should only be used to describe a product that goes through the whole cycle presented in Figure 2.1: A novel product becomes creative only after it has become part of the domain—that is, only after it has been positively valued by the field. If a product is rejected by the field, in this view, that product is not creative, whether or not it is novel. Thus, other scientists play a role in making a scientific discovery creative; fashion critics and clothing buyers for exclusive stores play a role in making a designer creative; and art critics and other artists play a role in making a painter creative. Creativity Is Novelty, Irrespective of Value As I said earlier, most researchers agree with Csikszentmihalyi that the value of a product plays a role in whether it should be called creative; I, however, do not (Weisberg, 1993, Chap. 8). Csikszentmihalyi has produced a valuable analysis of the way in which multiple factors come together in determining whether an idea will be accepted by the intellectual community. However, I would make one small change in the terminology used to label the cycle in Figure 2.1 and its components: I would say that the cycle determines whether an innovation comes to be valued, and I would restrict the use of the term creativity to the individual’s production of the innovation in the first place (see Figure 2.2). I make this distinction because I believe that one should separate the creativity of a product from its value, and therefore I will not use value as a criterion for calling something creative. The reasons I am reluctant to include value in the definition of creativity are twofold. One difficulty that I see in using the value of a product as part of a definition of creativity is that value can change over time. Sometimes a product is not valued when it is produced but comes to be valued by later generations. Examples of this are easy to find; the Impressionist painters, whose work is now among the most beloved and historically important in all of painting, were ridiculed by art critics when their works were first put on exhibit in Paris more than 100 years ago. On the other hand, sometimes a product is valued greatly when it is produced but is then put aside by future generations as being of no lasting value. Many of the artists whose works were valued most highly by the art community when the Impressionists first came on the scene are no longer remembered in any way; art history, curators of museums, and the art-viewing public have forgotten them. If we use value as part of the definition of creativity, we would have to say that a person could be creative at one time and then become not creative later, and vice versa (Csikszentmihalyi, 1988, 1999). A person could become creative after death, for example, if that person’s field comes to appreciate his or her works then. Even during one’s lifetime, one’s creativity could change, because of changes in how one’s work is viewed by the field. Figure 2.2 Modified Csikszentmihalyi cycle Csikszentmihalyi does not view as problematic the possibility of a change in a person’s creativity, but I think that allowing the creativity of a product (and of the person who produced it) to vary when its value changes creates difficulties for understanding creative thinking. One critical problem is that we would never be able to say for certain who or what is creative. Since it is possible that the value of any product can change over time, it means that anything can become creative or become not creative, and can keep switching back and forth. There is no guarantee that any products that we value today—the music of Mozart, the paintings of Rembrandt, the scientific theories of Einstein and Darwin—will be valued by future generations. Indeed, I know people who think that Mozart’s music has no value. This means that it is possible—perhaps not likely but still possible—that future generations may universally consider those works to be bereft of creativity. In Csikszentmihalyi’s view, Mozart is not creative to those people I know who are bored to tears by his music. This seems to me to miss the essence of what I think of as creative: the production of works that were novel for their creator at the time they were produced. I think it is more useful to keep the creativity of a product separate from its value, and to say that a product is creative so long as it is novel. If and when the field’s opinion of a work changes, then we can say that the work (or the person who produced it) has become more or less valued or appreciated, but I do not think that we should say that the work (or person) has become more or less creative. Accordingly, I will assume that you can be creative even if you produce a new scientific theory that is totally wrong (such as the triple helix of Watson and Crick and that of Pauling; see Chapter 1), or new music that no one ever likes, or a new airplane that never gets off the ground, or new clothes that no one ever wears. All that matters is that the product be novel for you and produced intentionally. A related problem that I see with the notion that value should play an important part in defining creativity is that acknowledging that people can become creative after they die, or can become noncreative, means that we can never carry out the psychological study of creativity. As an example, let us say that we want to study the thought processes or personality factors involved in creativity in painting. Whom are we to study? As a start, we could study living painters whose work has been deemed important by the field. Let us say we do that, and we draw some conclusions as to what factors are important in making someone a creative painter. However, let us assume that in 10 years those painters are no longer in favor; that is, their work is no longer valued, which means that, if we use value in our definition, they are no longer creative. Therefore, our conclusions from 10 years ago are no longer relevant to creativity. We have to study the characteristics of the painters who are now designated as creative, which means that we may always be going back to the beginning as far as the scientific understanding of artistic creativity is concerned. One might try to get around this problem by studying people long dead who have been designated as creative, assuming, for the sake of discussion, that we had information about those people’s lives, working methods, and personalities. However, even if a person has been dead for 100 years, say, or 200, or 500, and his or her work has been positively valued for all that time, there is no guarantee that the next generation will value those works. Even the long-dead could lose their status. I know that the opposite is true: I constantly see articles praising long-dead painters and composers who have been “discovered” by critics and orchestras, respectively. Thus, if we allow value to be part of our definition, there seems to be no way to achieve a solid foundation for our studies of creativity. In contrast, if we use my definition of intentionally produced novelty to designate a creative product, then the only issues we have to worry about are whether a person has produced a novel outcome and whether it was intentional. Both of those criteria are relatively easy to ascertain. It should also be noted that the question of assessing the novelty of a product and the intention of the individual must also be addressed if one defines creativity as the production of novel products that are of value. Thus, my definition is simpler to apply to a product, since there is no need to worry about value. Furthermore, by my definition, if you are designated as creative—and if the novelty and intentionality of your product have been measured accurately—your status as creative stays with you forever. Similarly, if you die noncreative, then that cannot change, unless, of course, some previously unknown novel work comes to light after your death. Therefore, we are free to study the characteristics of creative versus noncreative individuals, secure in the knowledge that those people we designate as creative will remain creative, barring unusual situations in which they are found to have copied someone else’s work and passed it off as their own or to have produced some novel product by accident and passed it off as something produced intentionally. It is interesting to examine the relationship between the two definitions of creativity being discussed: intentional novelty plus value versus intentional novelty. Consideration of the two sets of criteria indicates that they are related; one is contained within the other. In other words, anything that is designated as creative according to the novelty-plus-value definition currently used by most researchers will also be designated as creative using my definition, assuming that the novelty is brought about intentionally. If something is novel and valued, then it is, ipso facto, novel, and therefore we all would call it creative. Similarly, if something is not novel, then whether or not it is valued we would all agree that it is not creative. Thus, the only situation in which the two definitions would disagree is where some intentionally produced novel product cannot be said to have value (e.g., those triple helices); I would call that product creative, while most other researchers would not. Consider again briefly what sorts of products those valueless things might be. One might produce an invention that fails miserably, a piece of music that all people find unlistenable, a painting that no one cares to look at, or a scientific theory that turns out to make no correct predictions and does not serve to make sense out of the area in which it was developed. As a concrete example, one reviewer of this book said that my definition would not work, and as evidence he raised the idea of using nails for money. A 1 cm nail would be worth $1, and longer nails would be worth more. He said that there was no way that such an invention could be called creative, because it would be obvious to all that the idea was stupid, and he then dismissed the whole rationale of defining creativity based only on novelty. However, I think the issue is more complicated than the reviewer did. First of all, it is not clear to me that anyone in his or her right mind would seriously propose such an invention, so the example raised by the reviewer might only be a red herring. If someone in seriousness did propose using nails as money, surely that person, if asked, would have interesting reasons for such a proposal, which means that dismissing it as completely stupid is premature. I believe that when people spend time and effort on some product, then the outcome, if it is novel, should be considered creative. If we do not like the product, or if we find that it is useless, then we can say so in response to it, but those judgments are different from a judgment about whether the product is creative. One other reason for leaving value out of our definition, related to the point just made, is that I see no reason to believe that the psychological processes involved in producing a positively evaluated innovation are different in any way from those underlying a negatively evaluated one. We have already seen a number of examples of this in the case study of DNA in Chapter 1. First of all, compare Watson and Crick’s development of the double helix with that of their earlier triple-helix model. The latter was of no value and was rejected almost immediately, while the former has changed the world. However, one sees processes at work in development of those two models that are the same; the information that the processes were working on had changed, so that the outcomes were different. The issue of value seems to be irrelevant to the processes involved. Similarly, we compared Watson and Crick’s creative process with that of Franklin, who did not produce the ultimate outcome of value. The crucial element affecting her lack of success seems to have been that she was dealing with different assumptions and, most importantly, with a critical piece of information (her DNA X-ray photo that appeared asymmetrical; see Chapter 1) that led her in an unsuccessful direction. There does not seem to be any basic difference in the processes involved in production of more- versus less-valued products. As another example, a painter might produce two paintings, one of which becomes a seminal work—that is, part of the domain—and the other of which does not. This is the case with Picasso’s Guernica and Minotauromachy, the former of which is much more important (i.e., valued) in the history of art than is the latter. Similarly, an inventor might produce one machine that carries out some task well, and another that fails miserably at the same task (we will see examples of this in later chapters). Why should we assume a priori that different cognitive processes are involved in the production of the valued versus nonvalued innovations? It seems more reasonable in the case of valued versus nonvalued works of art, for example, to assume that the evaluations, both positive and negative, are the result of what others get out of the paintings rather than what the artist put into them. It is also of interest to note again that Picasso may have based Guernica on Minotauromachy, which again indicates that it is difficult to separate more- from less-valued innovations. Similarly, the inventor who produced a machine that failed miserably obviously missed some thing or things, but that does not mean that there was a basic difference in the processes involved in producing the successful versus unsuccessful one. In addition, in many cases of creative innovation at the highest level, one finds that an early product was unacceptable in some way, and the person modified it until it was acceptable. We already saw this in the case of DNA in Chapter 1: Watson and Crick first produced the triple helix, and only after much work was the double helix produced. In such circumstances, the negatively valued and positively valued products are part of the same process, and it seems difficult if not impossible to make a distinction between them. Thus, it seems that for a number of reasons things are much simpler for the student of creativity if we leave value out of our definition, so that is what I will do here. External Verification of Creativity Even if we agree to use goal-directed novelty as the definition of a creative product, that does not mean that social factors are irrelevant in determining whether some work is creative. A student once told me about an artist friend who at the end of each day destroyed that day’s work. She raised the question of whether he was creative. The answer is not a simple yes or no—the way the artist has structured the situation, we cannot say whether he is creative. Let us say that the artist tells his friend that he has had a great day and produced work that went far beyond anything he had done before. Since he has destroyed that work, however, we cannot tell whether that statement is true. In order to conclude that the work is indeed creative, we have to have that work available, so that we can tell that it is in fact new. Perhaps the artist is repeating himself but does not realize it—surely his memory is fallible, as all our memories are, and he may have forgotten that some time ago he produced a work just like the work he produced today, the one that he thought was so new. Thus, we need to have a product that other people can look at before we can call something creative. However, to reiterate, this need to have others look at a product is not, as Csikszentmihalyi believes (1988, 1999), because others’ judgment of the value of the work is important in determining whether it is creative. Rather, we need external observers because at this time only external observers can verify that the work is novel. If the artist in question had some database available—say, a computerized pictorial record of all his work—in a form that was easily searchable, so that he or his friend could determine accurately whether some just-produced work was novel, then he would not need the rest of us to determine if he had been creative that day. However, until such a database is available, any individual who destroys his or her work before anyone else sees it simply makes it impossible for us to determine whether anything creative has been produced. Similarly, if we had a complete videotape of the artist working, including any spontaneous comments he or she might make in the course of producing some work, then we could determine reasonably well whether some product came about by accident or intentionally. Again, until such records are available, we will have to ask the artist. A Working Definition We now have a working definition of creative thinking that we can use as the basis for the discussion in the rest of the book: Creative thinking occurs when a person intentionally produces a novel product while working on some task. Sometimes those intentional novel products are valued highly by society, and sometimes they are not, but all of them are creative products. A novel product intentionally produced by a person is a creative product, and the person who produces such a product is a creative person. Different Kinds of Creative Contributions If we agree for the moment that creative products are novel works intentionally produced, that does not mean that all creative works are equivalent. First, there are differences in the degree of novelty exemplified by different works. As one example, some of Picasso’s early works, such as the Cubist works he produced in collaboration with Georges Braque in 1912–1914, were much more innovative than was Guernica, with which we are already familiar. The Cubist works went much farther beyond anything Picasso (or anyone else) had done before, whereas, as we saw in the last chapter, Guernica did not represent a radical break with Picasso’s work of that time. We will at a number of points consider whether radically new works are produced by methods different from those involved in producing works that are more closely related to earlier works. The development of Cubism by Picasso and Braque is discussed in Chapter 5. A second distinction that can be made among creative products centers on the influence of the product on the field. Sternberg and his colleagues (e.g., Sternberg, 1999b; Sternberg, Kaufman, & Pretz, 2002) have proposed that an important aspect of a creative product is the effect that product has on the other members of the field. A creative product can change the direction of a field, or propel the field in any of a number of directions. Sternberg and colleagues have developed a classification system that they call a propulsion model of creative contributions. In their view, a creative work can be seen as having the effect of propelling the field in some direction, which may be the same as or different from the direction in which the field is currently moving. This propulsive effect can occur whether or not the creator has intentionally set out to bring it about. There are at least eight different effects a creative contribution can have on a field. Let us examine several of them, to get a feeling for the classification scheme proposed by Sternberg and colleagues. A creative product might simply build on what has already been done, without changing the basic direction in which the field has been moving, in what can be called forward incrementation. Guernica is an example of such a work, since it did not represent a break with Picasso’s then-current style. Similarly, the numerous researchers who built on the development of the double helix used that idea as the basis for forward movement, without introducing any radically new advances. A more radical change is seen in the development of the Cubist style of painting in 1912–1914 by Picasso and Georges Braque. This can be called a redirection of painting, since it took the then-current styles and proposed a radical shift away from them. Numerous painters adopted the Cubist style after it had been introduced, carrying on the innovations of Picasso and Braque and spreading them beyond the ways in which the originators had used them. Thus, in another example of forward incrementation, those artists used the Cubist style in ways not thought of by Picasso and Braque, but those painters and their works did not radically change the direction of painting, as the introduction of Cubism had. Sometimes an individual may make a contribution that is an incrementation beyond what had been done but the degree of advance is too much for the field to absorb, so the contribution is rejected; one might say that it was “ahead of its time.” In this case, the development is an advance forward incrementation. In 1913, Igor Stravinsky’s music to accompany the ballet The Rite of Spring, choreographed by Vaslav Nijinsky, premiered to an almost riotous audience response. Stravinsky’s music (and the choreography) had gone so far beyond the accepted taste that it could not be accepted, and it took years before the music became part of the modern repertoire. An important type of contribution is one of redefinition, in which the individual proposes a new perspective on the current state of the field. An example can be seen in medicine, when in the 1980s Drs. J. R. Warren and Barry Marshall developed the theory that peptic ulcers were caused by bacteria in the stomach (Thagard, 1998). Ulcers, in this view, were the result of bacterial infection. At that time, it was believed that ulcers were caused by excess stomach acidity, often triggered by stress. The typical treatment had been to try to reduce the person’s level of stress and to provide a bland diet that would reduce the irritation of the stomach lining. The bacterial-infection hypothesis was viewed by the medical establishment as absurd, but it is now generally accepted that bacterial infection plays an important role in the development of peptic ulcers. Redefinition can also be seen in art (Sternberg, Kaufman, & Pretz, 2002). One example is the development of Pop Art in the 1960s, when such artists as Roy Lichtenstein and Andy Warhol took objects from “low culture”—comic books and labels from soup cans, respectively—and used them as subject matter for paintings meant to be looked at as “art.” Thus, Pop Art redefined the subject matter of painting. This brief survey has demonstrated some of the ways in which creative products can shape the milieu in which they are introduced. Not only are creative products goal-directed and novel, they also have influences of various sorts on those working in the field and those who will do so in the future. As we consider additional examples of creative production in the chapters that follow, we will have occasion to consider the influence of a product. One important question from the present perspective is whether products that have different effects on their fields are brought about in different ways as far as psychological processes are concerned. For example, are different thought processes involved in redefinition versus redirection? Similarly, are there differences in the people that make contributions of different sorts? These questions will be considered in later chapters. Now that we have defined the important concepts that we will be dealing with, we can begin to look at the still-broader picture. Let us first examine the kinds of methods that researchers have used to study creativity; we will then review the types of theories that have been proposed to explain it. Method versus Theory in the Study of Creativity It is important to make a distinction between method and theory in the study of creativity. On the one hand we have the data that an investigator collects and the method used to collect those data, and on the other we have the theoretical interpretation that he or she applies to the data. One might agree with a researcher’s theoretical perspective but raise questions about the data he or she adduces in support of it—for example, if the researcher’s methods were flawed in some way. As one example, some investigators study the creative personality—the personality characteristics that might be unique to creative people—by studying the characteristics of undergraduates who have been designated as creative based on their scores on tests designed to measure creativity. There are theoretical predictions made about the creative personality that can be tested in this way. In this case, one might agree with the desire to measure the creative personality, and with the specific predictions being made, but disagree with the choice of the specific individuals being studied—that is, with the method being used to collect the data to support the ideas. One might believe that in order to study the creative personality one should study individuals who are acknowledged as creative in the “real world,” such as successful artists, scientists, or inventors, rather than undergraduates chosen as the result of test scores. It is also possible, on the other hand, for one to be interested in the data an investigator produces but disagree with the interpretation that the researcher puts on them. Some researchers study creative thinking by examining the performance of undergraduates on problem-solving exercises carried out in the laboratory. One might find the data interesting as they relate to undergraduates’ capacities to solve problems of various sorts but question the ability of those studies to tell us anything about “real” creativity. One might feel that true creativity was not being displayed in such small-scale exercises. In this case, the data might have been collected using valid methods, but the interpretation of those data is the problem. A similar phenomenon might come about when case studies, such as those presented in Chapter 1, are used to study creative thinking. Such case studies would overcome any possible objection concerning whether “true” creativity is being studied. However, you might disagree with my interpretation of the data from the case studies. I might say that the results indicate that ordinary thinking processes are involved in creative advances, and you might disagree, saying that the results demonstrate otherwise. For these reasons, it is necessary to keep distinct the method used to collect data and the interpretation put on those data. Methods of Studying Creativity Consider again the broad range of creative thinking of which humans are capable, ranging from the creation of poems to skyscrapers. One might expect that such a wide range of phenomena would be brought about in many different ways. In order to study the psychological processes underlying such a broad range of phenomena, one might expect that researchers would have to use a wide range of methods, and that is correct. And, as we shall see, each of those methods has strengths and weaknesses. Subjective Reports of the Creative Process The oldest method of studying creativity uses personal reports by individuals of extraordinary accomplishment concerning how they carried out their work (Ghiselin, 1952). Those reports come from creators from across the spectrum of creative fields, including poetry, literature, music, visual arts, and the sciences, and are in the form of letters, addresses before scientific societies or other groups, responses to questionnaires, and in-depth interviews. In the following passage, Ghiselin makes clear the motivation for use of such reports as the basis for theorizing concerning creativity. [A] large amount of comment and description of individual processes and insights has accumulated, most of it fragmentary, some of it not perfectly reliable. Among these materials the most illuminating and entertaining are the more full and systematic descriptions of invention and the reflections upon it made by the men and women most in position to observe and understand, the thinkers and artists themselves. (1952, p. 11) Let us examine several of those reports, because they seem to provide a fascinating glimpse into the world of the creative genius. The Romantic poet Samuel Taylor Coleridge (1772–1834) provided a description of how he came to write “Kubla Khan,” his famous poem that begins with a description of Khan’s elegant palace (“In Xanadu did Kubla Khan a stately pleasure dome decree”). Coleridge reports that he took an anodyne (a painkiller) because he was feeling indisposed. (This may not be true; it has been reported that he actually took opium, to which he was addicted at the time.) He then sat before the fire, reading a section about Khan’s palace in a well-known travel book describing exotic sights, and he dozed while reading. The following events then occurred (Coleridge refers to himself here as “the author”; qtd. in Ghiselin, 1952, pp. 84–85). The Author continued for about three hours in a profound sleep, at least of the external senses, during which time he had the most vivid confidence that he could not have composed less than from two to three hundred lines; if that indeed can be called composition in which all the images rose up before him as things, with a parallel production of the concurrent expressions, without any sensation or consciousness of effort. On awakening he appeared to himself to have a distinct recollection of the whole, and taking his pen, ink, and paper, instantly and eagerly wrote down the lines that are here preserved. At this moment he was unfortunately called out by a person on business from [the neighboring village of] Porlock, and detained by him above an hour, and on his return to his room, found, to his not small surprise and mortification, that though he still retained some vague and dim recollection of the general purport of the vision, yet, with the exception of some eight or ten lines and images, all the rest had passed away like the images on the surface of a stream into which a stone had been cast, but alas! without the after restoration of the latter. This report is interesting for at least two reasons. First, we are told directly by the poet how he carried out his work. Second, more specifically, we learn that a long poem—200 to 300 lines—came to him “without any sensation or consciousness of effort.” He just had to write it down, until the interruption by the businessman seems to have robbed him and us of the last section. Presumably, those 300 lines and the related images arose out of Coleridge’s unconscious, stimulated by his reading and perhaps by the opium. How else could a complete poem suddenly appear? This report, since it comes directly from the poet, would seem, at least on an initial scrutiny, to carry much weight. Another report concerning creative thinking is found in a letter by Wolfgang Amadeus Mozart (1756–1791) describing his working methods. Mozart began to compose music at the age of six, and he died at an early age, in somewhat mysterious circumstances, after composing more than 600 pieces of music. Reports of Mozart’s legendary abilities abound; he is said to have produced music effortlessly, without mistakes, simply writing down whole compositions as they came to him, without having to revise them. Mozart described his creative process as follows (Ghiselin, 1952, pp. 44–45). When I feel well and in a good humor, or when I am taking a drive or walking after a good meal, or in the night when I cannot sleep, thoughts crowd into my mind as easily as you could wish. Whence and how do they come? I do not know and I have nothing to do with it. Those which please me, I keep in my head and hum them; at least others have told me that I do so. Once I have my theme, another melody comes, linking itself to the first one, in accordance to the needs of the composition as a whole. There are two important aspects of this report. Like Coleridge’s, it comes directly from a creator of the highest level. Second, it too seems to point to the importance of unconscious processing in creative thinking. How else can one explain complete musical compositions being written down without error? Surely, they were worked out in Mozart’s unconscious before they were “presented,” in complete form, to consciousness. The two examples of unconscious processing presented so far have dealt with artistic creation, but evidence for unconscious processing has also been reported by scientists. A critical event in the development of modern chemistry was A. F. Kekulé’s (1829–1896) discovery of the structure of benzene, which forms the basis of many organic compounds. This molecule has a circular structure—it is a closed chain or ring of six carbon atoms—which serves as the foundation for many chemical structures that are crucial for the maintenance of life. Based on Kekulé’s reports of his work, one can separate this critical discovery into two stages. In the first stage, Kekulé had been trying to conceive of a molecular structure for certain compounds made up of carbon atoms at their core, and he did so by imaging how the atoms making up those molecules might interact. During my stay in London I resided in Clapham Road…. I frequently, however, spent my evenings with my friend Hugo Mueller…. We talked of many things but most often of our beloved chemistry. One fine summer evening I was returning by the last bus, riding outside as usual, through the deserted streets of the city…. I fell into a reverie, and lo, the atoms were gamboling before my eyes. Whenever, hitherto, these diminutive beings had appeared to me, they had always been in motion. Now, however, I saw how, frequently, two smaller atoms united to form a pair: how a larger one embraced the two smaller ones; how still larger ones kept hold of three or even four of the smaller: whilst the whole kept whirling in a giddy dance. I saw how the larger ones formed a chain, dragging the smaller ones after them but only at the ends of the chains…. The cry of the conductor: “Clapham Road,” awakened me from my dreaming; but I spent a part of the night in putting on paper at least sketches of these dream forms. This was the origin of the “Structural Theory.” (qtd. in Roberts, 1989, pp. 75–81) Kekulé now had the idea of a string of carbon atoms as the basis underlying the structures of many carbon-based compounds. The structure of benzene seems to have been problematic for Kekulé’s early structure theory, because the properties of benzene were very different from those of other molecules that were made up of strings of carbon atoms with hydrogen atoms attached to them. That raised the possibility that a different kind of structure was involved. In setting the stage for the discovery of benzene, Kekulé reports that he was sitting by a fire. During my stay in Ghent, I lived in elegant bachelor quarters in the main thoroughfare. My study, however, faced a narrow side-alley and no daylight penetrated it…. I was sitting writing on my textbook, but the work did not progress; my thoughts were elsewhere. I turned my chair to the fire and dozed. Again the atoms were gamboling before my eyes. This time the smaller groups kept modestly in the background. My mental eye, rendered more acute by the repeated visions of this kind, could now distinguish larger structures of manifold conformation; long rows sometimes more closely fitted together all twining and twisting in snake-like motion. But look! What was that? One of the snakes had seized hold of its own tail, and the form whirled mockingly before my eyes. As if by a flash of lightning I awoke; and this time also I spent the rest of the night in working out the consequences of the hypothesis. (qtd. in Roberts, 1989, pp. 75–81) Kekulé’s discovery was that the carbon atoms in benzene formed a closed ring. Kekulé reported that when he made the discovery he was dreaming, or in a dreamlike state, which would support the claim that unconscious processes were involved. In addition, Kekulé was dreaming of a snake, which may be evidence of a symbolic component in his thinking, and this again points away from ordinary conscious thought. Thinking of atoms as atoms presumably would not have produced a new structure—the closed ring—that opened up a new scientific domain (Csikszentmihalyi, 1996). We have now considered three firsthand reports, from creative individuals of the first rank, concerning their creative process. When individuals of the highest repute provide the information, it seems that we have learned much. On closer examination, however, those reports turn out to be of little value as evidence for unconscious processes in creative thinking. Consider again Coleridge’s “Kubla Khan,” the marvelous poem that reportedly came to him whole out of his unconscious during a drug-induced dream. The most troubling obstacle to our believing Coleridge is the existence of another version of the poem, written in Coleridge’s hand, which is different from the final version (Schneider, 1953). This other version seems to be an earlier one, because it contains some changes or corrections that appear in the published poem. Thus, contrary to what Coleridge reported, the poem was subject to the usual corrections and changes to which all works of art are subject (Weisberg, 1986). So this report, firsthand or not, should be put aside. Next we have Mozart’s letter, describing a process much like that of Coleridge, in which complete works appear to the creator unannounced and unbeckoned; again the creator simply transcribes them for us. However, Mozart’s letter is also not to be believed, because there is strong evidence that it was not written by Mozart (Weisberg, 1986). It is in a dialect of German that Mozart did not speak, and it refers to his sister by a nickname that he and his family did not use. So this letter too must be relegated to the trash. Finally, there is Kekulé’s report of his dream of a swirling snake biting its tail. There are four points to be made about this account. First, Kekulé’s report was part of an address he gave at a celebratory dinner commemorating his discovery of the structure of benzene, which had occurred some 35 years earlier. Surely we cannot put too much faith in a 35-year-old memory. Second, even if we accept the report, one can raise the question of whether Kekulé’s report says what it has usually been interpreted as saying (Weisberg, 1986). Kekulé is usually described as dreaming in front of the fire, but in his speech he used the German word Halbschlaft or “half-sleep” to describe his state, which seems to refer to daydreaming rather than fully sleeping. This would indicate that he imagined the snakes when he was conscious, rather than unconscious. Let us put this objection aside as well and look still further at Kekulé’s report, which is usually presented as his dreaming of snakes. Kekulé described the strings of atoms as being in “snake-like” motion. This is a curious adjective, because if one calls something “snake-like” it means that the object being described is not a snake. If I tell you that your new car has race-car-like handling, then you must have not bought a race car. If we can extend this analysis to Kekulé’s description of the content of his dream, then Kekulé was not describing snakes. He was comparing the movement of the strings of atoms to that of snakes, but he knew there was a difference. The final point concerning Kekulé’s report concerns whether the story is even true. It has been proposed that Kekulé made it up for presentation at that celebration (Wotiz, 1993). This last point has been the subject of some controversy (Rocke, 1985), but the fact that scholars can publish articles in respectable scientific journals debating whether Kekulé ever had the dream indicates that such reports are not the sorts of data that we can use as the basis for a theory of creative thinking. In the course of this book, we will come across several first-person reports on the creative process. I am reluctant to put much emphasis on such reports, however, because they suffer from several shortcomings, some of which were just pointed out. They are usually made long after the fact, which raises questions about their accuracy; like everyone else, great creative thinkers have fallible memories (Perkins, 1981; Weisberg, 1986, 1993). Even if the subjective report were obtained very soon after the events in question, which might reduce potential memory problems, in most cases we have no way of verifying the accuracy of the report, because there is usually no objective evidence to support it. Furthermore, the individuals providing those reports, although of undeniable eminence in their fields, usually have no training as behavioral scientists, which may limit their ability to provide valuable data, even if they are available (Ericsson & Simon, 1996). And then there is the problem of poetic license, which may be relevant in Coleridge’s case and that of Kekulé. Finally, even if there were no other issues, self-reports provide us with only qualitative descriptions of the creative process. They do not give us data that can serve in a rigorous scientific analysis. For these reasons, contrary to Ghiselin’s (1952) belief quoted above—that is, that the individuals who make creative advances are in the best position to tell us something about the processes involved—the position taken in this book is that the cognitive scientist, who is equipped with tools to analyze objective data, is the individual most likely to make valid observations about the creative process. In no other science—indeed, in no other area, even in experimental psychology—do we rely on such reports; we should also move beyond them in the study of creativity. (For a summary, see Table 2.1.) Biographical Studies In a large step away from reliance on subjective reports of the creative process, Gardner (1993) carried out biographical studies of seven of the most eminent creative individuals of the twentieth century: Sigmund Freud, Albert Einstein, Pablo Picasso, Igor Stravinsky, Martha Graham, T. S. Eliot, and Mahatma Gandhi. Each exemplifies one of Gardner’s proposed multiple intelligences: interpersonal (able to achieve a high level of self-understanding; Freud); logical-mathematical (Einstein); spatial (Picasso); musical (Stravinsky); bodily-kinesthetic (Graham); linguistic (Eliot); and intrapersonal (able to achieve a high level of understanding or of relating to others; Ghandi). More recently, an eighth intelligence has been added (Gardner, 2003)—naturalist (able to discern patterns or regularities in the natural environment; Darwin would be an example)—and possibly a ninth. Table 2.1 Methods used in the study of creativity: Strengths and weaknesses Method Strength(s) Weakness(es) Autobiographical self-reports Highest-level creators Lack of control over accuracy; evidence that several are untrue; limited to qualitative analysis Biography—creative thinking Highest-level creators; historical data accurate Only as good as data are complete; qualitative analysis Biography—personality or psychopathology Highest-level creators Qualitative analysis; retrospective assessment or diagnosis is difficult; only as good as data; data not in form amenable to analysis Case studies of innovations Highest-level advances; historical data accurate Qualitative analysis; only as good as data are complete; method and data may limit generality of conclusions Historiometric analyses Real creativity; hard data; quantitative methods Only as good as data; may be limitations on kinds of questions that can be asked (fine-grain analysis?) Quantitative case study Real creativity; hard data; quantitative analysis Lack of control; only as good as data; possibility of bias in selection of cases In vivo (e.g., Dunbar)—thinking Real situations Lack of control; quality of data; highest level? In vivo—personality Real creators Correlational study; highest level? In vitro (e.g., Dunbar)—discovery Control; approaches real situation Artificiality; is situation too structured by experimenter? In vitro—personality (undergraduates tested) Control Assessment of creativity depends on validity of test; correlational study Laboratory experiment—problem solving Control and analysis of phenomena Artificial situation; is it real creativity? Use of biographical data avoids some of the problems arising from subjective reports; most critically, biographies usually are based on verifiable historical records. Gardner used those biographies to derive a number of conclusions concerning how each individual brought about his or her groundbreaking work. For example, Gardner emphasizes the role of a support group in providing a sympathetic arena in which the individual can introduce radical ideas. The major strength of a biographical study is obvious: It provides direct study of individuals of the highest levels of creative accomplishment. One possible limitation to the biographical method, however, is the quality of the data that are available; any incompleteness or inaccuracy can severely restrict the studies that one can carry out. A quantitatively oriented investigator is also left unsatisfied with biographies as the basis for an analysis of creative thinking. Biographies, while undoubtedly informative, provide little in the way of quantitative data to serve as the basis for scientific theorizing; for example, there are no data tables or graphs in the 400-plus pages of Gardner’s book. Investigators have also used biographical information to make retrospective assessments of psychological characteristics of creative individuals of the highest order. As an example, Jamison (1993; see also Weisberg, 1994) has used biographical information as the basis for diagnoses of bipolarity (manic-depressive illness) in many individuals of great creative renown, such as Vincent Van Gogh and George Gordon, Lord Byron. We will discuss such retrospective diagnoses, which can serve as a kind of personality assessment, in several places later in this book. The obvious strength of such methods is that once again one is dealing directly with creative individuals of the first rank. Using such methods raises problems, however, of several sorts. As with other biographical studies, one is dependent upon the quality of the available information. In addition, attempts to assess psychological characteristics from biographical information must be indirect, since we cannot test an individual who is no longer alive, and such indirect assessments are subject to problems in interpretation of the historical record. For example, if a doctor’s report on a nineteenth-century poet describes him as being “in a frenzy,” does that phrase mean the same thing as it would if it were used today? So biographical studies have several problems that may limit their usefulness. However, if we wish to examine the personality or other psychological characteristics of eminent individuals who are long dead, we may have no choice but to use such information. We can learn something from such studies, but it is important to keep in mind the limits to the conclusions that we can draw from them. Historical Case Studies: Archival Data and Reconstruction of Process A number of studies have examined individual cases of creative achievement, such as Gruber’s (1981) analysis of Darwin’s development of the theory of evolution through natural selection (see also Holmes, 1980, 1996, and Tweney, 1989). Gruber’s groundbreaking and important study was based on archival data (i.e., Darwin’s notebooks), and his work stimulated much interest among psychologists in case studies of creative thinking (e.g., Perkins, 1981; Weisberg, 1986). One difference between Gruber’s historical case study of Darwin and Gardner’s (1993) biographical studies is Gruber’s concentration on Darwin’s theory. That is, one could contrast Gruber’s and Gardner’s perspectives by saying that Gruber presented a biography not of Darwin but of the theory of natural selection. In the historical-case-study approach, the emphasis is less on the creator than on the work. In addition, since Gruber’s study was based on Darwin’s own notebooks, there is no question as to the accuracy of the data. Gruber (1981) drew a number of important specific conclusions about Darwin’s creative process from his case study, which will be discussed later. He also made a general proposal on the basis of his analysis of Darwin: that the creative process is unique in each individual and no generalizations may be made about the creative process or the creative person. In Gruber’s view, each creative accomplishment is carried out by a unique individual working in a unique set of circumstances. This is obviously a conclusion of great potential importance. However, analysis of historical case studies may make it difficult to discover generalizations about the creative process, because one limitation of that method is that it provides little in the way of data to be analyzed in the search for generalizations. That is, historical case studies are usually limited to qualitative analyses. As in Gardner’s (1993) biographical studies, there are no data tables in Gruber’s study of Darwin. Gruber may be correct in his claim that no strong generalizations will come out of historical case studies of creative thinking, but the only way to know is to carry out more of them and to try to use a method that allows any possible generalizations to become clear. So historical case studies may tell us much about a specific person, but they may make it difficult to draw conclusions about creative people in general. In addition, carrying out a case study means that the researcher depends on the availability of information over which he or she has no control. What one gains in authenticity, one may lose in control. Historiometric Methods A number of researchers have applied quantitative methods to historical data in order to formulate and test causal hypotheses concerning creativity, using what has been called historiometric analysis (Simonton; e.g., 1999; Martindale, e.g., 1990; Hayes, 1989). The term historiometric means “measuring history.” As an example, Simonton investigated the influence of war and other social upheaval on creativity, by breaking the last two millennia into 20-year “epochs” and determining for each the frequency of social unrest and of creative accomplishment, based on such measures as the number of years in each epoch in which active war was carried out, and the number of creative individuals who flourished during each epoch. Using statistical methods, Simonton has attempted to distill causal relations from historical data and has concluded, for example, that occurrence of war involving a nation results in a decrease in creative accomplishment in that nation in the following epoch. Also, the occurrence of a significant number of individuals of high levels of accomplishment during one epoch is positively related to the level of accomplishment in the next generation, which Simonton takes as supporting the idea that one generation serves as role models for the next. Similarly, Martindale (1990) has measured changes in the content of French poetry over many generations of poets, in order to test hypotheses about the creative process. In a further variation on this perspective, Hayes (1981, 1989) examined the role of experience—what he called “preparation”—in the production of creative masterpieces. Based on available biographic information, Hayes measured the amount of time that elapsed between an individual’s beginning a career in the fields of musical composition, painting, or poetry, and the production of that individual’s first “masterpiece.” Hayes defined a masterpiece in objective terms as, for example, a musical composition that has been recorded relatively frequently, a painting that is cited in reference works, or a poem that is included in compendia. There was a consistent relationship between time in a career and production of the first masterpiece: All of the individuals in all of the domains that were studied required significant periods of time—approximately 10 years—before production of their first masterpiece. This finding has been codified as the “10-Year Rule.” Thus, Hayes provided quantitative evidence for the claim that immersion in a discipline is necessary before an individual can produce world-class work (see Weisberg, 1999, 2003, and Chapters 4 and 5 in this volume for further discussion). The strength of the historiometric method is obvious: One is dealing directly with creative accomplishment at the highest level. One weakness is one that we have already seen in examining other case studies: the availability of data. If no data are available concerning some individual or individuals, say, then one cannot set up a situation that will produce the relevant information. However, if data are available—as was the case for Simonton, who was able to obtain necessary data on the amount of war and the amount of creative innovation of many epochs of Western civilization—then historiometric case studies can provide a level of analysis beyond that of biographical or historical case studies. Obviously, Simonton (1999) was not able to manipulate the occurrence of war, and Martindale (1990) was not able to control the variables that might have affected the poets he studied, but through different types of statistical analysis those researchers were able to draw some relatively strong conclusions about causal relations in their studies. For example, if it is true that war results in a decrease in creative accomplishment, then one can predict a negative correlation between the intensity of war in one epoch and the amount of creative achievement in the next one. So in carrying out a historiometric case study one is not simply at the mercy of the already available data with no way to determine possible causal relations. Quantitative Case Studies Historiometric methods have usually been applied to the analysis of groups of individuals, over relatively long periods of time. This general approach can, however, be applied to the study of the creative process in individuals; my students and I have analyzed several individual case studies using methods similar to those of Hayes (1989), Martindale (1990), and Simonton (1999; Buonanno & Weisberg, 2005; Ramey & Weisberg, 2004; Rich & Weisberg, 2004; Weisberg, 1994, 1999, 2004). Two examples of this approach were seen in the case studies discussed in Chapter 1, the double helix and Guernica. Bringing a quantitative orientation to case studies (even when it is difficult, as in the case of DNA, to actually quantify things) sometimes results in discoveries that would not have been apparent from only qualitative presentations of historical information. A number of other quantitative case studies will be presented at various points in this book. Quantitative case studies have the obvious strength of the other types of case studies already mentioned: One is studying creativity at the pinnacle. However, one is also at the mercy of the data that are available; sometimes all the ingenuity in the world will not bring forth data on a case of interest. For example, if one were interested in determining whether Picasso was thinking about Minotauromachy when he was painting Guernica (see Chapter 1), one could not find an answer based on the sketches and photos available. One would need more, and, since that information is not available, it seems (at least at the present time) that one will never be able answer that question. A second potential difficulty with quantitative case studies, which actually is relevant for any type of case study, including biographical and historical case studies, is that there are concerns about selectivity in choosing the cases to examine. For example, in responding to the results from the examination of the development of Guernica, students and other researchers have me asked if those results are (1) representative of Picasso’s creative process (that is, did his other works develop similarly?); (2) representative of the creative process in painting (that is, did the works of other artists develop similarly?); (3) representative of the creative process in general (that is, did works in other domains—invention, say, or science—develop in similar ways?). People are concerned that the specific case study chosen might lead to a conclusion that is not really general. This is an important concern, and the only ways to counter it are, first, to investigate as wide a range of case studies, in as wide a range of domains, as possible. Other case studies will be presented later in this book. Second, those case studies should be carried out by as many different investigators as possible, so that the possible biases of any one investigator will not shape any conclusions that are drawn. Carrying out many case studies also has the additional benefit of making it more likely that any conclusions will have wide generality. Studying High-Level Creativity in Real Time: “In Vivo” Investigations Another way to try to study the creative process at the highest level is to observe it directly. Over several years, Dunbar (e.g., 1995, 2001) observed the ongoing activities in four high-level research laboratories in molecular biology. The directors of those labs, all scientists of high repute, gave Dunbar complete access to the laboratories’ activities. He regularly attended and recorded laboratory meetings, discussed ongoing work with the scientists involved, and was given copies of research papers in various stages of completion. On the basis of those observations, Dunbar made several discoveries concerning the processes underlying creative work in those laboratories (2001). As one example, he found that a scientist’s conception of his or her own work sometimes changes radically as the result of input from colleagues during laboratory meetings in which data and analyses are discussed. The scientist alone is less likely to try to deal with recalcitrant data. Dunbar was also able to quantify various aspects of the activities in the laboratories and so was able to test some rigorous hypotheses concerning factors influencing creative activity. Thus, one could say that Dunbar has shown how scientific research can be studied in vivo, and his investigations have produced a number of important results. One obvious strength of Dunbar’s research and other studies like it is that there is no question that the object of the study is real creativity. However, although the laboratories studied by Dunbar are directed by scientists of strong reputation, those individuals and their research groups are as yet nowhere near attaining the significance of Einstein, Darwin, and the like, and so there is still a question of whether the conclusions from Dunbar’s research would illuminate those more illustrious individuals. Therefore, we still need some method or methods that allow us to directly study the creative process in historically significant individuals. Dunbar’s method also does not allow the researcher control over possible extraneous variables, and it also does not enable the researcher to manipulate any variables of interest. Psychologists interested in the creative personality have carried out investigations similar to that of Dunbar, asking creative individuals in many domains—scientists, artists, architects, novelists, poets—to complete personality inventories of various sorts in order to determine personality characteristics associated with creative accomplishment (e.g., Feist, 1993, 1999). Research in this area will be discussed in Chapter 10. The individuals studied in those investigations are usually chosen on the basis of their career accomplishments. They may be nominated by members of the field of interest (e.g., chairs of departments of psychology might be asked to nominate psychologists who have made creative contributions to the profession). These studies have provided a wealth of information concerning the characteristics of individuals at the top of their respective fields. This information is no doubt potentially valuable in understanding creativity, although here too one can also raise the question of whether the results are relevant to individuals such as Picasso and Edison and the like, who cannot be studied in that way. There are also limitations to the kinds of conclusions one can draw from in vivo studies of the creative personality. For example, Feist (1993) concluded that an “arrogant thinking style”—the tendency of an individual to work alone and to be less than completely receptive to the opinions of others—was characteristic of the scientific researchers of high creativity whom he studied. The question arises as to whether having that style of thinking caused those individuals to be creative, and that question cannot be answered by Feist’s study. We do not know if the individuals had that style from the beginning—or at least early in their lives, before they began their scientific careers—in which case it might have played a role in causing their creative accomplishments. It is possible that the arrogant thinking style developed in response to or as part of their success as scientists, which means that the personality style was the result, not the cause, of their creativity. In these ways, in vivo studies of the characteristics of creative individuals are limited in the kinds of inferences they support. Laboratory Investigations of the Creative Process and Creative Individuals Modern psychology has a strong tradition as a laboratory experimental science, and researchers who study creative thinking have carried out experimental studies of various aspects of creativity. Typically, but not always, these center on the study of undergraduates carrying out some task requiring creative thinking, such as the problems in Figure 1.1C. A significant strength of experimental studies is that one can exert control over extraneous variables that might contaminate the interpretation of the results, and one also has control over the variables of interest, so one can draw conclusions concerning cause-and-effect relationships. In one example of an experimental study of creativity, Amabile, Hennessy, and Grossman (1986) asked children to tell a story in response to a series of pictures in a book. The stories were rated for creativity by teachers. Some of the children agreed to create the stories in order to earn an extrinsic reward: They were allowed to play with a Polaroid camera if they agreed that they would tell the story afterward. They “contracted” to create the story in order to be given access to the camera. A second group simply played with the camera and then created the story, with no connection made between the two activities. The children who created the stories because they had contracted to do so—the children who created because of an extrinsic reward—produced stories judged less creative than those produced by the children whose creative activities were motivated only by their interest in the task. Laboratory studies of creativity are potentially valuable because they allow researchers to make strong claims about cause-and-effect relationships among variables. Amabile and colleagues (1986), for example, were able to conclude that working for extrinsic reward had a negative effect—a causal effect—on creativity. Such conclusions are not possible in many of the other sorts of investigations already discussed (see Table 2.1). However, this strength does not mean that researchers studying creativity should rely only on experimental methods and abandon all others, because there are limitations on how much information one can get from laboratory studies. Such renowned creative individuals as Picasso, Mozart, Edison, and Einstein are also part of the population of creative thinkers one wishes to understand, but one may not be able to draw conclusions about such individuals on the basis of controlled investigations of undergraduates or schoolchildren. As I have mentioned several times, it is my belief that the same cognitive processes are involved in all acts of creativity. If correct, that view would mean that one could learn about Picasso or Edison by studying undergraduates in the laboratory. However, at this point that is only an unsupported belief. In order to provide support for it, it is necessary to also study creative thinking at the highest level, if at all possible, to show that the cognitive processes involved in highest-level creativity are indeed the same as those used by undergraduates in the laboratory or by schoolchildren working for intrinsic versus extrinsic rewards. So experimental methods are only one of several that creativity researchers use. “In Vitro” Investigations of the Creative Process An interesting method has been developed that attempts to bridge the gap between in vivo studies of creative individuals at work and controlled laboratory investigations. Dunbar (1995) calls the method the in vitro method, as an analogy to the situation in biological research where a phenomenon of interest is brought out of the organism in which it usually occurs and is studied in a glass (vitro) dish in the laboratory. If the biological researcher has isolated the basic mechanism of interest, the in vitro method allows him or her to exert some control while examining a situation of potential importance. In using the in vitro method to study creativity, one takes important pieces of information from a historically significant discovery and presents them to undergraduates under controlled conditions. The questions of interest are whether the students produce the same discovery and what manipulations are necessary in order to make it possible for them to do so. In one use of an in vitro design, Dunbar (1995) examined undergraduates’ responses to a situation in which they had to “discover” the regulatory function of certain genes in a microorganism. Monod and Jacob were awarded the Nobel Prize in 1965 for their discovery of the existence of regulator genes, which control the activity of other genes through inhibition. Regulator genes turn off other genes until the products of those genes are needed. Monod and Jacob believed originally that some genes served to stimulate other genes, and the discovery of inhibition was surprising to them. In Dunbar’s studies, undergraduates were first taught about genetic influences based on genes’ stimulating or activating other genes. That put them in a state of knowledge similar to that of Monod and Jacob when they had begun their research. The undergraduates were then given a new situation to work through, in which they had to determine the functions of three new genes. The entire experiment was done as a computer simulation, without live organisms. The students were able to conduct virtual experiments of various sorts, based on availability of information, in order to test hypotheses about how the genes worked. Unbeknownst to the students, not all the new genes worked through activation. In the first study, two of three genes in the organism worked through inhibition of other genes; in the second, one of three genes worked through inhibition. The results indicated that the students were able to develop hypotheses concerning the functioning of the new genes, but the ease with which this occurred depended on the relation between their knowledge and the new information. When two of the new genes were inhibitory, there was a much greater likelihood that the students would discover the inhibitory effect, compared to when only one was inhibitory and the other two were excitatory. Dunbar (1995) concluded that the students’ experience with excitatory genes got in the way of their exploring the possibility of inhibitory relationships between genes. When there was only one excitatory gene in the group of three new genes, there was a greater chance that the student investigators would explore other possibilities. Thus, Dunbar tested several hypotheses concerning how creative thinking might work by taking information from a case study and bringing it into the lab, allowing him to exert control over the potentially relevant variables, which he could not do in a historical case study of Monod and Jacob. The in vitro design may serve as an intermediate method between the tightly controlled lab experiment and the less-controlled historical case study. One caution concerning the in vitro design is that the situation simulated in the laboratory may not always be a good match for that in which the original researchers found themselves. That is, how close is the situation of the undergraduates in the laboratory to that of Monod and Jacob in the laboratory in 1965? If the situations are not comparable, then the in vitro lab situation may have little or no connection to the historically important situation that one hopes to understand. One recent interesting method somewhat analogous to the in vitro method involves developing computer models that make scientific discoveries (e.g., Langley, Simon, Bradshaw, & Zytkow, 1987). In those studies, computer programs are developed that can analyze data in various ways that are designed to be close to the methods available to a historically significant creative figure. The program is then given data that correspond to those that the original researcher had, and the question is whether the computer program will discover in those data what the original researcher did. We will discuss several programs of this sort later, but it is worth noting here that one of the objections raised to such studies is that the computer program is fed relevant data, whereas the original researcher had to determine in the first place exactly which data were relevant to the problem he or she was facing. This objection can also be raised concerning in vitro simulations with undergraduates, and it points to a possible problem: In order to gain control over the situation, the researcher may have to control things too tightly. Methods of Studying Creativity: Conclusions Research on creativity uses a wide range of methods (see Table 2.1); studies of the creative process have ranged from randomly chosen undergraduates solving problems or puzzles in the psychological laboratory, to real-time studies of scientists working in their laboratories, to case studies of significant advances, to in vitro laboratory simulations of seminal creative advances. Studies of the creative person have also ranged over a wide variety of methods and participants, ranging from examination of the personality characteristics of undergraduates selected as creative because of their responses on a creativity test, to studies of the personalities of scientists and artists who have been nominated by their peers as being highly creative, to attempts to study retroactively the personality of individuals of acknowledged greatness whom we cannot test or interview directly. In all those methods, the data are objective—that is, they are available for all to see—and thus can support theorizing concerning the creative process. We have also seen that each of those methods has strong points and weaknesses, so we cannot say that only one or two methods are useful in the study of creativity. Depending on the question one wishes to investigate, one chooses the method or methods best suited to answering that question, and one tries to keep in mind the possible weaknesses brought about by the methods one is using and to draw conclusions that are sensitive to the strengths and weaknesses of those methods. Researchers sometimes also rely on reports by creative individuals concerning how they achieved their seminal innovations. I am skeptical about the value of subjective reports concerning the creative process, even those reports given by individuals who have reached the highest levels of creative achievement. As we have seen, many questions can be raised about several subjective reports that have been cited many times in the literature as evidence for various aspects of the creative process. We will in a number of places have occasion to review other such reports, because they have been brought forth as support for theoretical proposals. In each case, the reports will be examined carefully in order to determine whether they can support the conclusions being drawn from them. Such scrutiny is necessary if our attempts to understand creative thinking are to be on a firm foundation. So far in these two chapters we have examined in some detail the development of two creative advances of the first rank, defined the relevant concepts, and discussed the strengths and weaknesses of the broad range of methods used to study creativity. We now turn to a brief review of the broad range of theories that investigators have proposed to explain creativity. An Introduction to Theories of Creativity For as long as humans have thought about where new ideas came from, it has been believed that truly novel ideas that produce creative leaps forward must come from extraordinary sources. Often, the very people who produce those ideas have no awareness of where the ideas came from (“How on earth did I think of that?”). Therefore, in order to understand how that idea came about, many of the creators as well as many theorists have postulated processes outside ordinary conscious thinking that produce the ideas and present them to the conscious thinker. I have found it useful to divide the history of theorizing on creativity into several more or less separate streams, each of which focuses on one general idea or issue, as shown in Table 2.2. Reality, of course, is not as clear-cut as the outline in Table 2.2: The streams are not separate; there is cross-talk between them, as will become evident from later discussions. However, for now the outline in Table 2.2 will serve to orient us to the issues to be dealt with. The Gods and Madness The question of the origin of new ideas has been of interest for several thousand years; early scholars, among them Plato and Aristotle, speculated on how creative ideas came about (Murray, 1989). It was proposed originally by the Greeks that creative ideas were gifts from the gods. Specifically, the Muses—nine daughters of Zeus, each of whom was in charge of a separate domain—played a central role in producing novel ideas. This meant that not only did the ideas originate outside the normal thinking process, they actually originated outside the person. The person served as the messenger or conduit through which the ideas were presented from the gods to the rest of us. The residue of this school of thought is seen whenever someone says that he or she “got an inspiration” or was “inspired” in describing the appearance of a creative idea. Inspiration means “breathing in”; one received inspiration from the Muses, because they breathed creative ideas into people. It was believed that an individual in the throes of creative activity was out of his or her mind, in the sense that an outside source was providing the ideas (Murray, 1989). Plato, for example, described the poet in this way. His description did not mean that the poet was crazy; rather, the creation was occurring as the result of processes outside of the poet’s mind—the inspiration from the gods. However, by the generation after Plato, his student Aristotle had come to the conclusion that states of mental illness could play a role in creativity. In more recent times, beliefs about the sources for creative ideas moved away from the supernatural to internal processes, but the underlying notion—that processes beyond ordinary day-to-day thinking are involved—still remains. Examples of such processes are (1) unconscious thinking, (2) psychopathological thinking, (3) intuitive leaps of insight, and (4) divergent thinking. In addition, some theorists have gone beyond a concentration only on creative thinking processes and also emphasize the roles of personality and environmental factors in creativity. Table 2.2 Theories of creativity Theoretical stream Issue(s) The gods and madness Muses: Plato, Aristotle ⇒ genius and madness: modern interest in creativity and psychopathology Unconscious thinking: associative unconscious; unconscious processing Freud: unconscious conflicts in creativity, associative unconscious Genius and madness: creativity and psychopathology Poincaré: unconscious processing; incubation and illumination Wallas: stages of creative thinking; Modern interest in associative unconscious and unconscious processing Leaps of insight in creativity: the Gestalt view Insight in problem solving and creative thinking Productive versus reproductive thinking Psychometric theories Guilford: testing creativity; divergent thinking; creative personality Other creativity tests Confluence models of creativity: Cognition (general creative thinking and domain-specific thinking); personality; environmental (social) factors Amabile; Sternberg & Lubart Evolutionary theories Campbell: blind variation and selective retention Simonton Cognitive theories Newell, Shaw, & Simon: creative thinking and problem solving Expertise in creative thinking Perkins: Ordinary thinking in creativity Unconscious Thinking We are all familiar with the Freudian conception of the unconscious. Freud applied his ideas to creativity. In addition, a different conception of unconscious processing has also been discussed in the context of creativity (Poincaré, 1913). The Freudian View: Associative Unconscious The notion that unconscious thinking is important in creativity has been with us for a long time, at least since Freud applied his theoretical ideas to the understanding of creativity, and others have carried forward various aspects of the Freudian view to the present (e.g., Csikszentmihalyi, 1996; Gedo, 1980; Koestler, 1964; Rothenberg, 1979; Simonton, 1988; 1999). From the Freudian perspective, unconscious needs and conflicts played an important role in determining both the subject matter that creative individuals dealt with and the way they portrayed it. As an example, consider Leonardo’s portrait of the Mona Lisa, with that enigmatic not-quite-smile that is more distant than welcoming. Freud proposed that the emotional tone of that smile was the result of unfulfilled needs stemming from Leonardo’s early childhood. He was orphaned at an early age, so Freud concluded that the reason that the Mona Lisa looks distant is because Leonardo was expressing through the painting his feelings about his lost mother, who would forever be out of reach. Thus, in Freud’s interpretation, deep feelings from Leonardo’s childhood affected his choice of subject matter and how he portrayed it. A modern example of the influence of the Freudian view on theorizing about creativity is seen in Gedo’s (1980) analysis of Picasso’s creativity. Gedo proposed that the origins of many of Picasso’s works could be traced to childhood trauma. As one example, Gedo proposed that Guernica was an expression of several conflicts from early in Picasso’s life, concerning Picasso’s relationship with his mother and his younger sister. Gedo believes that the women in Guernica represent Picasso’s mother and the dead baby represents Picasso’s younger sister, whose arrival on the scene served to take the spotlight from Picasso in a female-dominated household. Therefore, the portrayal of the baby as dead represents Picasso’s unconscious wish that his sister be removed. He could not actually get rid of her in real life when she arrived, but, years later, in an attempt to satisfy the still-simmering need from childhood, he did so in his art. In theories of creativity based on the Freudian concept of the unconscious, the creator cannot tell us how and why certain ideas surfaced in his or her work, because unconscious connections among ideas lead from one to the next. Since we are not aware of those unconscious connections, the creator on a conscious level knows nothing about where his or her ideas came from. Thus, I will refer to this class of theories as postulating an associative unconscious. The Freudian view, with its emphasis on unresolved conflicts and early trauma, is related to the theory that psychopathology plays a role in creativity. It is not a very large leap to assume that, since the creative individual’s conflicts are close enough to the surface that they show themselves in his or her works, those conflicts may manifest themselves elsewhere in the person’s life as well—specifically, in psychopathology. Among the sorts of mental illness that have been postulated as being related to creativity are schizophrenia and bipolarity (manic depression). In addition, Eysenck (e.g., 1993) has proposed that people who are high in the personality trait of psychoticism have a greater tendency to be creative. Psychoticism is an underlying inherited tendency to become mentally ill when placed in a stressful environment. Psychoticism is not psychosis: People high in psychoticism are normal, although a bit eccentric and hard to get along with. Eysenck found that normal people high in psychoticism exhibit certain characteristics—among them a looseness of thinking—that he believed are potentially important for creativity. We will examine the Freudian theory of creativity and the related views on genius and madness in Chapter 7. Unconscious Processing A different conception of the role of unconscious thinking in creativity was proposed at the end of the nineteenth century by Henri Poincaré (1854–1912), a world-renowned mathematician and scientist (Miller, 1996). In studying his own creative achievements, Poincaré concluded that thought processes that occurred outside his consciousness had played a critical role in his own creative thinking. Poincaré’s view centers on the phenomena of illumination and incubation. Illumination is the sudden appearance in consciousness of a creative idea or solution to a problem when one has not been thinking about the matter consciously—an Aha! experience. Poincaré (1913) reported several illuminations when describing his creative achievements. The occurrence of illuminations was taken as evidence for unconscious processing by Poincaré and many who have followed his lead. If Aha! experiences do not come from conscious thinking, then, so the argument goes, where else can they come from? It has been proposed that unconscious incubation—thinking about the problem unconsciously while consciously thinking about something else—is the explanation for sudden illumination. Poincaré’s view is not the same as the Freudian view, because Poincaré did not assume that the connections among ideas were any different in unconscious thinking. That is, Poincaré’s view was not based on the premise of the associative unconscious. The only difference between conscious and unconscious processing, according to Poincaré, was that multiple thought processes could go on at once; in modern terms, the creator was assumed to be capable of carrying out parallel processing, with consciousness comprising only one stream of that processing. Thus, it is important to differentiate between unconscious associations (postulated by Freud and theories evolving from his work) and unconscious processing (postulated by Poincaré and his followers). Furthermore, those two views are not mutually exclusive; both the associative unconscious and unconscious processing may be true. We will discuss those issues further in Chapter 8. Wallas (1926) elaborated Poincaré’s ideas in a four-stage model of creative thinking, which can still be found in modern theorizing (e.g., Csikszentmihalyi, 1996). Modern psychologists have become increasingly interested in the role of unconscious processing in cognition, so the ideas on unconscious processing in creativity introduced by Poincaré have become more mainstream than ever before. The idea of unconscious processing has also had an impact on analyses of creativity beyond psychology. As one example, Kantorovich (1993), writing in the philosophy of science, has invoked unconscious processing (e.g., Simonton, 1995) as support for a theory of scientific discovery. Also, Miller (1996) has used unconscious processing as the basis for an analysis of developments in the history of science. Chapter 8 will examine evidence for unconscious processing in creativity. Leaps of Insight in Creativity: The Gestalt View An idea related to the role of unconscious processing in creative thinking is the notion that problems are sometimes solved, and creative ideas in general sometimes come about, as the result of leaps of insight, a notion introduced broadly into psychology by the Gestalt psychologists early in the twentieth century. Who among us has not had such an experience, if only when suddenly remembering a name that has slipped our mind? Leaps of insight, or Aha! experiences, come about when a new idea seems to flash into consciousness from nowhere, bringing with it a way of looking at a problem that is totally different from what one had just been thinking about. According to the Gestalt psychologists (e.g., Wertheimer, 1982), true creative advances require that the person use productive thinking to go beyond what had been done before. Staying with what had been done before was dismissed as mere reproductive thinking, in the sense that it reproduced what had been done before. Furthermore, if one relied on the past and “mechanically” reproduced habitual responses, one would not be able to deal with the particular demands of any novel situation that one might face, and would therefore be doomed to failure. The notion of leaps of insight is related to other views of extraordinary thinking, because sometimes those leaps are explained by positing an unconscious processes. Poincaré, for example, who as we know was one of the first to postulate unconscious processing in creative thinking, brought that theory forth in order to provide an explanation for several leaps of insight that he experienced. In more modern versions of this view, Csikszentmihalyi (e.g., 1996) and Simonton (e.g., 1988, 1999) both postulate unconscious processes as the basis for creation of new ideas. However, the notion of leaps of insight can be examined independently of questions concerning conscious versus unconscious processes in creative thinking, because the occurrence of leaps of insight does not necessarily mean that unconscious thought processes are involved. We will examine the question of leaps of insight in creativity in Chapter 6, separately from the question of unconscious processes. Confluence Theories of Creativity: Divergent Thinking and the Creative Personality There was a significant change in the direction of research on creativity around 1950, as the result of Guilford’s (1950) presidential address to the American Psychological Association. Guilford, an expert on intelligence testing, surprised many people by proposing in his address that psychology had not spent enough time examining thinking that went beyond the kind of thinking measured by IQ tests—that is, creative thinking. He outlined a theory of how creative thinking worked, and, using intelligence testing as a guide, he proposed a set of tests that could be used to measure creative-thinking ability and to identify individuals with creative potential. One important component of creativity is the ability to see that a problem exists in some area. For example, if two individuals use the same appliance and one is dissatisfied with its performance, he or she might attempt to create an improved version of that product. That person has demonstrated sensitivity to problems, which may be necessary to set the creative process in motion. A person who sees no problem with the appliance will have no chance to create something. Concerning the role of knowledge in creative thinking, Guilford reasoned, similarly to the Gestalt psychologists, that an important step in the creative process must be a breaking away from the past, which is the function of what he called divergent thinking. As the name implies, this type of thinking diverges from the old and produces novel ideas, which can serve as the basis for a creative product. Once divergent thinking has served to produce multiple new ideas, then convergent thinking can narrow down the alternatives to determine the best one. The highly creative individual is assumed to be high in divergent-thinking ability (e.g., Csikszentmihalyi, 1996); that is, the highly creative individual is creative, at least in part, because he or she is capable of producing many novel ideas. It must be emphasized that there are two uses of the term divergent in the context of discussions of creative thinking. On the one hand is the ordinary usage, as when we say that a great creative advance, like Watson and Crick’s discovery of the double helix, diverged from what others at the time had been thinking about. In the sense of the new product being different from the old (diverging from the old), that use of the term is based on ordinary language and is completely straightforward. However, there is a second use of divergent, as exemplified in the phrase divergent thinking, proposed by Guilford (e.g., 1950) as part of his theory of creative thinking. In this case, one is referring to a special kind of thinking. In Guilford’s view, creative thinking works in two stages: Divergent thinking is the first stage, which produces numerous ideas that then serve as the input to convergent thinking, the second component of the process. Convergent thinking takes the ideas produced by divergent thinking and narrows them down into a workable product. It is this second meaning of divergent—divergent thinking as a theoretical term—that I will be referring to numerous times in this book. Creative products obviously diverge from the past, and therefore they are by definition the result of thinking that can diverge from the past (the first use of diverge). However, that conclusion does not mean that creative products are ipso facto the result of divergent thinking as presented in Guilford’s theory, where it means a special type of thinking skill. That is, one might be able to produce new things without using a special kind of thought process that breaks away from the past. Guilford’s (1950) work was the stimulus to the psychometric stream of creativity research, which has focused on measuring the psychological characteristics of creative people (psychometric means “measuring the mind”; Feist, 1999; Plucker & Renzulli, 1999). Many people took Guilford’s proposal to heart, and his ideas were carried forth in a number of ways. First, Guilford and others used his tests in attempts to measure the thought processes underlying creative thinking. Second, other investigators developed their own tests (e.g., Torrance, 1974; Wallach & Kogan, 1965), which differed from Guilford’s in various ways. In addition, current theoretical analyses of creative thinking sometimes assume that it is based on divergent thinking, even if that term is not explicitly used. For example, Simonton (1999, p. 26) proposes that the creative process must begin with “the production of many diverse ideational variants.” Those “ideational variants”—that is, numerous and varied ideas, produced presumably as a result of divergent thinking in Guilford’s terms—provide the basis for the thinker’s ability to deal with the new situation that he or she is facing. Guilford’s ideas on testing for the potential to think creatively, and research that followed from those ideas, will be critically reviewed in Chapter 8. In addition to his discussion of types of thinking processes that might underlie creativity, Guilford proposed in his presidential address that a person’s personality was involved in making that person creative. That proposal stimulated work that tried to uncover the creative personality, that is, those aspects of personality that were prevalent in people of great creative accomplishment and that were not present to the same degree in “ordinary” people (e.g., Feist, 1999). It was proposed that those personality characteristics were important in the person’s innovations. For example, it was suggested that a flexible personality structure allowed a person to think flexibly, which was assumed to be necessary for the person to think creatively. Research on the creative personality will be reviewed in Chapter 9. The psychometric perspective has led to the development of confluence models of creativity, which assume that creative products arise when there is a confluence—a coming together—of several factors, all of which are needed for creative production to occur: Creativity requires a person with a particular thinking style, knowledge base, and personality, who is in a particular environment (e.g., Amabile, 1983, 1996; Simonton, 1999; Sternberg & Lubart, 1995). It is assumed that thinking skills of various sorts are involved in creative thinking, both general creative-thinking skills and skills specific to the domain in which the person is working, such as painting versus science versus poetry. In addition, personal characteristics are assumed to be critically important in determining whether the person will put those thinking skills to use in the task of producing innovation. Finally, the environment can play a role in fostering or interfering with creative production. I will briefly sketch several influential confluence models here; they will be discussed in detail in Chapter 11. Two examples of confluence models of creativity are outlined in Table 2.3. Amabile’s Componential Theory of Creativity Amabile (e.g., 1983, 1996) developed an early theory that proposed that creativity was the result of the coming together of several components, some related to the person and others related to the environment, including the social environment. She was one of the first theorists to incorporate social-psychological factors in thinking about creativity. The original theory developed by Amabile is presented in Table 2.3; in Chapter 9, I will discuss newer developments in this viewpoint. The first component in Amabile’s (1983) theory consists of domain-relevant skills, which are skills relevant to the domain in which the individual is working. They include knowledge and technical skills, such as the ability to play a musical instrument or paint, as well as any talent that the domain demands. Some of those skills are based in innate abilities, while others are acquired through formal and informal education. The second component of the theory is creativity-relevant skills, which are skills that go beyond any specific domain and that can be applied to any domain in which one might be attempting to produce innovation. Creativity-relevant skills incorporate methods of breaking set during problem solving—that is, abandoning an unsuccessful approach to a problem—as well as knowledge of heuristics, or rules of thumb, for generating novel ideas. An example of such a heuristic is the guideline “When all else fails, try something counter-intuitive” (Newell, Shaw, & Simon, 1962). Table 2.3 Two confluence models of creativity The creativity-relevant skills are interesting because many if not all of them are based on the assumption that creativity depends on breaking away from the past. That is, one is not creative by staying within what one already knows. It seems that this notion is contradicted by the development of Guernica and the discovery of DNA, the two case studies presented in Chapter 1. In neither case did the creative thinkers reject the past: Rather, they built solidly on it. Obviously, two case studies cannot serve as the basis for a very general conclusion concerning the role of the past in production of the new, but those cases raise questions about the notion of creativity-relevant general skills that serve to enable the thinker to break from the past. This example also demonstrates the usefulness of having a database at the ready to apply to claims made by theories of creativity, as we will do in many places in this book. According to Amabile (1983), the person’s attitude toward the task is critical in determining whether he or she will respond creatively to it. If the person finds the task intrinsically motivating—that is, if he or she is interested in the task for its own sake and not because of some extrinsic reward that might come about as a result of successful performance—the chances of the person’s producing an innovative response will be maximized. One hears individuals who work in creative fields—writers, artists, scientists—say again and again that they do what they do because they love it, and the fact that they make a living doing it is a bonus. There is also some evidence that being extrinsically motivated (carrying out a task in order to gain some reward, such as making money or winning a prize) may interfere with creative work, although there have been conflicting findings in this area (Eisenberger & Cameron, 1998). In addition, if the person believes that he or she has independently chosen to work on the task, the outcome will be more creative than if the person believes that he or she is working on the task because of external pressures; the attitudinal / motivational components of the individual’s status at any time are affected by the social environment. Amabile’s discussion represented an early attempt to broaden the factors taken into account when attempting to understand creativity. Sternberg and Lubart’s Investment Theory of Creativity Sternberg and Lubart (e.g., 1995; see also Rubenson & Runco, 1994) have proposed an analysis of creative thinking based on economic principles, which assumes that the creative thinker buys low and sells high. Buying low means that the creative thinker tends to propose ideas that are unpopular but have potential for growth. That is, the ideas could become popular with a little help. Due to the creative person’s perseverance and ability to convince others of the value of the new ideas, they will become accepted. At this point, the creative thinker will sell high: He or she will give up work on the now-popular idea, and move on to some now-unpopular idea, to start the whole process again. Carrying the economic analogy further, Sternberg and Lubart propose that the person capable of creative production must possess several resources, some of which are as follows: a set of intellectual abilities, with three of particular importance: the ability to see problems in new ways and go beyond ordinary ideas (which is similar to Guilford’s (1950) divergent thinking); the ability to recognize which ideas are worth pursuing; and the ability to persuade others of the value of one’s ideas knowledge of the domain, although too much knowledge can interfere with generation of new ideas a personality that allows you to think independently, which is necessary if you are to defy the crowd and advocate ideas that most others do not agree with an environment that supports and rewards creative ideas Evolutionary Theories of Creativity: Blind Variation and Selective Retention Campbell (1960) proposed an analysis of the creative process based on Darwin’s theory of evolution through natural selection. In Darwin’s theory, species change randomly from one generation to the next due to such factors as mutation, which is a blind process. When a mutation occurs, there is no intelligence that directs it. Some of those changes, although produced blindly, have positive effects on the survival and reproductive abilities of the organisms that possess them. Those organisms will have a greater chance of passing on those characteristics to the next generation, since they have a greater-than-average chance of reproducing. Thus, those changes will be “selected” by nature, and the species will evolve. In Campbell’s view, a similar process is at work in creativity. First comes the blind or random generation of ideas in response to some problem. In Campbell’s view, if true creativity is involved in dealing with some situation, then it must come about through the rejection of the past as the basis for constructing the new. Otherwise, the response will not be truly creative. Once an idea or ideas have been generated through a blind or chance process, each is then subjected to testing to determine if it meets the present needs. One or more ideas may then be retained, to be used in similar situations at a later time. Thus, Campbell proposed an evolution of ideas analogous to evolution of species, with similar underlying mechanisms. Simonton (e.g., 1999) has taken Campbell’s basic ideas and elaborated them into a wide-ranging confluence theory of creativity, which incorporates the blind-variation-selective-retention mechanism with other components, such as cognitive factors, personality characteristics, and environmental influences on the creative process. Simonton’s theory has broad implications, as he has attempted to incorporate many phenomena within his theory. In the discussion of historiometric methods earlier in this chapter, we discussed his analysis of the influence of war on creativity. That is but one of the ingenious methods that he has developed in order to bring under scrutiny many phenomena that one would have thought were outside the range of methods of scientific psychology. We will examine the evolutionary perspective in detail in several later chapters, in conjunction with discussions of unconscious processing in creative thinking as well as discussions of confluence theories of creativity. Cognitive Perspective: Creative Thinking and Ordinary Thinking One idea that we have come across frequently in the theories of creativity that we have just reviewed is that a critical part of the creative process is to break with the past. In order to produce innovation, one cannot depend on what one knows, because true creativity demands something new. Because of this need to break with the past, many theories of creativity postulate extraordinary thought processes of one sort or another, because ordinary conscious thinking is closely tied to the past. The idea that creativity must break with the past has become part of our culture, as evidenced in the often-heard comment about the need for “out-of-the-box” thinking in situations demanding creativity. The “box” that one must get out of is the constraints that the past, in the form of our experience and habits, imposes on us. Examples of such out-of-the-box processes are productive thinking and leaps of insight, divergent thinking, and set-breaking skills. The general perspective that underlies this view can be summarized as an assumption that a tension exists between creativity and experience. The view of creativity that underlies this book assumes, in contrast, that creative products come about through the use of ordinary thinking processes; creative thinking is simply ordinary thinking that has produced an extraordinary outcome (Weisberg, 1986, 2003). From this perspective, when one says of someone that he or she is “thinking creatively,” one is commenting on the outcome of the process, not on the process itself. Although the impact of creative ideas and products can sometimes be profound, the mechanisms through which an innovation comes about can be very ordinary. This perspective is called the cognitive view because it was developed by Newell and Simon (e.g., 1972; Newell, Shaw, & Simon, 1962), leaders of the cognitive revolution in psychology that began in the 1950s. The cognitive revolution was so called because Newell and Simon advocated the study of cognitive processes—internal mental processes—as the avenue to the understanding of human functioning. One early expression of the cognitive view was a paper by Newell, Shaw, and Simon (1962) in which they proposed that creative thinking was basically the same as the thinking involved in solving ordinary problems. My take on the cognitive perspective is a bit different from that of Newell and Simon; in my view, in order to understand creative thinking, we must consider ordinary thinking in a wider sense than just problem solving (Weisberg, 1986, 2003; see also Perkins, 1981). There may be times when we think creatively without specifically solving a problem, but we may still do so using ordinary thought processes. Furthermore, problem solving is a complex activity, made up of simpler cognitive components, and those components should also be considered if we are to understand the structure of creative thinking. Accordingly, I will begin the introduction to the cognitive perspective on creativity in Chapter 3 by examining the idea that ordinary thinking serves as the basis for creativity. This requires that we begin with a consideration of what “ordinary thinking” entails. We can then examine the DNA and Guernica case studies for evidence that ordinary thinking was involved in those seminal advances. This will lead to an examination in Chapters 3, 4, and 5 of details of the cognitive perspective on problem solving and creative thinking. Theories of Creativity: Conclusions We have now examined several theoretical perspectives that have played important roles in directing research on creativity. This relatively brief review has provided enough specifics to position each perspective in history relative to the others. As mentioned earlier, the theoretical positions are presented here as more or less separate entities, with no communication among them. However, there is often overlap among theories. For example, Simonton’s evolutionary view is part of a larger confluence theory that he has developed, and the evolutionary perspective contains as one component the notion of unconscious processing (e.g., Simonton, 1999). Given this inevitable complexity, Table 2.2 will still serve as a useful outline to the views that will be discussed in more detail later.

CHAPTER 8Out of One’s Mind, Part II: Unconscious Processing, Incubation, and Illumination In the last chapter, we examined one component of the “out of one’s mind” perspective: the question of genius and madness, and the possible influence of psychopathology on creative thinking. We concluded that there was only mixed support for the simple notion that psychopathology causes people to think creatively and that there is a complex relationship between psychopathology and creativity. The causal link might in some cases actually work in the opposite direction, with creativity causing psychopathology. This chapter examines a related notion: that the unconscious plays a critical role in creativity. The logic behind proposals that unconscious thinking is the basis for creativity is the same as that which underlies theories on genius and madness: It is assumed that ordinary conscious thinking cannot produce novel ideas, so some other source is designated. Outline of the Chapter The present chapter will consider several variants of the notion that unconscious cognitive processing is crucial in creative thinking. This work centers on the phenomena of illumination and incubation. Illumination is the sudden appearance in consciousness of a creative idea or solution to a problem when one had not been thinking about the matter consciously—an Aha! experience. Who among us has not had such an experience, if only when remembering suddenly a name that had slipped our mind? The occurrence of illuminations has been taken as evidence for unconscious processing, because if Aha! experiences do not come from conscious thinking, then, so the argument goes, where else could they come from but the unconscious? It has been proposed that unconscious incubation—thinking about the problem unconsciously while you are consciously thinking about something else—is the explanation for sudden illumination. However, not all researchers are comfortable with the concept of unconscious processing as an explanation for psychological phenomena, and several have proposed explanations for sudden illuminations that do not rely on unconscious processes. I will take a historical perspective in examining the variations on the theories of unconscious thinking that have been proposed by psychologists studying creative thinking. In the literature on creativity, the notion that we can be carrying out unconscious thinking has its origins in numerous reports presented by creative individuals concerning how they produced their creative works. We will begin with a set of those reports produced by a thinker of great renown, the mathematician-scientist Henri Poincaré, concerning his creative-idea production. As we shall see, Poincaré’s views have been of great importance in modern theorizing about creative thinking. We will then consider what modern researchers have had to say about unconscious processing by tracing the development of modern views as they have built on and elaborated Poincaré’s ideas. I will critically analyze the various views of unconscious processing in creative thinking and the evidence brought forth to support them. The conclusion of the chapter is that evidence for unconscious processing in creative thinking is very weak. The final section of the chapter will examine recent theorizing that has attempted to explain sudden illumination without assuming that unconscious processing occurs. We will consider several alternatives to the view that unconscious processing underlies creative thinking. Unconscious Associations and Unconscious Processing There are two components of the idea that the unconscious plays a role in creative thinking. The first component emphasizes unconscious connections among ideas; that is, this theory, which of course stems from Freud, asserts that our ideas are sometimes linked for reasons of which we are not aware. For example, an adolescent may have a dream in which he plays hockey against his father and wins, although neither he nor his father can even ice-skate, much less play hockey. It takes a trained therapist to unearth the real meaning of the dream. In Freudian terms, such a dream would symbolize the boy’s Oedipal wishes, which center on his desire to remove the father from his life and have his mother all to himself. This wish, which is too threatening to be allowed to become conscious, must be expressed symbolically in the dream in a form that is nonthreatening. Similarly, we discussed in the last chapter Freud’s proposal that creative products often have meaning on a symbolic level, which also goes far beyond their surface appearances. Consider again the smile on the Mona Lisa of Leonardo da Vinci. In Freud’s view, Leonardo painted the Mona Lisa with that reserved smile because he had lost his mother as a boy. Because of that loss in childhood, Leonardo had an unconscious yearning to be united with a mother figure, a yearning that could never be assuaged, since he was no longer a child. Therefore, he painted women who looked emotionally withdrawn and slightly out of reach, although he himself may not have known why. We can call this component of the Freudian view the associative unconscious, because the links, or associations, that lead from one idea to the next are not open to conscious awareness. The associative unconscious is one facet of the out-of-one’s-mind theory, because the person has no conscious awareness of, or control over, the links among his or her ideas. If you ask someone why he or she thought of some idea, and if that idea was the result of the functioning of unconscious associative links, the individual will say that he or she does not know where it came from. This contrasts with the situation in which one can explain to someone else how one’s recent ideas led one to the next. The notion of the associative unconscious can be extended directly to creative thinking: The unconscious can link ideas that never would be brought together in conscious thinking. Those unconscious links are active while the artist is working on a project, for example; while Leonardo might have felt that he was consciously working out how to paint the woman in the Mona Lisa, unconscious links were providing hidden direction. If you asked the artist why he was painting that woman in that way—with that particular smile, say—he might have replied that he found her of interest, and little further. The true basis for that interest may lie deep below the surface, in emotional links laid down many years earlier. A second aspect of the unconscious has also received attention from modern researchers. This notion, which I will call simply unconscious processing, assumes only that we can be working on more than one project at once, using what is called parallel processing. In this view, we can be carrying out some activity of which we are perfectly conscious—say, driving to work and listening to the radio—while at the same time, on an unconscious level, processing may be occurring on some entirely different task—say, solving some problem that has arisen at work. The distinction between unconscious processing and the associative unconscious is between processes that actually carry out some sort of cognitive activity, albeit on an unconscious level (unconscious processing), and the material they work on, which may be organized through links that are hidden from conscious awareness (unconscious associative connections). The theory of unconscious processing can also be seen as a variation on the out-of-one’s-mind theory. While it is neither psychopathology nor the Muses that produce novel ideas (see Chapter 7), those ideas are still being produced by processes over which the person has no conscious control. The idea of unconscious processing has been elaborated in two ways when applied to creative thinking. On the one hand, it has been proposed that unconscious processing is of the same sort as ordinary conscious processing: All the associative links are the same, and the unconscious is simply conscious thought “gone underground.” Nonetheless, unconscious processing can still produce creative leaps: A person can be thinking about one thing when suddenly there flashes into consciousness a novel idea, an Aha! experience, which is relevant to a completely different topic. This leap occurs because the person had been processing in parallel, without knowing it. Once the idea occurred, however, since the associative links are the same as those used in ordinary conscious thinking, the person who produced it should be able to understand whence it came. It has also been proposed that the links through which unconscious processing works are different from those that underlie conscious thinking; that is, unconscious processing can work with unconscious associative links (the associative unconscious). Thus, in this two-component view, a creative leap can come about because (1) the processing has occurred on an unconscious level, which results in the thinker’s being surprised by the sudden leap; and (2) the leap is based on connections that the person could never think of using conscious thought, which is a second source of surprise. As we will see, both components of the unconscious have been discussed by researchers. The discussion so far is summarized in Table 8.1. We see there the two dimensions along which we can analyze unconscious processing: (1) whether multiple streams of thought are possible, and (2) whether the links from one thought to the next are comprehensible to the thinker. We have discussed two possibilities on each of those dimensions, so there are four possible structures for the thought processes underlying creativity (as well as all other thinking). Those possibilities will be considered in this chapter. We now turn to a seminally important analysis of the role of the unconscious in creative thinking, that of Poincaré (1913). Poincaré’s Theory of Unconscious Creative Processes It could be argued that the modern psychological study of creative thinking began with Poincaré (1854–1912), who carried out world-class work in a variety of fields in the latter part of the nineteenth century (Miller, 1996). Poincaré’s great accomplishments were recognized in a singular way by the French Academy of Sciences, an honor society whose membership is reserved for those who have made contributions at the highest level to their disciplines. Individuals of great accomplishment are usually elected to the Academy in one discipline in the sciences. Poincaré, in contrast, was elected to the Academy in all five of its different disciplines, and was also elected president of the Academy. (A mathematician I know once described Poincaré as the last man to know everything.) Table 8.1 Two dimensions of the unconscious. Processing mode Associative links Conscious Unconscious One stream (no unconscious processing) (A) One train of thought; links consciously worked out and are understood; outcome not surprising to the thinker. Example: Painter decides to paint a given subject and will be able to explain why. (B) One train of thought; some links not understood. If those links contribute to outcome, thinker will not be able to explain how that outcome came about. Example: A painter will not be able to explain why a painting turned out how it did. Multiple streams (unconscious processing) (C) Multiple trains of thought; sudden solution to problem is surprising, because person does not know that he or she is thinking about something outside of consciousness; after it occurs, derivation of outcome can be understood. Example: Person will solve a problem in an Aha! experience and will be able to explain where solution came from. (Theorist: Poincaré) (D) Multiple trains of thought; sudden solution is surprising, because person does not know that he or she is thinking about something outside of consciousness. Some links not understood; if they contribute to outcome, derivation of outcome will be impossible. Example: Person will solve problem in Aha! and will not be able to explain where solution came from. (Theorists: Freud; Koestler; Simonton; Csikszentmihalyi) Not surprisingly, Poincaré was also interested in the creative process, and he presented a report describing how several of his most important mathematical breakthroughs came about (Poincaré, 1913). Poincaré’s ideas on this issue are important, because, as we shall see, much modern theorizing concerning creative thinking is based directly on them. Some of the mathematical concepts mentioned by Poincaré in the following excerpts may not be familiar to you, but you can still get the cognitive import of the report even without understanding all the terms. One point to keep in mind when reading Poincaré’s self-reports on his creative process is the inherent weaknesses of those sorts of reports as evidence for scientific theorizing, as already noted in Chapter 2. A self-report, because of its particular nature, is usually unverifiable. This problem exists with Poincaré’s reports: We cannot determine if they are accurate. However, since those particular reports have been of singular importance in the development of theorizing about creative thinking, they are worth careful review. In addition, as we examine the theories that have built on Poincaré’s ideas, we will have the opportunity to investigate whether more recent work is built on firmer ground than self-reports. Poincaré’s Self-Reports The first critical segment of Poincaré’s work involved his attempt to prove that a certain sort of mathematical function could not exist (Miller, 1996). He was actually interested in the existence of those functions, but he set out to prove the opposite. This is a not-untypical method among mathematicians: They attempt to demonstrate that a mathematical object of potential interest cannot exist, hoping that in doing so they will find a contradiction in their reasoning that allows the conclusion that the object must exist—thus reaching the outcome they actually desired from the beginning. Poincaré worked without success for 15 days on this task. His routine was to work on mathematics 4 hours per day, from 10 AM to 12 PM and 7 PM to 9 PM (Miller, 1996). One night, after a typically unsuccessful day, he drank black coffee and could not sleep. He then had an extraordinary experience. Ideas rose in crowds; I felt them collide until pairs interlocked, so to speak, making a stable combination. By the next morning I had established the existence of a class of Fuchsian functions…. I had only to write out the results, which took but a few hours. (1913, p. 387) Thus, during this sleepless night of thinking, Poincaré established that one example of the presumed-impossible functions could be shown to exist. He called them Fuchsian functions in honor of Lazarus Fuchs, a mathematician whose work had influenced his dissertation research (Miller, 1996). Although Poincaré was obviously conscious when those ideas arose, he felt that the thinking was of an extraordinary sort, since it occurred during sleeplessness brought on by the coffee. He felt himself to be merely an observer of what was happening, playing no role in directing the thought process. He therefore concluded that he was observing the workings of his own unconscious, which, as he reports, involved ideas being combined until “stable” combinations were found—that is, new ideas that “held together,” presumably because they were of potential value. After discovering Fuchsian functions, Poincaré went to Coutances, a city near his home at Caen, to attend a geological conference (Miller, 1996). This previously scheduled trip interrupted his mathematical work. While away, he made another discovery, which was totally unexpected. The incidents of travel made me forget my mathematical work. Having reached Coutances, we entered an omnibus to go some place or other. At the moment when I put my foot on the step, the idea [of the equivalence of Fuchsian functions and the transformations of non-Euclidean geometry] came to me, without anything in my former thoughts seeming to have paved the way for it…. I did not verify the idea; I should not have had time, as, upon taking my seat in the omnibus, I went on with a conversation already commenced, but I felt a perfect certainty. On my return to Caen, for conscience’ sake, I verified the result at my leisure. (1913, p. 388) So, in the midst of a conversation having nothing to do with mathematics, Poincaré had the realization that the recently discovered Fuchsian functions were identical to a set of functions already existing in mathematics, the transformations of non-Euclidean geometry. At the moment of illumination, he was engaged in a conversation about something else, so, as far as he could determine, none of his previous conscious thoughts led up to it. In order to explain this sudden illumination, Poincaré concluded that he must have been thinking about those concepts all along, but on an unconscious level. It is also noteworthy that Poincaré felt certain the idea was correct without having to verify it. Where might such a feeling have arisen? We will shortly learn how Poincaré explained the occurrence of that feeling of certainty. When he returned from his trip, conscious work demonstrated that his illumination had indeed been correct. A similar phenomenon occurred soon after Poincaré returned to Caen. [After returning to Caen] I turned my attention to the study of some arithmetic questions apparently without much success and without a suspicion of any connection with my preceding researches. Disgusted with my failure, I went to spend a few days at the seaside, and thought of something else. One morning, walking on the bluff, the idea came to me, with just the same characteristics of brevity, suddenness and immediate clarity, that the arithmetic transformations of indeterminate ternary quadratic forms were identical with those of non-Euclidean geometry. (1913, p. 388) Again Poincaré made a connection between two concepts, and again that connection seemed to be brought about outside of his conscious thought. Poincaré believed that those incidents demonstrated the importance of unconscious processes in creative thinking. He concluded that a sudden illumination was “a manifest sign of long, unconscious prior work. The role of this unconscious work in mathematical invention appears to me incontestable” (1913, p. 389). Poincaré’s Theory of Unconscious Processes in Creative Thinking From his observations during his sleepless night of crowds of ideas colliding until pairs interlocked, Poincaré concluded that the unconscious works by attempting to build combinations of ideas. Although Poincaré was discussing mathematical invention, it has been assumed by him and by others (e.g., Campbell, 1960; Csikszentmihalyi & Sawyer, 1995; Koestler, 1964; Miller, 1996; Simonton, 1995) that similar processes are at work in all creative thinking. Because Poincaré’s theory of unconscious processing is a central component of much modern theorizing concerning creative thinking, I will examine it in some detail to put modern work in historical context. Poincaré (1913) first examines the question of the definition of creativity in mathematics. In fact, what is mathematical creation? It does not consist in making new combinations with mathematical entities already known. Any one could do that, but the combinations so made would be infinite in number and most of them absolutely without interest. To create consists precisely in not making useless combinations and in making those which are useful and which are only a small minority. Invention is discernment, choice. The mathematical facts worthy of being studied are those which, by their analogy with other facts, are capable of leading us to the knowledge of a physical law. They are those which reveal to us unsuspected kinship between other facts, long known, but wrongly believed to be strangers to one another. Among chosen combinations the most fertile will often be those formed of elements drawn from domains which are far apart. Not that I mean as sufficing for invention the bringing together of objects as disparate as possible; most combinations so formed would be entirely sterile. But certain among them, very rare, are the most fruitful of all. (p. 386) Thus, for Poincaré creation ultimately involves discovering valuable combinations of ideas. The combinations that are potentially most fruitful are those that form analogies between facts that, because of their remoteness, had not previously been considered as being related. We have seen two examples of this already, in Poincaré’s reports concerning his discoveries of the equivalences of the transformations of non-Euclidean geometry with two other entities. However, it is not enough just to combine ideas, even those from domains that are far distant: One must also have a way of avoiding the “sterile” combinations that usually result from bringing together disparate ideas. Mechanisms of Combination of Ideas There are at least two possible ways in which valuable combinations of ideas might be produced: (1) The thinker might, by some great skill or intuition, produce only potentially valuable ideas; or (2) the thinker might produce large numbers of combinations, valuable and sterile alike, and then choose for further contemplation only those that are valuable. The conscious experience of the thinker corresponds to the first alternative: He or she is aware of only potentially useful ideas, as Poincaré (1913) asserts: The sterile combinations do not even present themselves to the mind of the inventor. Never in the field of his consciousness do combinations appear that are not really useful, except some that he rejects but which have to some extent the characteristics of useful combinations. All goes on as if the inventor were an examiner for the second degree who would only have to question the candidates who had passed a previous examination. (pp. 386–387) Although one’s conscious experience is not crowded with useless combinations of ideas, Poincaré believed that the actual creative process works differently: Many ideas are produced by unconscious processing, useful and worthless alike, but only potentially useful ideas become conscious. This is the view summarized in Table 8.1C. He continues: Figure the future elements of our combinations as something like the hooked atoms of Epicurus. During the complete repose of the mind, these atoms are motionless, they are, so to speak, hooked to the wall…. On the other hand, during a period of apparent rest and unconscious work, certain of them are detached from the wall and put in motion. They flash in every direction through the space (I was about to say the room) where they are enclosed, as would, for example, a swarm of gnats or, if you prefer a more learned comparison, like the molecules of gas in the kinematic theory of gases. Their mutual impacts may produce new combinations. (Poincaré, 1913, p. 393) So we have hooked atoms flashing every which way in the chamber of the unconscious, until new combinations—new ideas—are produced. However, in Poincaré’s view there is still a potential problem: The number of possible combinations of ideas is still so numerous that even fast unconscious combinatorial processing would not allow us to winnow the potentially important ideas from the chaff of sterile ones. Therefore, logic demands that there be some limitation on the number of ideas entering into even unconscious combinations. According to Poincaré, the earlier conscious work, which seemed to the thinker to produce no progress, actually makes a positive contribution to creative thinking, because it serves to restrict the combinatorial process to at least those ideas that have some chance, even if it is a remote one, of producing fruitful combinations: What is the role of the preliminary conscious work? It is evidently to mobilize certain of these atoms, to unhook them from the wall and put them in swing. We think we have done no good, because we have moved these elements a thousand different ways in seeking to assemble them, and have found no satisfactory aggregate. But, after this shaking up imposed upon them by our will, these atoms do not return to their primitive rest. They freely continue their dance. Now, our will did not choose them at random; it pursued a perfectly determined aim. The mobilized atoms are therefore not any atoms whatsoever; they are those from which we might reasonably expect the desired solution. Then the mobilized atoms undergo impacts which make them enter into combinations among themselves with other atoms at rest which they struck against in their course. However it may be, the only combinations that have a chance of forming are those where at least one of the elements is one of those atoms freely chosen by our will. Now, it is evidently among these that is found what I call the good combination. (Poincaré, 1913, p. 389) The unconscious combinatorial process thus has two important characteristics. First, some of the ideas that it deals with are those that were set into motion by being considered during preliminary conscious work on the problem. This serves to focus the process at least partly on those ideas that might be potentially useful. Second, those selected ideas are used for high-speed unconscious combination with inactive of ideas, to make it possible for the still-numerous potential combinations to have a chance at formation. Thus, in Poincaré’s view, the unconscious does not do anything that conscious processing could not do if there were but time available. Criteria for a Combination’s Becoming Conscious We now have the unconscious combinatorial process at work, and sometimes it is successful: A potentially useful combination is hit upon and bursts suddenly into consciousness, where it is experienced as an illumination. This leads to the question of the criteria for determining whether an unconsciously produced combination should be examined further in consciousness. This requires some sort of judgmental process, which is another function, beyond the simple combination of ideas, carried out by the unconscious. In Poincaré’s words, It is certain that the combinations which present themselves to the mind in a sort of sudden illumination, after an unconscious working somewhat prolonged, are generally those useful and fertile combinations…. [T]he privileged unconscious phenomena, those susceptible of becoming conscious, are those which, directly or indirectly, affect most profoundly our emotional sensibility…. The useful combinations are precisely the most beautiful, I mean those best able to charm this special sensibility that all mathematicians know, but of which the profane are so ignorant as often to be tempted to smile at it. What happens then? Among the great numbers of combinations blindly formed by the subliminal self, almost all are without utility: but just for that reason they are also without effect on the esthetic sensibility. Consciousness will never know them; only certain ones are harmonious, and, consequently, at once useful and beautiful. They will be capable of touching this special sensibility of the geometer of which I have just spoken, and which, once aroused, will call our attention to them, and this give them occasion to become conscious. (1913, pp. 391–392) Thus, an idea becomes conscious when it strikes the (unconscious) sensitivity of the thinker as being “beautiful” or “harmonious.” We now have an explanation for Poincaré’s certainty that his idea on the omnibus was correct: It had already been subject to evaluation by the unconscious sensibility. This sensibility is also why we are never consciously aware of the many sterile combinations that, according to Poincaré, our unconscious must produce: Their sterility insures that they will not get past the unconscious gatekeeper. Poincaré’s Theory of Unconscious Processing in Creative Thinking: Summary Poincaré concluded on the basis of his introspections that unconscious processes were crucial in his creative process. Most important, the occurrence of illuminations made it clear to him that processing had been going on under the surface. In addition, the fact that the ideas that occurred to him were always at least potentially valuable led him to the conclusion that some sort of unconscious evaluation process was being carried out. Row A of Table 8.2 presents a summary of this view. It may not be an exaggeration to say that modern psychological theorizing about creative thinking is built solidly on the foundation of Poincaré’s theory. His view has been adapted and modified by more-recent researchers in light of new developments, but the core of those new theories remains much as Poincaré proposed. The remainder of this chapter will examine the development of modern theories of unconscious thinking in creativity, which means in essence that we will trace the influence of Poincaré’s ideas on modern psychology and examine variations on the themes that he raised. The underlying issue to deal with is how one explains the occurrence of illuminations during problem solving, which is also the agenda set by Poincaré. Table 8.2 Summary of theories of unconscious processing Theorist Postulated mechanism of unconscious thinking (A) Poincaré Random combinations of ideas; good ideas are those that combine domains previously thought unrelated. All combinations produced in unconscious are possible in conscious thinking, if the person is given enough time. Aesthetic sense determines which ideas reach consciousness. (B) Wallas Formalized Poincaré’s ideas into four stages; second stage is unconscious incubation. (C) Hadamard Poincaré’s ideas and Wallas’s stages; questionnaire to mathematicians and scientists (including Einstein). Discussed broader evidence for unconscious, such as recognition of faces and production of discourse. (D) Koestler Bisociation as the process underlying creative thinking: brings together two previously independent streams of ideas. Unconscious connections based on Freudian primary-process thought. (E) Simonton Darwinian view, based on Campbell; unconscious combinations go beyond conscious possibilities: Freudian primary-process thinking. Creative thinker’s associations organized into Mednick’s flat associative hierarchies. (F) Csikszentmihalyi Parallel processing based on associations among ideas in unconscious, versus logical and constrained order of ideas in conscious thinking. Unconscious processing produces unlikely and illogical combinations. Reports by creative individuals concerning how their thought process works provide evidence for unconscious processes. Wallas’s Stages of the Creative Process Wallas (1926) formalized Poincaré’s ideas into a well-known series of four stages of the creative process. The first stage, preparation, centers on initial conscious work on the problem, in which the thinker immerses him- or herself deeply in the problem, becoming familiar with it and attempting solutions. If this work is unsuccessful and leads to an impasse, the person then breaks off work on the problem. The unconscious, however, keeps working. This stage of unconscious work is called incubation, in an analogy to what happens inside an egg when it is warmed by a hen. The discovery of a new and potentially useful combination by the unconscious leads to the third stage, a conscious experience of illumination. That is, the incubated egg hatches. Finally, the idea that produced the illumination requires verification so that its adequacy can be determined. As we saw with Poincaré’s illuminations, this stage requires conscious thought. Wallas based his discussion on reports written by individuals of creative accomplishment, such as Poincaré and Hermann von Helmholtz, the great physicist, who also provided a report on how he carried out his work (Helmholtz, 1898). Wallas offered advice to thinkers based on these sorts of reports, emphasizing that, if one hopes to be successful, one should sometimes completely stop thinking about a problem and give the unconscious processes time to do their work. Wallas’s ideas are a straightforward elaboration of those of Poincaré, with Wallas’s original contribution being an explicit labeling of the stages that Poincaré discussed in less formal terms (see Table 8.2B). Hadamard’s Studies of Unconscious Thinking in Incubation Hadamard (1954), a mathematician who was a disciple of Poincaré, also presented a detailed description of the phenomenon of inspiration and the role of unconscious processes in creative thinking. He gathered together information from individuals in the fields of science (including Einstein, who answered questions from Hadamard concerning his thought process), mathematics, and the arts that was relevant to the general question of inspiration and provided support for Poincaré’s conclusions on the critical role of the unconscious. Hadamard also brought forth other types of information that he believed supported the role of unconscious processes in thinking. As one example, Hadamard points to so seemingly simple a phenomenon as recognizing the face of a friend. In order to recognize that face, we use many different features, or pieces of information, but we are aware of none of the complexities of the process, and we cannot describe in any way what we are doing. Thus, there is a gap between our unified conscious awareness (“There’s Sue”) and the complex recognition processes that, in Hadamard’s view, must be occurring. It therefore is necessary to assume that much unconscious processing is involved. The act of synthesis that occurs when you take the multitude of available features and from them produce the unified perception of a familiar face points to a crucial difference between conscious and unconscious processing. Consciousness is unified and singular—you are aware of only that single face—while the unconscious is manifold—multiple things can be processed at once. In modern terms, consciousness is a serial processor, while the unconscious processes information in parallel. Another example discussed by Hadamard (1954) is the production of connected speech. You are talking to a friend, say, and you produce a string of connected sentences in order to describe some event. When you produce the first sentence, asks Hadamard, where is the second sentence in what will be a coherent string of speech? “Certainly not in the field of my consciousness, which is occupied by sentence number one; and nevertheless, I do think of it, and it is ready to appear the next instant, which cannot occur if I do not think of it unconsciously” (Hadamard, 1954, p. 24). Hadamard also makes distinctions among several “layers” of nonconscious processing (pp. 24–28), based on, for example, our ability to produce a string of related sentences. He proposes that the to-be-spoken-sentences are waiting in an unconscious that is close to the surface and at the disposal of conscious processes. Table 8.2C summarizes this view. In short, Wallas made explicit the stages Poincaré had already discussed only informally. Hadamard marshaled additional evidence for the theory of unconscious processing and tried to demonstrate the wide role of such processing in other areas of cognition. Koestler’s Bisociation Theory Koestler (1964) presented an analysis of creative thinking that combined Poincaré’s analysis with Freudian theory. He analyzed many creative advances and concluded that they often involve bringing together two independent streams of associations into one idea, a process he called bisociation, in contrast with association, which involves only one stream of linked ideas. As an example, we can examine Koestler’s report of Guttenberg’s invention of the printing press with movable type. The printing press is reported to have been brought about when Guttenberg attended a wine festival, where, helped along by some of the wine, he realized that the press used to crush grapes could be used to apply type to paper. Thus, the movable-type printing press was born. Koestler, like others (e.g., Hadamard, 1954), emphasized that the root of cogitare, Latin for “to think,” is to shake together, which fits with Poincaré’s image of “hooked atoms” zooming this way and that in the chamber of the unconscious until a pair collides and becomes hooked together. However, unlike Poincaré, Koestler believed that the connections used by unconscious thinking are different from those used by conscious thinking. Koestler assumed that conscious thought processes operate through verbally based logic as well as associative connections based on experience and habit. Creative thinking, on the other hand, demands connections among ideas that go against logic and habit. In Koestler’s view, Freudian primary-process thought provides the vehicle whereby those new bonds can be forged (1964, p. 183ff.). He provides examples of how some of the primary-process mechanisms we discussed in the last chapter function in creative thinking. Use of optical puns, for example, is seen when a thinker conceives of a string of molecules as a snake, as was reported by Kekulé in describing his discovery of the ring structure of the benzene molecule (see discussion in Chapter 2). Kekulé’s analysis could also be classified as symbolization of an abstract theoretical idea in a concrete image. Einstein reported to Hadamard that his thought was almost never based on words, which supports Koestler’s claim that one must use nonverbal modes to break away from the tyranny of logic and verbal habits. Einstein also reported that an important insight concerning relativity theory came about as a result of his imagining what would happen if he were moving at the speed of light in pursuit of a light beam. This is an example of concretization, although Koestler (1964) acknowledges that in many cases primary-process thinking is not neat and clean and that often many of the different subprocesses are being used together, which may make analysis a bit difficult. Thus, Koestler elaborated Poincaré’s view by combining it with Freudian ideas on primary-process thinking. (See Table 8.2D.) So we now have two streams of theory concerning how unconscious processing functions in creative thinking: “pure” Poincaré and Poincaré plus Freud (see Figure 8.1). As we shall see, modern views are typically variations on one or the other of those streams. Campbell’s Evolutionary Theory of Creativity: Blind Variation and Selective Retention The strongest and most direct influence of Poincaré’s theorizing on modern psychology can be seen in a stream of research that began with the work of Campbell (e.g., 1960), who developed a theory of creativity based on the notion that creative ideas come about through a process of evolution analogous to the natural-selection process operating in Darwin’s theory of organic evolution. In Darwin’s theory, blind variation and selective retention determine how species evolve. That is, there is first a random change or variation in the genetic material, caused by, for example, mutation brought about by radiation. This variation is blind, in the sense that the evolutionary process has no foresight concerning which variations might be best for a given species in a given environment. Variations occur randomly; some are useful and some are not. Those that are useful are retained, since the organisms that possess them are more likely to survive and pass their genes on to the next generation. Figure 8.1 Map of historical development of theories According to Campbell (1960), similar processes determine how creative ideas come about. In Campbell’s view, three conditions are necessary for creative thinking, in the form of solution of a novel problem. (1) There must be some means of generating new ideas, or ideational variation, analogous to the occurrence of mutations in organic evolution. (2) Once new ideas are produced, those variations are subjected to a consistent selection process that retains only those that are successful, again analogous to natural selection in organic evolution. (3) The variations that have been selected must be preserved and reproduced by some mechanism so that they are available to succeeding generations. Campbell (1960) concluded that, in order for there to be truly effective variation in ideas when facing a new problem, the ideational variation must be fully blind, as is the production of mutations in evolution. He argues that when an organism is faced with a truly novel problem there is no recourse but to produce behavioral responses that are blind, that is, not only independent of the specific problem situation but also independent one from the next. The response attempts that are produced will have no correspondence to the problem being faced, in the sense of being directed toward solution of that problem. Nor will the various attempts be related to each other: Later attempts will not be “corrections” of earlier ones. There will thus be no predictability in when the solution will emerge; it will be just as likely to follow an attempt that is totally off the mark as one that is very close to correct. If one does see foresight and intelligence on the part of an organism approaching a problem, this is of course due to experience. However, argues Campbell (1960), in the history of that organism and / or of the species, there must have been a point before experience, at which totally random variations were produced. Otherwise, according to Campbell, there could not have been the true novelty in behavior that is needed to deal with situations demanding it—that is, those situations in which true creative advances are made. Campbell extensively quotes Poincaré (1913) in order to provide support for the blind aspect of ideational variation. As an example, he quotes Poincaré’s description of his sleepless night as evidence for the random way in which ideas are combined in creative thinking. Campbell’s theorizing is the place where Poincaré’s theory becomes part of the mainstream of psychological theorizing about creative thinking, although Campbell does not explicitly discuss unconscious processing. Simonton (e.g., 1988, 1995b, 1999), whom we will discuss next, explicitly incorporated unconscious processing into Campbell’s point of view. Simonton’s Chance Configuration Theory Simonton (e.g., 1988, 1995b, 1999, 2003), one of the most prolific and influential psychologists currently writing on the creative process, has carried forth Poincaré’s ideas as elaborated by Campbell, including the evolutionary analogy and its emphasis on a random component in the production of ideas (see Table 8.2E and Figure 8.1). Simonton noted that Campbell too built on a long tradition. William James, for example, also made explicit an analogy between production of new ideas and production of new organic forms in evolution through random mutations: The new conceptions, emotions, and active tendencies which evolve are originally produced in the shape of random images, fancies, accidental outbirths of spontaneous variation in the functional activity of the excessively unstable human brain, which the outer environment simply confirms, refutes, or destroys—selects, in short, just as it selects morphological and social variations due to molecular accidents of an analogous sort. (1880, p. 456) In elaborating Campbell’s view, Simonton (1988) proposed that the creative process operates on what he referred to as mental elements, which are fundamental psychological units that can be manipulated in some manner, comparable to Poincaré’s (1913, p. 393) “hooked atoms of Epicurus.” Closely following Poincaré and Campbell, Simonton proposes that those elements must be free to enter into combinations through the process of chance permutation. Those permutations are carried out in the unconscious. After combinations of elements have been formed, some selection process must be introduced, because not all combinations should be retained. Again in reasoning that closely parallels that of Poincaré, Simonton proposes that the permutations that are formed differ in stability. The greater the stability of a combination, the greater the chance that it will be selected, and the greater attention the combination will command in consciousness. Thus, Simonton’s continuum of stability is analogous to Poincaré’s notion of selection on the basis of the thinker’s unconscious aesthetic sensitivity. Confronting the question of why certain combinations are more stable than others, Simonton proposes that certain mental elements possess “intrinsic affinities” for each other. That is, there is an attraction of some sort between pairs of elements, something like a magnet’s attraction for a paper clip. Intrinsic affinity comes about because sometimes two configurations are structured so that their elements can line up one-to-one. This means that chance combination can result in the thinker’s becoming aware of an analogy between hitherto unrelated phenomena, again a point emphasized by Poincaré and others who followed him, such as Koestler (1964). As one example of this realization of a previously hidden analogy, one can recall Poincaré’s illumination that the Fuchsian functions were equivalent to the transformations of non-Euclidean geometry. This illumination occurred because those elements turned out to have one-to-one correspondence, which triggered the realization that the concepts were identical. In art, a chance combination can produce a particularly striking metaphor, which will bring together heretofore unrelated domains of experience, where a poet, for example, might discover a previously unrealized relationship between his or her lover, say, and a beautiful flower. Individual Differences in Cognition: Mednick’s Associative Hierarchies In Simonton’s view, individuals can vary along two dimensions that are relevant to their ability to produce novel ideas. First, people differ in the total number of mental elements they possess: The “genius” possesses more elements in his or her database than does the “normal” individual. Obviously an individual with more mental elements has a greater chance of producing a valuable combination. It is not enough, however, to possess a large number of mental elements; the elements must be organized in the manner that is optimal for creative production. That optimal structure depends on the associative organization among the elements in a person’s database. In analyzing the associative relationships among mental elements that distinguish the creative from the noncreative individual, Simonton (1988, 1999) bases his ideas on the theory of Mednick (1962), who published a relatively short but very influential paper on the “associative” mechanisms underlying creative thinking. Here too we can see Poincaré’s influence, as Mednick defined the creative thinking process as “the forming of associative elements into new combinations which either meet specified requirements or are in some way useful. The more mutually remote the elements of the new combination, the more creative the process of solution” (p. 221). This definition echoes Poincaré. In Mednick’s (1962) analysis, any situation that demands creative thinking can be analyzed as a set of stimulus elements, each of which has many other elements or ideas associated to it. The thinker’s analysis of the situation results in the elements of the situation evoking some subset of all those associated elements. These evoked elements, now “active” in consciousness, can combine to produce a new idea; thinking a new thought is the result of a new combination of elements. Mednick’s analysis leads to a simple conception of how individuals may differ in factors that lead to creative ideas. The basic component is the associative hierarchy, or the organization of an individual’s associative responses to a given situation. Some people, those who will not think creatively, have restricted hierarchies, in which there are one or two dominant responses. Those responses tend to occur often and quickly, and they therefore tend to block production of less-frequent responses. Such individuals tend to produce stereotyped and familiar responses to a situation, and so they are at a disadvantage when novel responses (i.e., relatively infrequent responses) are demanded. Creative individuals, in contrast, possess associative hierarchies in which a relatively large number of responses, of more or less equal probability, are available. Such individuals have a much greater likelihood of coming up with a relatively unusual response to a situation, which could result in a creative outcome. Mednick’s two types of associative hierarchies are shown in Figure 8.2. Simonton (1995b), adopting Mednick’s notion, proposes that new ideas are produced by what he calls a nonorderly “free-associative procedure” (p. 471), which he equates with Freudian primary-process thinking. He notes that “the products of this mechanism are unpredictable and uncontrollable, the associative meanderings freewheeling” (1995b, p. 471). Some Illustrations: Introspections In support of his view, Simonton presents a number of reports by creative thinkers on their creative thought process. As one example, Einstein, in the already-mentioned response to Hadamard (1954, p. 142), says that “combinatory play seems to be the essential feature in productive thought.” That statement may remind one of Poincaré’s report of his sleepless night, when “ideas arose in clouds” and collided one with another until pairs collided. This can be looked upon as a sort of combinatorial play. So Einstein and Poincaré provide examples of Simonton’s postulated free-associative thought of the individual at the highest level of creative genius. Simonton also presents William James’s (1890, p. 456) description of the thought processes in minds of the highest order, which we have already seen in Chapter 4, which also supports the same viewpoint: Figure 8.2 Mednick’s theorySource: Adapted from Mednick (1962). Instead of thoughts of concrete things patiently following one another in a beaten track of habitual suggestion, we have the most abrupt cross-cuts and transitions from one idea to another, the most rarified abstractions and discriminations, the most unheard of combination of elements, the subtlest associations of analogy; in a word, we seem suddenly introduced into a seething cauldron of ideas, where everything is fizzling and bobbling about in a state of bewildering activity, where partnerships can be joined or loosened in an instant, treadmill routine is unknown, and the unexpected seems only law. Simonton proposes that there is a continuum of problem solving. At one end is routine problem solving, such as working out a problem in long division, for which a straightforward procedure is known to everyone. At the other end are the sorts of problems involved in scientific research, for which there are no well-known procedures. In the latter situations, the free-associative process is crucial. In Simonton’s (1995b) words, For the kinds of problems on which historical creators stake their reputations, the possibilities seem endless, and the odds of attaining the solution appear nearly hopeless. At this point, problem solving becomes more nearly a random process, in the sense that the free-associative procedure must come into play. Only by falling back on this less disciplined resource can the creator arrive at insights that are genuinely profound. (pp. 472–473) Here again Simonton (1995b) makes an explicit connection between his theorizing and the ideas of Poincaré and Koestler: As Poincaré … remarked: “[The most useful ideational permutations] are those which reveal to us unsuggested kinship between other facts, long known, but wrongly believed to be strangers to one another…. [Hence,] among chosen combinations the most fertile will often be those formed of elements drawn from domains which are far apart….” So commonplace are fantastic syntheses that Koestler made them the cornerstone of his theory: bisociation…. [P]robably the only way two irrelevant realms can be brought together is by the crazy confluence of rather haphazard and whimsical trains of association…. Thus, the more offbeat is the bisociation, the greater must be the role of chance in generating it, on the average. (p. 473) In explaining how these unusual combinations of ideas can be brought about, Simonton falls back on an already-familiar premise: The combinations occur in the unconscious. However, on the basis of recent research in cognitive psychology concerning unconscious processing (e.g., Greenwald, 1992), Simonton (1995b) does not assume that unconscious processing is very sophisticated. Rather, he says, “in all likelihood, the unconscious mind is the repository of some rather primitive cognitive and affective associations that can form linkages that the conscious mind would deem preposterous” (p. 475). Thus, the unconscious provides the links that can be used by a not-very-sophisticated associative process to bring together ideas that would never have occurred to the thinker in the conscious state. It is also interesting to note that in Simonton’s view there are several circumstances in which the unconscious free-associative process can become accessible to consciousness. First, if the thought process involves vivid imagery, it will sometimes be attended to consciously, as was the case with Kekulé’s discovery of the benzene ring. Also, if the creator can somehow be in a state wherein his or her consciousness is not occupied with some other task, the workings of the unconscious might be glimpsed. An example of this is Poincaré’s sleepless night, when he was not consciously engaged in any other activities and so had the unique opportunity of observing his unconscious at work. Usually, however, when one is engaged in other activities, the unconscious operates outside of awareness. What is usually necessary for the optimal operation of the unconscious is that the individual be engaged in some mundane activity, such as walking, that requires minimal resources and attention to external stimuli. This makes available the capacity for the unconscious to do its work. An example would be Poincaré’s illumination while walking on the seaside bluff. Thus, the boundary between conscious and unconscious processing is not hard and fast; it shifts with the circumstances, granting a fortunate few access to the workings of their own minds while functioning at the highest level. Simonton’s Theory of Unconscious Processing in Creativity: Summary In Simonton’s theorizing we can see clear evidence of the continuing influence of Poincaré’s theory on modern research. The basic idea of Simonton’s theory is that while we are engaged on one task, we can still be working on another. In addition, Simonton’s theory also contains a free-associative initial component, which is based on Campbell with links to Koestler and thus can be traced back to Freud. Those connections are outlined in Table 8.2E. The notion of unconscious processing, à la Poincaré, is only one component of Simonton’s theory, in which he attempts to bring together many aspects of the creative process, the creative person, and the milieu in which that person is functioning. We will examine the broader aspects of Simonton’s theory in Chapter 11, in which several broad theories of creativity are reviewed. Csikszentmihalyi’s Theory of the Unconscious in Creative Thinking The influence of Poincaré’s theorizing in psychology can be seen also in the work of Csikszentmihalyi (1996; Csikszentmihalyi & Sawyer, 1995), who has presented an analysis of creative thinking based on interviews with almost 100 individuals who have made significant lifelong creative contributions in the arts, the sciences, technology, and business. In providing the background to his study, Csikszentmihalyi (1996) presents Wallas’s stages as the basis for his own theorizing. The second phase of the creative process is a period of incubation, during which ideas churn around below the threshold of consciousness. It is during this time that unusual connections are likely to be made. When we intend to solve a problem consciously, we process information in a linear, logical fashion. But when ideas call to each other on their own, without our leading them down a straight and narrow path, unexpected combinations may come into being. (p. 79) In considering possible explanations of what might be happening during the stage of incubation, Csikszentmihalyi rejects the Freudian primary-process view as “spectacularly implausible” (1996, pp. 100–102). He discusses as an alternative a more “cognitive view,” which assumes, as demonstrated by the passage just quoted, that unconscious thinking involves making connections among ideas based on laws of simple association. This results in essentially parallel processing, which produces what may seem to be random combinations of ideas, rather than the logical and constrained serial processing of conscious thought. Thus, the unconscious will succeed through the production of combinations of ideas that would not have been produced in conscious thinking. This is because conscious thinking is linear and logical, in Csikszentmihalyi’s view, whereas the unconscious uses associative connections that go beyond the bounds of strict logic. Csikszentmihalyi, like Simonton, moves away from Poincaré’s view and includes vestiges of the Freudian view, as filtered through Koestler, in his theory. Csikszentmihalyi does, however, echo Poincaré’s and Simonton’s notions concerning how ideas produced in the unconscious become conscious as illuminations or insights (see Table 8.2): “The insight presumably occurs when a subconscious connection between ideas fits so well that it is forced to pop out into conscious awareness, like a cork held underwater breaking out into the air after it is released” (Csikszentmihalyi, 1996, p. 104). This is analogous to Poincaré’s discussion of how combinations that are aesthetically pleasing are passed from the unconscious to consciousness. As a specific example of how creative discovery can come about, Csikszentmihalyi (1996) presents the case of Kekulé’s discovery of the structure of benzene, which was discussed in Chapter 2: The German chemist August Kekulé had the insight that the benzene molecule might be shaped like a ring after he fell asleep while watching sparks in the fireplace make circles in the air. If he had stayed awake, Kekulé would have presumably rejected as ridiculous the thought that there might be a connection between sparks and the shape of the molecule. But in the subconscious, rationality could not censor the connection, and so when he woke up he was no longer able to ignore its possibility. (p. 101) Interviewees’ Opinions on the Unconscious As new evidence to support the notion of unconscious thinking during incubation, Csikszentmihalyi (1996; Csikszentmihalyi & Sawyer, 1995) presents the opinions of many interviewees concerning the importance of unconscious processing in their creative process. Many of the individuals have no doubt that creative leaps based on unconscious processing are important in their creative work, and Csikszentmihalyi and Sawyer (1995) note the potential importance of the reports of their interviewees concerning the creative process. Whereas examples of insight in everyday life tend to be elusive and debatable, they are both more public and more convincing when they occur to scientists whose work results in Nobel Prizes or to artists and writers who enhance our lives with their creative endeavors. (p. 330) Csikszentmihalyi and Sawyer (1995, p. 331) report that their respondents described their moments of insight as part of a four-stage process, corresponding to the stages of Wallas (1926): preparation, incubation, illumination, and verification. In contrast to other analyses of unconscious processes in creativity, Csikszentmihalyi and Sawyer’s analysis emphasizes the large individual differences that can occur in incubation and illumination (1995, p. 336). Some individuals report working on a problem—including, presumably, unconscious work—for a period of years; others might start working on some problem in the morning and have an illumination in the afternoon. As one renowned example, Charles Darwin worked for many years on the theory of evolution through natural selection before achieving his breakthrough. Csikszentmihalyi and Sawyer propose that the long-time-frame versus short-time-frame insight processes are so different that they might actually represent two ends of continuum. They discuss this continuum in terms of a distinction between presented and discovered problems. In a presented problem, which usually involves a shorter time frame, the person begins to work on a problem that already exists. In a discovered problem-solving process, which may extend over long stretches of time, the problem is not one people have been dealing with before the individual in question came on the scene. In the view of Csikszentmihalyi and Sawyer (1995), great creative insights, those that result in shifts in the field, belong to this long-term-insight category. It is interesting to note that they present Darwin’s discovery as an example of a long-time-frame process. There is no doubt that Darwin’s work resulted in a paradigm shift, that is, a radical change in theorizing in biology (see discussion in the last chapter of Kuhn’s [1962] notions of revolutionary—i.e., paradigm-shifting—versus normal science), but, contrary to the analysis proposed by Csikszentmihalyi and Sawyer, Darwin solved a presented problem. That is, he was dealing with an issue—how to explain the evolution of species—that had been of interest to several generations of theorists before him, including his own grandfather (Eiseley, 1961). As noted in Chapter 5, Eiseley concluded that Lyell set the problem that Darwin and Wallace solved. Thus, the distinction between presented and discovered problems may not coincide with that between short- and long-term incubation processes. Interaction between Conscious and Unconscious Processes Concerning the details of the creative thinking process and, more specifically, the question of the interaction between conscious and nonconscious processes, Csikszentmihalyi and Sawyer (1995) note that several research traditions have suggested that unconscious processing has greater capacity than conscious processing. In these views, more than the unitary Freudian subconscious is working on the problem; the unconscious is seen as many smaller entities, each of which might be working on a different problem, so many problems might be being worked on at once. This conception raises a further question (Csikszentmihalyi & Sawyer, 1995, p. 339): If consciousness is a serial processor with limited capacity, while the unconscious is a parallel processor with much greater capacity, how can the individual coordinate them? Somehow the individual must be able to direct what Csikszentmihalyi and Sawyer call the undirectable subconscious process, so that useful insights result. Perhaps paradoxically, many of the individuals surveyed by Csikszentmihalyi and Sawyer claimed that they had indeed developed the ability to direct those unconscious processes. Furthermore, although we have all heard of situations in which insight comes about through the action of some external stimulus (the apple that supposedly fell on Newton’s head and stimulated the development of his theory of gravitation, for example), Csikszentmihalyi and Sawyer found no evidence for such occurrences in their interviews. Their respondents described their insights as “welling up from the subconscious” (1995, 343), without a specific external stimulus. This is parallel to Poincaré’s illuminations on boarding the omnibus and walking on the bluff: In both cases there was no external stimulus that triggered the illumination. Data: Content of the Interviews Let us now turn to the content of the interviews carried out by Csikszentmihalyi and Sawyer (1995). The respondents were, as mentioned, individuals who had lifetimes of creative output in many domains and are still active. Even when they were dealing with short-term presented problems—for which, in Csikszentmihalyi and Sawyer’s view, incubation is less of a factor than in longer-term discovered problems—many of the respondents “structured their days to include a period of solitary idle time that follows a period of hard work. Without this time, they would never have their best ideas” (p. 347). The respondents were thus paving the way for the operation of their unconscious. Keeping the mind idle sometimes involves simple repetitive physical activity, as it did for the respondent who said, “Generally, the really high ideas come to me when I’m gardening” (p. 348). All the respondents reported that small-scale insights came during this set-aside time, which an economist describes: “We have this little cabin…. We have a little ritual in the morning; [my wife] takes a short bath, and then I have a 40-minute bath and do some exercise” (p. 348). Some recipients carry notebooks with them to take advantage of the ideas when they come. A banker reported: “It often happens [i.e., illuminations occur] when I’m sitting around a hotel room; I’m on a trip and nothing’s going on and I sit and think. Or I’m sitting on a beach … and I find myself writing myself notes” (p. 348). A similar process concerning an insight on a larger scale was described by a physicist-mathematician who was trying to bring together two seemingly incompatible theoretical approaches to quantum mechanics that had been proposed by two other physicists, Richard Feynman and Julian Schwinger. I spent 6 months working very hard, to understand both of them clearly, … and at the end of 6 months, I went off on a vacation, took the Greyhound bus to California…. [A]fter two weeks in California, where I wasn’t doing any work, just sightseeing, I got on the Greyhound bus to come back to Princeton and suddenly, in the middle of the night, when we were going through Kansas, the whole thing sort of suddenly became crystal clear, so that was sort of the big revelation for me, the eureka experience. (Csikszentmihalyi & Sawyer, 1995, p. 359) Concerning the period of incubation, those people who reported long-time-frame insights often reported that the insights occurred during an extended period of time away from work, such as a vacation or sabbatical (Csikszentmihalyi & Sawyer, 1995, p. 352). One respondent, as we have seen from his quotation, carried around a notebook; he had made several contributions that changed the structure of the banking industry. He reported that his major creative insights had always come while he was on vacation, often while on the beach (Csikszentmihalyi & Sawyer, 1995, p. 354). One of his insights is even called the “memo from the beach”; in it, he outlined the structure of the first consumer banking enterprise, in 1974: I was on a vacation, and I started out saying, “I’m sitting on the beach thinking about the business,” and it went on for 30 pages. And it turned out to be the blueprint. I didn’t sit down and say, “I’m gonna write a blueprint;” I said, “I’m sitting on the beach thinking,” and I sort of thought through the business in a systematic way … and I shared it with my colleagues. (p. 354) A similar process occurred for this individual during a second insight, which led to a corporate reorganization. This one occurred while he sat on a bench in Florence: In September I had been kind of tired … and I had gone to Italy for a week, just gotten away…. I’d get up early in the morning, and I’d wander around, and I sat on a park bench, between 7 in the morning and noon…. I had a notebook, and I wrote myself long essays on what was going on and what I was worried about. (Csikszentmihalyi & Sawyer, 1995, p. 354) In further support of the importance of the subconscious in creative thinking, many of those interviewed by Csikszentmihalyi and Sawyer (1995) had on their own developed theories of the creative process that emphasized the importance of “off time.” These veterans of creative-thinking campaigns had no doubt that the process requires a period of incubation to allow the subconscious to carry out its work. As one example, the physicist-mathematician who earlier described his middle-of-the-night eureka experience in Kansas began his interview by saying: “I’m fooling around not doing anything, which probably means that is a creative period” (p. 352). Concerning the elaboration stage of the process, in which an insight that has burst into the light must be evaluated and shaped to fit reality, almost all the respondents reported that such work was necessary. One respondent, however, an economist, environmentalist, and poet, reported having illuminations that did not need further elaboration; that is, things came to him complete. He reported, “The last 9 days I was there [in California], I dictated the book a chapter a day and revised it very little actually. I’d been thinking about it for over a year, and it just came through. It was like having an intellectual orgasm, it just comes [laughs]” (Csikszentmihalyi & Sawyer, 1995, p. 356). On the basis of their research, Csikszentmihalyi and Sawyer (1995, p. 358) concluded that the creative leap or insight is the result of a period of incubation, during which information is processed in parallel at an unconscious level. If incubation is successful, an insight will occur, and there is usually then a period of conscious evaluation and elaboration of that illumination. (See Table 8.2F.) This view is compatible with those already discussed. Csikszentmihalyi and Sawyer (1995) build the case for stages of creative thinking from evidence from their interviews. First of all, those high-creative individuals have developed their own theories of creative thinking that rely on the opportunity for unconscious processing. In addition, the interviews provide evidence of the role of unconscious processing in producing illuminations. Unconscious Thinking in Creativity: Conclusions We have seen that there is large-scale agreement among theorists that the stages postulated by Wallas (1926) many years ago, on the basis of Poincaré’s reports, are a useful description of the process of creative thinking. More specifically, as summarized in Table 8.2, theorists are in general agreement that during incubation unconscious processing is taking place. They also believe that this unconscious processing makes possible connections among ideas that are beyond the capabilities of ordinary conscious thinking. In Poincaré’s original conception, those new connections came about because the unconscious was able to work more quickly than was conscious thinking, but he assumed that there were no differences in the associative links that served to guide thought. Most modern theorists, however, have moved closer to something resembling the Freudian view in various ways—even if they do not accept all of it (e.g., Csikszentmihalyi’s [1996] explicit rejection of Freudian ideas)—and assume that the unconscious is able to make connections among ideas that conscious processing cannot bring about. There is little disagreement among modern theorists concerning the belief that unconscious processing is able to bring together combinations of ideas that are beyond those available to conscious thinking. Again, Table 8.2 summarizes this work, and the historical connections among the various theories discussed so far in this chapter are outlined in Figure 8.1. The Question of Subjective Reports It was noted in Chapter 2 that one can question the value of self-reports as evidence for psychological theory. In that context, it is important to note that all the theories we have discussed so far have been built on such reports. We began with Poincaré, whose theorizing was based on evidence from his own experiences. Poincaré also analyzed his creative process logically and concluded, for more than one reason, that unconscious processes must have been operating. First, he experienced illuminations, wherein the solution of a problem appeared suddenly when he was not working on it. It is difficult to understand how such experiences could occur unless he had been working on the problem at a nonconscious level. Also, he believed that what he called “invention”—creative thinking, in our terms—depends on combinations of ideas. However, he was not aware of the innumerable combinations of ideas that, in his view, must have been occurring. Therefore, he decided, those combinations were being carried out by processes of which he was not aware. Furthermore, also based on the notion of creativity as combination, it must have been the case that many sterile combinations were formed during Poincaré’s creative process (1913). He was aware only of potentially valuable combinations, however, which meant that something must have been blocking the occurrence of those useless combinations in his conscious awareness. That blocking must have occurred at an unconscious level. Finally, the combinations that he did become aware of possessed beauty in mathematical terms. This feature led him to the conclusion that all the combinations were being subjected to a process of judgment. He was aware of no such process, but since, based on his logical analysis of the situation, it must have been happening, he concluded that it must have been carried out on an unconscious level. Wallas (1926) provided little new evidence beyond that of Poincaré, and Hadamard (1954) added reports from Einstein and others, but we still have only self-reports. Hadamard also concluded on the basis of his logical analysis of such psychological phenomena as facial recognition and speech production that unconscious processes must have occurred. Koestler (1964) used self-reports as support for his theory, which incorporated ideas from Freud and Poincaré. Campbell (1960) used Poincaré’s reports as evidence to support his theoretical assumption that there must be a blind-variation process in creativity since otherwise there could be no production of new ideas. Campbell did not provide any new data; his adoption of Poincaré’s ideas was driven by logic (his perceived necessity of explaining the production of truly new ideas in creative thinking) and by the requirements of his theory (the need to incorporate a process analogous to the blind variation that occurs in organic evolution). Simonton (1988) accepted the arguments of Poincaré and Campbell; he added Koestler’s ideas, thereby incorporating aspects of Freud’s view. Simonton’s adoption of Mednick’s (1962) theorizing is also not based on any new data. Indeed, as we have seen, Mednick also cited Poincaré. Lastly, Csikszentmihalyi and Sawyer (1995) also built their theory on that of Poincaré, as filtered through Wallas’s stages. For data they rely on subjective reports from 100 highly creative individuals. Given the potential weakness of subjective reports as support for psychological theory, let us now turn to a consideration of experimental studies that have been designed to provide support for the notion of unconscious thinking in creativity. Laboratory Investigations of Incubation and Illumination A number of empirical studies have attempted in several ways to provide support for the notion of unconscious processing in creative thinking. First, attempts have been made to provide verification for the stages of creative thinking postulated by Wallas (1926), and second, researchers have tried to find evidence specifically for the occurrence of illumination, which would provide at least indirect evidence for incubation and unconscious processing. Patrick’s Studies of Stages of Creative Thinking C. Patrick (1935, 1937) carried out two often-cited early investigations directed at the question of stages in creative thinking. In the first study, accomplished poets, chosen because their work had already been published, were given a picture and asked to write a poem in response to it. In the second study, artists whose work had been exhibited throughout the world were asked to draw a picture in response to poetry of Milton. In these studies, performance of the creative participants was compared with that of a control group, who were matched in age, intelligence, racial background, and sex, but who had not exhibited creative achievement. Patrick saw each individual in a session at his or her home, during which he or she carried out the creative task while thinking aloud; Patrick took down in shorthand everything the participant said. The task was carried out in a single session. After the poem was written or the picture completed, Patrick interviewed the creative participants concerning their usual methods of working and whether they typically went through periods of incubation that were followed by illumination, among other questions. Although the creative productions were in response to stimuli provided by the experimenter, most of the poets and artists reported that the methods they used in Patrick’s study were similar to those they usually used. The poets and artists took about 20 minutes on average to complete their projects, as did their respective control groups. In order to determine if these sessions provided evidence for Wallas’s (1926) hypothesized stages, Patrick divided the single work session into quarters. In the first quarter, she found that her participants made the most shifts from idea to idea, which she took as evidence of preparation. During the second quarter of the session, she noted whether the topic of the poem involved an idea that had been raised during the first quarter but dropped as the participant turned to other ideas. Recurrence of a previously rejected idea, which was most frequent during the second quarter, was taken as evidence of incubation’s having occurred. Similarly, in the study of artists, the theme for the painting often involved an idea that had been discussed earlier, then put aside, and then returned to at a later time; again, Patrick took this as evidence of incubation. During the third quarter, the poem or the picture was given general theme or shape, and typically this stayed the same throughout work on the project. This was evidence of illumination, as an idea for the work took shape as the result of incubation. Finally, in the fourth quarter of the working session Patrick found the highest frequency of revisions of the work, which she took as evidence of verification. Thus, in Patrick’s view, her work provided support for the existence of the stages proposed by Wallas. Although C. Patrick’s (1935, 1937) work has often been discussed in the context of Wallas’s (1926) stages, the context in which she herself placed it, questions can be raised about whether this research is relevant to the general theory of stages of creative thinking and the more specific theories of unconscious incubation. First of all, the fact that the work was carried out during a single session means that the individual obviously was thinking about the creative project the whole time. In the classic examples of incubation, those provided by Poincaré (1913), he first thought about the problem intensely and then stopped working. Poincaré also reported that during that break he stopped thinking about the problem. The solution to the problem then came to him in another context, in which he claimed that he had not been thinking at all about the problem. Thus, Patrick’s data from a single session are not relevant to questions about creative-thinking situations in which a person reaches an impasse and then works—or, as the case may be, does not work—on some problem over a period of time. One piece of information from C. Patrick’s (1935, 1937) studies that might be relevant to the general issue of Wallas’s stages, as well as to the specific question of the occurrence of unconscious incubation, comes from the questionnaires she administered to the poets and artists after they produced their creative works. Most of the members of both groups reported that incubation occurred in their work; that is, they thought about an idea, put it aside, and then had the idea come to mind as the theme of a work. Many of the participants reported, however, that they thought about the idea from time to time during the period of incubation—that is, during the time in which unconscious processing was assumed to be occurring. Needless to say, if poets and artists think about ideas off and on when they are not formally working on some project, any purported illumination could be the result of these bouts of conscious thinking. If this finding has any validity and generality—and it should be noted that the finding comes from unverified responses to Patrick’s questionnaire, because she provided no other evidence to support the claims of her participants—it would indicate that, in some cases at least, it is not necessary to postulate the occurrence of unconscious processes as the basis for illumination. The conscious processing could be doing all the work. This issue will be discussed further shortly. For the reasons we have discussed, C. Patrick’s studies of poets and artists shed little light on stages of creative thinking in general or on unconscious processing in particular. The design of her studies made it impossible for unconscious processing and incubation to occur. Eindhoven and Vinacke’s Study of Stages in Creative Thinking One problem with C. Patrick’s (1935, 1937) studies, as just mentioned, is that the painters and poets were given only a single session in which to carry out the task of creative production, so a true incubation period was not possible. Eindhoven and Vinacke (1952) extended Patrick’s research to try to overcome this difficulty. As in Patrick’s study, painters were asked to create a picture in response to a poem, but they were given more time to work on the project: They could take up to four sessions over a week’s time. Of the 13 artists studied, almost all took at least two lab visits to complete the painting, and more than half returned to the lab for a third visit. Similarly, of the 14 nonartists who participated as a control group, almost three quarters returned for three visits. In addition, although most of the artists completed their visits during 1 week, several took 2 weeks, and for one artist several months elapsed between his two visits. No time limit was placed on the visits, but an hour was suggested. During the time between visits, the participants were asked to keep a diary in a pocket notebook of any experiences and ideas relevant to the project. Any sketches done outside the laboratory were done on paper provided by the researchers and were brought in to the next session. The first important result from the Eindhoven and Vinacke (1952) study is the finding that multiple sessions were typically used in order to carry out the project, which indicates that C. Patrick’s (1935, 1937) limiting her participants to a single session in her early studies might have resulted in important data being missed. Concerning the specific activities carried out by the participants, results indicated that overall the artists produced many more sketches than did the nonartists. Furthermore, the artists tended to produce new sketches only during the first session and then concentrated on using one of those sketches as the basis for their final work. The nonartists kept producing new sketches over the several sessions. The early sketches produced by the artists tended to be small, as might be expected of a preliminary work. The nonartists produced larger works throughout all the sessions. Eindhoven and Vinacke (1952) also examined the content of the sketches and found that the motif of the first sketch—the salient theme or feature, such as an abstract composition versus a landscape, or the presence or absence of human forms or artifacts—tended to be repeated in later sketches, indicating that the artists had decided early in the process on the motif and then stayed with it. More specific aspects of the subject matter—the specific landscape, say—were developed over the series of sketches. Examining the first 5 minutes’ work on a picture and comparing it with the last 5 minutes’, the researchers found, perhaps not unexpectedly, that both artists and nonartists spent the early time laying out general aspects of the picture—overall composition, outlines of principle objects—and spent later time filling in detail. Those findings parallel those from the study of Picasso’s creation of Guernica presented earlier (see Chapter 1). Picasso too decided on the subject matter very early in the process and developed the specifics of the painting over a series of sketches. The participants’ diaries, although of potential importance, were kept by only seven participants (of whom five were artists), so little information could be gleaned from them. However, poststudy interviews indicated that a majority of participants thought about the project outside the laboratory. That finding supports Patrick’s (1937) finding that artists reported that they often thought about a project when they were away from their studios. In summarizing their results, Eindhoven and Vinacke (1952) draw several important conclusions. First, the typical path taken by their participants involved a significant period of time during which the final product gradually evolved, as the individual increasingly focused on a particular sketch and from it produced the final product. Concerning Wallas’s (1926) four stages of the creative process, Eindhoven and Vinacke raise a number of points. They note first that Wallas’s description of the stages of the creative process might lead one to expect them to be universal; that is, in all situations demanding creative thinking, all people would behave in the same way, and that behavior would be in accord with the stages. However, the artists in this study behaved in general very differently from the nonartists, which indicates that, irrespective of the general issue of stages in creative thinking, the same process is not occurring in all individuals faced with a situation that requires creative thought. As regards the specifics of Wallas’s (1920) postulated stages in the creative process, Eindhoven and Vinacke concluded that the stages described by Poincaré (1913) and Wallas (1926) could not be isolated as separable entities in their results. The hypothesized stages actually blended into each other in complex ways, so it was difficult if not impossible even to talk of stages. Furthermore, when Eindhoven and Vinacke discuss the stages, it becomes obvious that the term stage itself is not particularly apt as a description of what was happening in their study. As an example, they discuss the illumination stage and comment that it was very difficult in their data to find a distinct point where an idea suddenly came into consciousness (although one artist reported that an idea for a sketch came suddenly one night). Eindhoven and Vinacke propose that one might redefine illumination as a process leading to a definite idea or reorganization of previous ideas. They would then find illumination in their study, and one could then talk of a series of illuminations, as the first sketch is transformed into final form. In response to that proposal, however, one can ask if we would be studying illumination in the sense proposed originally by Poincaré and those who followed him, like Wallas (1926) and Hadamard (1954). If the notion of illumination has to be changed that much in order to find evidence for it, it might be better to simply give up the idea of stages and perhaps even to relinquish the notion of illumination. It might be more informative to simply examine how a product is transformed over time without worrying about specific stages in the process, since there might not be any stages to be found. As we can see, the studies by C. Patrick (1935, 1937) and Eindhoven and Vinacke (1952) have not provided strong support for the general theory of stages in creative thinking or the specific question of whether unconscious processing occurs in creative thinking. We can now turn to a number of studies that have attempted to demonstrate the occurrence of incubation in the laboratory. These latter studies have used experimental situations designed to capture the essential features of those situations in which incubation and illumination have been reported by creative thinkers. Attempts to Demonstrate Incubation in the Experimental Laboratory The basic experimental design that can allow one to demonstrate the occurrence of incubation in the laboratory is shown in Table 8.3A. First, the individual tries to solve a challenging problem (period a in Table 8.3A). If the session ends with the target problem unsolved, then the study of incubation can begin. In an analogy to Poincaré’s reports, the initial work should end with the person reaching an impasse, but that is not always ascertained. The individuals who do not solve the problem are assigned to one of two conditions: The incubation group has time away from the problem (intervening period b in Table 8.3A), while the control group keeps working on the problem. After the incubation period, the target problem is again presented (period a′ in Table 8.3A) and the incubation group goes back to work on it. The two groups spend the same total amount of time working on the problem (a + a′, as shown in Table 8.3A). Performance of individuals given a break is compared with that of individuals who work continuously on the target problem; if taking a break results in better performance, it can be taken as evidence for incubation having occurred when the person was not working on the target problem. This conclusion is based on the condition that the participants did not think about the target problem during the break. Not thinking about the problem during the break is comparable to Poincaré’s (1913) report of his incubation experience: He reached an impasse and then left his work at home and went traveling. He says that his travels made him forget his mathematical work. If people do not stop working on the problem during the break, then the question of incubation occurring becomes moot, because the so-called incubation group simply kept on working during the time that they were supposed to take a break from it. Obviously, that continuous work will account for any increase in performance compared to the continuous-work control group, because when you add in the additional time during the break, the incubation group worked longer. As we shall see, the condition that the incubation group not work on the problem during the break is not always met. Table 8.3 Laboratory studies of incubation An important question in laboratory studies of incubation is what happens during the period away from the problem, when any incubation effects would presumably occur. In most studies the intervening or incubation period is structured in one of two basic ways (see Table 8.3B). In some studies, the participant leaves the laboratory for the incubation period, with the understanding that the target problem will be returned to later. Those participants are asked to carry out their normal activities during the time away from the lab and not to think consciously about the unsolved problem during that time. In other studies, the participant stays in the laboratory and another task is presented, and the participant works on that task for the duration of the incubation period. This second task can be called a distractor task, since it is designed to distract the person from thinking about the unfinished target problem during the break. Distractor tasks are used in experiments because experimenters are concerned that people might either intentionally or unintentionally think back to the unsolved problem during the break. As just noted, such conscious thinking about the problem would make invalid any attempt to demonstrate incubation. As shown in Table 8.3C, the distractor activities in incubation studies are usually of two sorts. In some studies the participants work on a distractor activity that has nothing at all to do with the problem they had been working on. As an example, a person working on a mathematical problem might be asked to rate pictures of faces for attractiveness during the incubation period. In other studies, the person works continuously on a series of problems of the same sort, so the incubation period is filled with other problems. The person presumably cannot think about the unsolved Problem1 during the incubation period because he or she is working on Problem2, Problem3, and so forth. When Problem1 is presented for the second time (see Problem1 in Table 8.3C2), one could say that the participant is returning to it after an incubation period. It is important to emphasize that there are two interrelated issues tied together in these studies. The interest in the possible facilitative effects of taking a break stems, of course, from Poincaré’s (1913) conclusion that unconscious processing was occurring when he was away from his desk and not thinking about his work. However, as just noted, the question of whether taking a break facilitates problem solving is independent of whether the facilitation comes about because of unconscious processing. Let us say that we carry out a study such as that presented in Table 8.3A (ignoring at this point how the incubation period is structured), and we find that taking a break facilitates solving the target problem. There then arises the question of the mechanism whereby this facilitation comes about. Finding that taking a break from work facilitates problem solving is a necessary but not sufficient step in demonstrating that unconscious processing occurs during creative thinking. There might be other reasons why taking a break helps, reasons that have nothing to do with unconscious processing. In short, the first question is whether taking a break facilitates problem solving; if this question is answered in the affirmative, then we can turn to the question of the mechanism that brings about the facilitation. So before we theorize about the mechanism underlying illumination, let us examine research that has been designed to demonstrate the facilitative effect of taking a break during problem solving. There have been a number of laboratory studies that have used the design outlined in Table 8.3A–8.3C, and the results have been mixed even in demonstrating that taking a break facilitates solving the target problem. One difficulty is that many of those studies suffer from flaws in their designs, which makes it difficult to draw conclusions as to whether incubation occurred. As one example, we can examine an early study of incubation by C. Patrick (1938). As noted in the last section, Patrick’s two studies of Wallas’s (1926) stages in problem solving were problematic (1935, 1937) because she tested each person in a single session, which made it impossible to investigate all the stages. In a later study, C. Patrick (1938) asked individuals to propose scientific methods to investigate the effects of heredity and the environment on humans. The control group worked continuously on the problem, while the incubation group was given a diary and was told to return 2 to 3 weeks later with proposals. They were told to use the diary to record thoughts or ideas that arose during the break from the problem. A basic difficulty with this study becomes apparent immediately: The incubation group did not really take a break from the problem but simply worked on the problem in whatever manner they wanted to over a long period of time. The experimenter exerted no control over what the participants did during the incubation interval, and one might conclude that the instructions to write ideas in the diary encouraged the people to think about the problem straight through. So this study can tell us nothing about incubation. Several studies that do not suffer from such basic flaws are summarized in Table 8.3D; for reviews of additional studies see Olton and Johnson (1976) and Dodds, Ward, and Smith (in press). A study by Olton and Johnson (1976) using the Farm problem (see Table 8.3E) examined possible effects of several different sorts of activities during the incubation period (see Table 8.3D). Those activities ranged from unstructured free time to listening to set-breaking instructions (designed to help people break out of any ruts they were in) to being moved to a room in which there were pictures that contained geometric forms that were analogous to the solution. None of the conditions resulted in incubation: All the groups solved the problem at the same rate as the group that worked continuously. In addition, Olton and Johnson derived a method to score people’s progress toward the solution, so they had a second measure beyond simple solution of the problem. This progress score also showed no evidence of incubation. This study was also an attempt to replicate one by Dreistadt (1969), which had reported incubation using the Farm problem, so the inconsistent results across the two studies point to difficulties in demonstrating incubation. In another attempt to demonstrate incubation in the laboratory, Olton (1979) used a specially selected group of participants—experienced chess players—as well as a problem that was selected by a chess master to be challenging: working out the final moves in a chess game. That problem was of the sort that those participants on their own would typically spend much time on. In that way this study mimics what happened with Poincaré (1913): Presumably he was working on problems of intrinsic interest. Olton gave the chess problem to the participants and instructed them to work on it for about 1 hour; if they had not solved it by the end of that time, they were to break off work for about 2 hours. So the break occurred when the participants felt ready for it. During the break the participants could do anything they wanted, as long as they did not think about chess. The problem seemed to have been highly involving, with some people actually voluntarily cutting short the break to continue working on it. A control group worked on the problem without a break. Although it seems that overall Olton (1979) did a relatively good job of producing in the laboratory a microcosm of situations in which illuminations have been reported, the results of the study were disappointing as far as demonstrating incubation: 50 percent of both the incubation and control groups solved the problem. One participant in the incubation condition reported an Aha! experience during the break, à la Poincaré, but overall the break did not help the incubation group. As Olton notes, “We simply didn’t find incubation” (p. 17) Browne and Cruse (1988, Experiment 2) also used the Farm problem and examined three distractor activities. One group of participants drew geometric forms during the incubation period, in the hope that one of those shapes might cue the solution to the problem, based on analogy. This was the analogical hint condition. A second group was given relaxation instructions and listened to music during incubation, and the last group carried out difficult mental work (memorizing a passage) during the incubation period, to try to make it very difficult for them to think about the problem during the break. After the experiment, the participants filled out a questionnaire asking them what they had done during the incubation period, to gather evidence for people thinking about the problem consciously when the experimenters thought they were supposed to be taking a break from the problem. Incubation effects were found for the analogical hint and relaxation conditions; there were no effects for the participants carrying out difficult mental work. Browne and Cruse concluded that the positive effects were due to the participants’ working on the problem when they were supposed to be not thinking about it, and the participants’ responses to the questionnaire supported that conclusion: The people in the two conditions that resulted in incubation effects reported that they had thought about the problem during the break. Browne and Cruse concluded that perhaps all results previously attributed to some underlying mechanism working during the incubation period were in actuality the result of people consciously thinking about the problem without the experimenter’s realizing it. A. S. Patrick (1986) carried out a study using verbal insight problems such as those shown in Table 8.3D). One group worked on the problems but cycled through them for 2 minutes at a time, as outlined in Table 8.3C2, until they had worked a total of 8 minutes on a problem or had solved it. A second incubation group also cycled through the five incubation problems for a maximum of 8 minutes, but before beginning the next cycle of problems they also spent an additional 5 minutes talking with the experimenter about activities not relevant to the experiment. So their incubation time between each attempt on a given problem was longer than that for the incubation group that just cycled directly through the problems. The final group cycled through the problems and carried out a difficult mental rotation task for 5 minutes at the end of each cycle before returning to the problems. This study seems to meet the criteria for a well-designed study: Participants reached impasse before stopping work on each problem; several problems were used to test for incubation; and experimental controls made it difficult for people to think about the problems they were not working on during the incubation periods. However, once again strong incubation effects were not demonstrated. A. S. Patrick (1986) looked separately at high-ability versus low-ability participants, whose ability was measured by their performance on the initial series of problems, and he found that only one condition—problems plus mental rotation with high-ability participants—produced a significant increase in performance compared with continuous work. In none of the other remaining five comparisons (two ability level and three incubation conditions) was the incubation group significantly better in performance than the continuous-work group. So this study produced at best weak evidence for incubation. The final incubation study presented in Table 8.3D was carried out by Segal (2004). He used the geometry problem presented in Table 8.3E, and he began the incubation period only after the participant had indicated that he or she had reached an impasse. The incubation periods were either 4 or 12 minutes and the distractor activities were easy (reading newspapers) or difficult (working on a crossword puzzle). Incubation was found with the hard task for both incubation-period durations; the effect was less strong for the easy task, with only the short interval resulting in increased performance compared with the continuous-work group. So again we have a study that found some evidence for an incubation effect, but the lack of an overall strong pattern of incubation raises questions about the robustness of the phenomenon. Laboratory Studies of Incubation: Conclusions The results from laboratory studies of incubation are at most mixed: It has been extremely difficult to demonstrate consistently within a single study even that taking a break facilitates problem solving, so one never gets to the question of whether unconscious processing might be the cause of that facilitation. One might dismiss the negative findings in Table 8.3D because the laboratory studies are sterile and are far removed from the real-life situations in which incubation results in illumination. However, it should be noted that at least one study specifically designed to meet those objections—the chess study of Olton and Johnson (1976)—produced negative results. In addition, if incubation is a phenomenon of critical importance in problem solving, as for example Poincaré’s (1913) reports might lead one to believe, then one would expect that it would be easy to demonstrate it in the laboratory. If, for the sake of discussion, we take the evidence from the laboratory studies at face value, it leads to the conclusion that there is no strong and consistent laboratory evidence to support the notion of incubation during creative thinking, and therefore there is also no strong and consistent laboratory evidence to support the related notion of unconscious processing. In this context it is interesting to note again that Browne and Cruse (1988; see Table 8.3D) interpreted their results as indicating that what has been called incubation—the facilitation of problem solving by taking a break from the problem, perhaps brought about by something like unconscious processing during that time period—comes about because people think about the problem consciously during the break. Olton (1979) raised a similar possibility, that what he called “creative worrying” (actively thinking about a problem when one was supposed to be not thinking about it) was the basis for reports of incubation. I will consider that possibility further at the end of the chapter. There is also one other point to note concerning the lack of evidence for incubation and illumination in the studies just discussed. Those studies provided a very weak test of Poincaré’s notions, and yet they still found little support for unconscious processing. It will be recalled that Poincaré (1913) reported that his illuminations came to him as Aha! experiences when he was away from his study—boarding an omnibus in Coutances or walking on the bluff at the seaside. The studies summarized in Table 8.3D almost never report such Aha! experiences; all they measure is whether taking a break and not thinking about a problem helps people when the problem is presented again. This is not quite a parallel to what happened in Poincaré’s case: He solved the problem when he was not working on it. Thus, even in the studies in Table 8.3D that found that taking a break facilitated problem solving, one can still raise the question of whether such results provide evidence for Poincaré’s incubation, since nothing like spontaneous illuminations occurred (e.g., Olton and Johnson, 1976, reported only one person as having an Aha! experience when away from the problem, which is the sort of evidence that would have provided strong support for Poincaré’s theorizing). Thus, the weak results in the studies just discussed are doubly troublesome for the question of unconscious processing during creative thinking: They provide support neither for Poincaré’s strong claims about unconscious processing—illumination occurring while away from the problem—nor for the much weaker claim that taking a break should facilitate solution. These negative conclusions leave us with only the anecdotal reports of illumination as evidence of unconscious processing in Poincaré’s strict sense. It now becomes of interest to look again at the anecdotal evidence for the presence of unconscious processing and illumination. Evidence for Incubation and Illumination: A Critique We have just seen that there is at best weak empirical support from the laboratory for the occurrence of unconscious processing in creative thinking. We have also seen that over the last 100 years many students of creative thinking—including, over the last 40 years, many cognitive psychologists—have built theories of creative thinking in which unconscious processing is given pride of place. The postulation of unconscious processing in creative thinking is based mainly on anecdotal reports. Those reports now have an especially heavy burden to bear: They must convince us that unconscious processing is at the core of creative thinking. The evidence that was discussed earlier in this chapter, when each of the various theories was outlined, is presented in Table 8.4. Questions about Poincaré’s Self-Reports Poincaré’s reports, as we have seen, have surfaced again and again over the years in the discussion of unconscious processing in creative thinking (Table 8.4A). We have already considered some problems with self-reports, including Poincaré’s, as data for a scientific theory. There are several additional questions that can be raised about how much confidence we should have in those particular reports. First of all, Poincaré’s public discussion of his discoveries was presented some 30 years after the events in question occurred. Furthermore, two of Poincaré’s reports—stepping on the omnibus and walking on the bluff—dealt with Aha! experiences, that is, events of an extremely brief duration. It seems very difficult to believe that Poincaré’s recounting of those experiences could be accurate after so long a period of time; it is hard enough to recount Aha! experiences immediately after they occur. In Chapter 5, we discussed Perkins’s (1981) study of insight in solution of the Antique Coin problem, and we examined how Perkins had designed his study to ensure that his reports would be of maximum accuracy. We noted that he made sure that the reports came as soon as possible after solution and that the participants were given some practice in the reporting task. Fleck and Weisberg (2004, 2006) and Chrysikou and Weisberg (2005), who also collected verbal protocols, also spent time designing their studies and training their participants (for detailed discussion of conditions for collecting protocols, see Ericsson and Simon, 1996). In this context, it must be noted that Poincaré’s reports were made by an individual who, although no doubt of the highest level of intellect, had no specific experience in the behavioral sciences. If one assumes that training in the behavioral sciences prepares one even a little for observing and reporting behavior, even one’s own—and the discussion of Perkins’s studies indicates that training is necessary to ensure that behavioral data are accurate—then Poincaré’s reports, coming from a mathematician-scientist with no training as a behavioral scientist, again come into question. Table 8.4 Critical review of evidence for unconscious processing in creative thinking Theorist Evidence Critique (A) Poincaré (1) Boarding the omnibus; on the bluff (2) Sleepless night (3) Logical analysis as the basis for concluding unconscious processes occurred: infinitude of possible combinations (1) Self-report; original Aha! experiences took seconds to unfold; reported 25+ years later. (2) Self-report; also, Poincaré was conscious (3) No empirical evidence; question of infinite regress; homunculus. (B) Wallas and Hadamard (1) Poincaré’s stages (2) Self-reports (1) See above. (2) See above. (C) Koestler (1) Analysis of creative advances: Unlikely combinations of ideas (e.g., Guttenberg) (2) Kekulé (3) Einstein’s report of “combinatorial play” (1) Self-reports. (2) Self-reports—see above. (3) Einstein: Self-reports. (a) Combinatorial play: conscious. (b) Riding on light beam: conscious. (D) Simonton (1) Logical analysis: numbers of combinations; unlikely combinations needed for important problems (2) Poincaré; Kekulé (3) Other opinions: William James, etc. (1) No new evidence presented—reasoning based on Poincaré and Campbell. (2) Self reports—see above. (3) Not empirical evidence. (E) Csikszentmihalyi; Csikszentmihalyi and Sawyer (1) Reports of interviewees: Nobel Prize winners, etc., are “more public and more convincing” (2) Insights while gardening: “Generally, the really high ideas come to me when I’m gardening.” (3) Physicist on the bus: “In the middle of the night, when we were going through Kansas, the whole thing sort of suddenly became crystal clear.” (4) Banker’s “memo from the beach”: “‘I’m sitting on the beach thinking about the business,’ and it went on for 30 pages.” (1) Why are reports more convincing when reported by Nobel Prize winners? Still subjective reports. (2) Gardening: Is she not thinking? (3) He was not consciously thinking? (4) He was thinking and writing memos, and was surprised at length and the result—how is that evidence for unconscious processes? It should also be noted that the first of Poincaré’s (1913) reports, which describes the sleepless night when he observed his ideas colliding and combining, actually says nothing at all about the unconscious, since Poincaré was conscious during that episode. As we have seen, Poincaré concluded that, since he did not take an active part in directing the thought process, he must have been observing his unconscious at work, but there is absolutely no evidence for that conclusion. The objective evidence (assuming that we can take his report at face value) is that he was conscious. So the hooked atoms of Epicurus, which Poincaré and others (e.g., Simonton, 1995) refer to either directly or indirectly in building theories of unconscious processing in creative thinking, occurred in conscious thinking, and so seem irrelevant to theorizing about the unconscious. Ultimately, Poincaré’s evidence consists merely of several first-person anecdotal reports that were made public many years after the events in question. In most areas that cognitive psychologists study they do not use subjective reports or anecdotes as the basis for theorizing. Poincaré’s Logical Analysis Poincaré (1913) also used logical analysis to support the proposal that unconscious processing served creative thinking. He only became conscious of potentially useful combinations, but he believed that numerous combinations must have been formed somewhere during his work. Therefore, he concluded that those useless combinations that he never became aware of were formed by unconscious processing and did not possess the aesthetic qualities needed to be passed on the consciousness. However, this logical analysis, while perhaps reasonable on its face, is based on several assumptions, about the thinking process in general and creative thinking in particular, that are not the only possible assumptions. The critical assumption underlying Poincaré’s conclusion is that thinking works by making combinations of ideas, one after the other, until something potentially useful captures attention. Poincaré’s theory of thinking is therefore an example of what we have called a bottom-up view of the thought processes (see Chapter 3), in the sense that it works by combining the basic elements or ideas, without any planning; planning is called top-down processing, because plans are based on ideas generated from the top. We are also familiar with the bottom-up / top-down distinction from the discussion in Chapter 5 of heuristic-based restructuring in insight. Contrary to Poincaré’s bottom-up conception, it is possible to conceive of creative thinking as working in a more top-down manner, so that the goal the thinker is working toward plays a role in determining how ideas are combined. In this view, many combinations would never be tried out because they would not ever be considered. As an example, consider Watson and Crick’s discovery of the double-helix structure of DNA (see Chapter 1). They did not start by combining available ideas without planning. Their thinking shows evidence of planning, in several ways. First, and most important, they used Pauling’s success with alpha-keratin as the basis for the general helical orientation that they took toward their work. Therefore, all the ideas they considered were germane in some way or other to the general idea of helices. This was not, however, because unconscious processing produced those combinations and passed them along to consciousness, as Poincaré assumed. Rather, it was because of the top-down nature of the process: Their conscious decision to deal with helical structures meant that, from the very beginning of the process, only certain ideas would be considered, which radically limited the breadth of (conscious) search. Other case studies in Chapters 1 and 5 also support that conclusion. Poincaré’s conclusion that unconscious processing must have been involved in his thinking was based on the fact that he was never conscious of the innumerable ideas that were totally irrelevant to his problems. His critical assumption was that he combined all those ideas somewhere, so if he was never conscious of them, they must have been combined in unconscious processing. However, that assumption may simply be incorrect. If there are not innumerable hidden combinations of ideas that need to be explained, we do not need unconscious processing. As we can see, Poincaré’s evidence of unconscious processing is of little substance when examined with a critical eye. His empirical reports were given years after the fact and even turn out to contradict the notion of unconscious processing, since at least one—the sleepless night—was based on conscious processing, and his logical analysis supporting the necessity for postulating unconscious processing is based on assumptions that can be questioned. The views of Wallas and Hadamard (Table 8.4B) are built directly on that of Poincaré, and therefore they add nothing new to the discussion. Wallas’s (1926) stages are built on Poincaré’s reports and on those of others, but those other reports are no more rigorous and reliable than are those of Poincaré. Hadamard (1954) attempted to go further, but the evidence he provided was of the same sort as Poincaré: self-reports, although from a wider range of individuals. The numerous individuals whose reports of illuminations fill the literature (e.g., Ghiselin, 1952; see also Csikszentmihalyi & Sawyer, 1995) may believe that their creative thought involves unconscious thinking, but that belief, per se, does not help in the scientific study of the phenomenon. These criticisms are also relevant to Koestler (1964; Table 8.4C), all of whose theorizing about the unconscious is based on self-reports. Modern Views of Unconscious Processing We have examined modern theorists’ advocacy of unconscious processes, and we have found that those theories are based on data that are not different in kind from those available to Poincaré. Simonton (1995; Table 8.4D; see also Miller, 1996), for example, simply assumes that Poincaré’s analysis is valid on its face, without providing further evidence to justify that assumption. Csikszentmihalyi and Sawyer (1995) present many subjective reports as data (Table 8.4F), but those are simply additional after-the-fact self-reports by thinkers and do not add anything new to what we learned from Poincaré. As noted earlier, Csikszentmihalyi and Sawyer (1995, p. 330) claim that self-reports concerning unconscious processing are of value when they come from Nobel Prize-winning scientists or from artists and writers of renown. However, I would raise the same objections to those reports as have already been raised to those of Poincaré. Do we have any independent evidence for what those reports claim? And why would we want to believe the report of a Nobel Prize-winning chemist, say, about her unconscious processing? What does she know about such things, other than that sometimes she may suddenly have an idea occur to her and she cannot trace where it came from? That chemist’s credentials qua chemist are of little value here, since the study of creative thinking is a domain totally unrelated to her area of expertise. As was noted earlier in the discussion of Poincaré’s lack of training as a behavioral scientist, it is not a given that anyone can produce and collect behavioral data that are valid as a basis for the construction of theories. Furthermore, questions can be raised concerning the interpretation of several of the reports presented by Csikszentmihalyi and Sawyer (1995). One respondent quoted earlier talked about ideas coming to her while she is gardening (p. 348). That report has nothing to do with unconscious processing: She is gardening, but that does not preclude her thinking about some other things at the same time. Gardening is exactly the kind of mindless physical activity that allows one to do one thing while thinking about something else. Similarly, the physicist-mathematician who was trying to synthesize the work of Feynman and Schwinger described how, on the bus during the night in Kansas, he was able to unify those two seemingly incompatible approaches to quantum mechanics. Here is the crucial part of the passage quoted earlier: “[S]uddenly, in the middle of the night, when we were going through Kansas, the whole thing sort of suddenly became crystal clear, so that was sort of the big revelation for me, the eureka experience” (Csikszentmihalyi & Sawyer, 1995, p. 359). When I read that passage, it seems to me that he was conscious the whole time, thinking about the problem, when the solution suddenly came to him. If my interpretation is correct, this report also has nothing to do with the unconscious. Finally, consider the banker’s “memo from the beach,” in which he outlined the structure of the first consumer banking enterprise (Csikszentmihalyi & Sawyer, 1995, p. 354). He reported: “I didn’t sit down and say, ‘I’m gonna write a blueprint;’ I said, ‘I’m sitting on the beach thinking,’ and I sort of thought through the business in a systematic way … and I shared it with my colleagues” (p. 354). Again, this report seems to have nothing to do with unconscious processing. He was thinking about the business, and he wrote a memo. Perhaps that memo was more detailed and elaborate than he had expected; however, that has no relevance to the issue of unconscious processing. He simply wrote a memo, which is a conscious act. I do not wish to belabor the point further with an analysis of each quoted example. It is important to emphasize, however, that those reports are being brought forth as evidence for unconscious processing, so they should be subject to the same scrutiny that is afforded any scientific evidence. When examined in this way, the modern evidence for unconscious processing is no stronger than that brought forth by Poincaré a century ago. It follows that the empirical support for the idea of unconscious processing in creative thinking consists of a number of anecdotal reports, several of which have been repeated so frequently over the years that many investigators now assume that the case is closed. Anecdotal reports, however, are not adequate grounds on which to build a scientific theory of creative thinking. I will therefore conclude that, as the evidence stands, unconscious processing cannot explain the occurrence of illuminations. Illumination without Unconscious Processing? Assuming that the self-reports of illuminations in the literature are accurate, we still have a phenomenon to explain. Accordingly, I now turn to explanations of illuminations that do not postulate unconscious processes. Selective Forgetting Many years ago, Woodworth (1938) formulated an explanation of facilitation in problem solving brought about by taking a break that did not rely in any way on the notion of unconscious processing. He argued that, since the problem had not been solved as a result of the initial attempts, those attempts were in the wrong direction. If we assume that some dynamic cognitive factors had played a role in determining the initial incorrect directions taken by the thinker, then with the passage of time those factors might be replaced by others. In other words, assume that when the problem is first presented the individual makes some decisions about how the instructions are to be interpreted and what sorts of solution attempts should be made. It is not unreasonable to assume that the factors or cues that influence those decisions might change over time, because when the person returns to the problem after a break, he or she is no longer the same person as before the break. Therefore, the direction taken by the thinker when the problem is re-presented might be different, but only because the intervening time allowed those misdirecting factors to become less important; one could say that the misdirecting factors were selectively forgotten during the interval away from the problem (see also Simon, 1986). According to this view, nothing positive, of either a conscious or unconscious nature, occurs during the interval; the interval merely serves to allow a change in at least some of the cues eliciting the mode of attacking the problem. This view has more recently been revived by Smith (1995; see also Segal, 2004, for a slightly different view), who proposed that initial problem-solving attempts may result in thinkers’ falling into mental ruts, unsuccessful approaches to a problem that interfere with thinking of new ones. According to Smith, the cues eliciting those unsuccessful approaches must be forgotten before anything different can be thought of with regard to the problem. Time is needed for this forgetting to occur, and that is the purpose served by the incubation period. Smith tested this hypothesis using an experimental design like that shown in Table 8.5. People were first exposed to word problems and presented with cues to the solution. Sometimes the cues were designed to be helpful (see Table 8.5A), but in the more interesting condition (Table 8.5B) the cue was designed to interfere with solution to the problem. Smith hypothesized that those misdirecting cues would have to be forgotten before solution could occur. After attempting and failing to solve the problems with the misdirecting cues, people were given breaks of various lengths away from the problems. After the breaks, the unsolved problems were presented again, and the participants were asked to try to solve them again, and they were also asked to try to recall the cue that had originally been presented with each problem. The results (see Table 8.5C) indicated that the chances of solving the problems increased as the break was lengthened. Most important, Smith also found that the chances of recalling the misdirecting cues decreased as the break lengthened, and that solving the problem was inversely related to the chances of recalling the misdirecting cue. That is, the highest solution rates occurred when the original misdirecting cue could not be recalled. The results thus support the hypothesis that getting out of mental ruts depends on changing the way the problem is analyzed, and that in turn depends on the cues used to access information used to deal with the problem. Table 8.5 Smith’s study of incubation Source: Smith and Blankenship (1989). Selective Forgetting and Illumination: Critique One problem with the selective-forgetting explanation of illumination and incubation is that it requires that the problem be reconsidered at a later time in order for the new cues to work. However, consider once again the situation that initiated this stream of work, Poincaré’s (1913) illuminations on the omnibus and on the bluff by the seaside. In those situations, at least as he reports them, the problems were not re-presented to him; he did not sit down again at his desk for another session of work. Rather, the problems spontaneously came to mind. In Smith’s version of the selective-forgetting hypothesis, there seems to be no mechanism that would allow the occurrence of spontaneous solutions to a problem. Therefore, the selective-forgetting hypothesis is at best incomplete: It can explain how breaking away from a problem can result in more efficient solution when the problem is encountered at a later time. It does not, however, seem able to deal with the phenomenon of sudden illumination, in which there is no re-presentation of the problem. In those cases, solution seems to be spontaneous, and the selective-forgetting view has no place for such a phenomenon. Failure Indices: The Opportunistic-Assimilation Model Seifert and colleagues (Seifert, Meyer, Davidson, Patalano, & Yaniv, 1995) proposed a theory, also based on memory, that attempts to deal with the spontaneous nature of illumination after breaking away from a problem. They propose a mechanism whereby seemingly spontaneous solutions to problems can be explained even when the problem has not been re-presented. In their view, which has been adopted by others (e.g., Simonton, 2003), when an impasse in problem solving leads to breaking off work, there occurs storage of the unsolved problem in memory. That is, the person would remember that he or she had not solved such and such a problem. However, according to Seifert and colleagues, the problem is also stored with a “failure index,” which makes it a unique type of memory. Those failure indices specify the general type of information needed to solve the problem. One could say that the problem is stored as containing a gap, along with a general description of the type of information that might fill that gap. When the individual encounters an environmental or mental event that matches the information needed for a given problem, that problem is retrieved, and the thinker experiences an Aha! moment. That is, Seifert and colleagues assume that a specific environmental event can retrieve a problem from memory, thereby initiating or cueing solution. Seifert and colleagues called their view an “opportunistic-assimilation” model of illumination and insight: The individual assimilates relevant environmental events that he or she happens to run across, so he or she is opportunistic in taking advantage of what the environment presents. Seifert and colleagues (1995) developed a laboratory analogue of what they believe occurs in the real world under circumstances that other investigators have labeled incubation and illumination. The experiment consisted of three phases. In phase 1 (see Table 8.6A) the participants were given a set of general-information questions to answer (those were the problems to be solved). An example is “What is a nautical instrument used to measure the position of a ship?” (Answer: sextant.) Approximately one third of the time, the participants were not able to answer the question and abandoned their attempts. After all the problems were presented, the experiment moved to phase 2, which involved a word-recognition task. A series of verbal stimuli was presented, and for each the participant judged whether it was a word. Examples are mantiness (no) and lacuna (yes). As far as the participants knew, there was no connection between phase 2 and phase 1; however, some phase 2 words were correct answers to phase 1 problems. The next day, the participants returned for phase 3, which again involved a set of information questions, some of which were new and some of which were repeated from phase 1. In addition, as just noted, for some of the old and some new questions, the answers had been exposed during phase 2, although of course the participants were not told of this. Table 8.6 Study by Seifert et al. (1995) and its relation to Poincaré’s reports The results of the study are shown in Table 8.6B, and they indicate that presentation of the target words during phase 2 facilitated solution of the old problems when those problems were re-presented during phase 3. Furthermore, the old questions without answers during phase 2 were not answered more than the new questions introduced in phase 3, which means that there was no incubation effect without exposure of the target word during the break (i.e., during phase 2). Based on these results, Seifert and colleagues (1995) concluded that incubation does not play a role in problem solving. Rather, a break may help problem solving because the course of daily events may lead to an accidental encounter with an external object or event that is especially relevant to solving the original problem (p. 87). Illumination as the Result of Spontaneous Retrieval of a Problem: Critique of the Failure Index Hypothesis Seifert and colleagues (1995) concluded that the results in Table 8.6B supported their failure index explanation of illumination. However, it should be recalled that their model is designed to explain Poincaré’s (1913) original report, which described spontaneous retrieval of the solved problem when one had not been thinking about it. Their model assumes that the availability of a cue in the environment can result in spontaneous retrieval and solution of a previously blocked problem. Their experimental situation and results, on the other hand, provide evidence only that re-presentation of an unsolved problem (in phase 3 of the experiment, after the phase 2 word-recognition task) can retrieve a recently seen word (from that phase 2 word-recognition task) that solves it. That is, it seems on close analysis that the results are not relevant to the failure index model or to Poincaré’s reports of his experiences. The structure of Poincaré’s experiences, as he reported them, and the structure of the model of Seifert and colleagues are outlined in Table 8.6C, and the lack of correspondence between the experiment and Poincaré’s reported experiences and the model can be seen. In my interpretation of Seifert and colleagues’ (1995) model, presentation of the solution words during the word recognition task (phase 2) should have spontaneously retrieved the blocked problems at that time, which would have been a demonstration of an analogue of Poincaré’s illumination experiences under controlled laboratory conditions. Presentation of the relevant word during the word-recognition part of the experiment meets the definition of an external event that should have retrieved the relevant problem. Exposure to those words facilitated solution when the problems were re-presented, which means that at least some of those words were effective cues for some problems. However, it would seem to follow from Seifert and colleagues’ model that those cue words should have spontaneously retrieved the relevant problems during the word-recognition phase of the experiment (phase 2), but they did not. The failure of spontaneous solutions to occur during the phase 2 word-recognition task in the study by Seifert and colleagues (1995) seems to contradict their model. The experimental data indicate that a recent encounter with information relevant to a problem can assist solution when the problem is presented later, but they do not support the view that Poincaré’s illuminations were due to environmental cues that provided information critical to the retrieval and “spontaneous” solution of the problems he was working on. A number of other studies have also investigated the predictions of the opportunistic-assimilation view, and overall the results have not supported them. Dodds and colleagues (Dodds, Smith, & Ward, 2002) used three-word verbal insight problems (see Table 6.5B, p. 299) to examine the influence of environmental cues on solution of a previously unsolved problem. Participants first worked through a series of 20 problems, with 30 seconds given for each. The problems had been pretested, so the experimenters knew that most people would not be able to solve most of them. The participants were then told that because the problems were difficult, later in the session they would get a second chance to work on any problems they had not solved. The participants then worked through a series of activities, and before beginning them, half the participants received instructions that information from those activities might help them on the word problems when the problems were presented again. The series of activities began with several problems of various sorts that were not related to the word problems. There was then a set of anagram-like tasks in which the participant was to make three words from the letters of each of 20 words on a list. The stimulus words in that anagram list were either the answers to the earlier problems, words semantically related to the answers (e.g., the word oatmeal was related to the answer “cookies”), or unrelated words. After the anagram task, the initial list of 20 verbal problems was presented again. Based on the opportunistic-assimilation model, one would expect that the presence of the answers and the semantically related words during the session—immediately before the participants attempted to solve the critical problems the second time—would facilitate solution on the second try, but that was not found. Over a series of experiments, Dodds and colleagues (2002) concluded that the presentation of those solution words and related words was only helpful when the participants had been warned that the activities would contain information relevant to the problems. The mere presence of the solution word shortly before the problem was considered again did not facilitate solution. Dodds and colleagues concluded that their results did not support the opportunistic-assimilation view (Seifert et al., 1995), which would seem to predict that the recent presence of the relevant words should facilitate solution. In addition, as noted earlier, Seifert and colleagues (1995) proposed their model to account for Aha! experiences such as those reported by Poincaré (1913). As noted in my earlier critique of the experiment by Seifert and colleagues, it stands to reason that the appearance of the solution word during the anagram task in the study by Dodds and colleagues (2002) should have brought about Aha! experiences in their participants. However, not only did that not happen, but, as we just saw, the presence of the solution words did not even facilitate solution when the problems were presented shortly thereafter. This study supports the negative conclusions I derived earlier from critical examination of the study by Seifert and colleagues. A study by Anolli and colleagues (Anolli, Antonietti, Crisafulli, & Cantoia, 2001) found similar results. In this study, a problem similar to the Radiation problem (Table 4.1, p. 156) was used as the target. Participants worked through a series of pages in a booklet, and the target problem was on the last page. The earlier pages contained a series of activities, including several filler stories that were not related to the target problem. In addition, on some pages there was a question that asked about some element of a story on an earlier page, so the participant had to remember something about a story presented earlier. Most important, on one of those pages was a situation that was analogous to the target problem (relevant base situation, comparable to the General problem and the Radiation problem; see Table 4.1). That base provided information that was potentially relevant to solution of the target problem. The target problem was presented on page 7 of the booklet given to participants in the control group, whose group simply started to work on it. Participants in the other groups saw the target on page 8 of their booklets, with different information on page 7. The Reminder group saw on page 7 a question they were to answer about the base situation presented earlier, but nothing was said about the potential relevance of the base to the upcoming target. The Hint group read on page 7 a hint: The base information could suggest a solution to the upcoming target; there was no specific question or information about the base on the page. The Reminder + Hint condition answered the question about the base situation and was also told that that information was relevant to the target. The target was then presented to those groups, and the question of interest was which of those conditions, if any, resulted in facilitation of solution compared to the control group. Anolli and colleagues (2001) found that solution of the target problem was facilitated only when participants heard the hint that the base information was relevant. Simply activating the base information by asking the participant a question about it—the reminder condition—had no effect on solution rates. This result would seem to contradict the opportunistic-assimilation model, which would predict that there would be facilitation by the person’s thinking about that base information just before attempting the target problem. Furthermore, one might also predict that presentation of the target problem might result in spontaneous retrieval of the base (an Aha! experience), because they are structurally related. However, Anolli and colleagues reported no occurrences of such phenomena, which probably means that they did not occur. So again we see a critical difference between what Poincaré (1913) reported and what happened in an experiment. In contrast to the studies just reviewed, Christensen and Schunn (2005) presented evidence that at first glance supports the opportunistic-assimilation prediction of spontaneous retrieval of an unsolved problem by an environmental cue during an incubation period. Cues for earlier unsolved problems were presented as the participant worked through a series of problems. If the person did not solve the third problem in the series, for example, a solved problem that was analogous to that problem was presented, but not until after several intervening problems. So the unsolved problem and the cue were separated in time. As an example, if the person had failed to solve the Radiation problem, then the General problem with its solution would be presented as a cue (see Table 4.1). When the cue problem was presented, the participant’s task was to stop solving the other problems and rate the difficulty of that new problem. That task was presumably not relevant to the problem-solving task being carried out. The participant then returned to solving the problems in the series. No mention was made of the relation of the cue problem to the unsolved problem. No restrictions were placed on the participants, so they could turn back to any unsolved problems at any time, including after a cue had been presented. The question of interest was whether presentation of the cue problem would cause the participant to return to the unsolved problem to which it was related. That is, would the presentation of the cue result in the retrieval of the unsolved problem to which it was analogous? The results of the study supported the prediction from the opportunisticassimilation theory: Presentation of a cue that was analogous to an unsolved problem resulted in retrieval of that problem and its being solved. There were control cues presented that were not analogous to any of the unsolved problems, and those cues had no effect on performance. These results seem to provide a laboratory analogue of Poincaré’s (1913) reports of his illuminations: The person broke away from the problem and then an environmental event cued its solution. So perhaps that was what happened with Poincaré: When he was away from his work—boarding the omnibus or walking on the bluff—some cue in the environment provided him with the information he needed to make his leap, which he mistakenly attributed to unconscious processing. However, there is one critical difference between Christensen and Schunn’s (2005) experimental situation and the situations in which Poincaré (1913) experienced his illuminations. Poincaré’s spontaneous retrievals—his illuminations—occurred in contexts very different from those in which he had been working on the problems: that is, on the omnibus or on the bluff above the beach. In the Christensen and Schunn study, the cues were presented by the experimenter in the same situation as the unsolved problems. It is true that the cue was not a problem to be solved—but it was a problem and its solution—and it is also true that nothing was said about a possible connection between the cue and any of the problems—but the cue was presented in the middle of the experimental session. That co-occurrence information was something that Poincaré did not have. In order to replicate the critical condition of Poincaré’s experiences as he reported them, it is necessary to present the cues in a context different from that in which the problems were presented, to see if the cues, per se, can retrieve the unsolved problems. As discussed in Chapter 4, Spencer and Weisberg (1986) showed that context had a strong effect on whether spontaneous analogical transfer occurred during problem solving: When the General problem and the Radiation problem were presented in different contexts, there was no transfer from one to the other. Thus, the Christensen and Schunn results, while interesting and potentially important, do not provide unequivocal support for the opportunistic-assimilation view of Seifert and colleagues (1995). Models of Illumination: Conclusions A number of models have been developed to explain the occurrence of illuminations in Poincaré’s creative endeavors. However, neither of the models we have considered, the updated selective-forgetting view of Smith (1995) and the failure index view of Seifert and colleagues (1995), seems to be able to account for the phenomena in question. The selective-forgetting view can explain why taking a break might facilitate solving a problem when the person returns to the problem, but it cannot explain how and why a person might suddenly and spontaneously experience the solution to a previously blocked problem. The failure index view explains such a phenomenon by assuming that the occurrence of an environmental event relevant to the solution of a previously blocked target problem can serve as a cue to retrieval and solution of the problem. A laboratory study designed by Seifert and colleagues to provide an analogue to Poincaré’s experiences, however, did not produce the predicted results. The presentation of relevant stimuli during a break did facilitate solution when the blocked target problems were re-presented at a later time. However, initial presentation of those relevant external stimuli, in the absence of the blocked problems, did not, contrary to the failure index view, result in spontaneous retrieval of the solved problems. Such a finding would have been analogous to what happened to Poincaré. So we must look elsewhere—beyond the unconscious and beyond memory retrieval by events in the environment—for an explanation of illumination in problem solving. Other studies (e.g., Anolli et al., 2001; Dodds et al., 2002) have not supported the view of Seifert and colleagues. Creative Worrying? Conscious Thinking During a Break Since we are still left with the problem of explaining illuminations—assuming that Poincaré’s reports are worth considering—let us look at one more hypothesis. Olton (1979) has proposed that much of what has been attributed to unconscious incubation can be explained by assuming that thinkers never really stop thinking about their problems, especially significant problems. They constantly return to them, if only for short periods of thought. Olton calls this process “creative worrying.” We have discussed a closely related view already in this chapter (e.g., Browne & Cruse, 1988). If creative worrying occurs, one would not have to even consider unconscious thought as a possible explanation for illumination in problem solving: Any progress a person made on a previously blocked problem would be explainable as a result of conscious thought. Some evidence to support this view was presented briefly earlier, where we saw that C. Patrick (1935) and Eindhoven and Vinacke (1952) reported that participants in their studies said that they had been thinking about their problems when they were away from them (see also Browne & Cruse, 1988). The creative-worrying explanation does not seem to be acceptable as it stands, however: After a bout of conscious work on a problem, we are usually able to report that work, as we saw in reviewing the findings from C. Patrick (1935, 1937), Eindhoven and Vinacke (1952), and Browne and Cruse (1988). However, Poincaré (1913) and many other individuals have provided anecdotal reports that indicate that they carried out no conscious work on their problems before solutions presented themselves. Therefore, if one wishes to conclude, as I have done earlier, that no unconscious processing occurred at those times, then the conscious processing to which I am turning for a possible explanation of illuminations must have been different from ordinary, garden-variety reportable conscious thinking. In the next section, I propose an extension of the notion of creative worrying that may be able to account for illumination without assuming unconscious processing. The discussion that follows is totally speculative and, at this time, without any support whatsoever. I present these speculations because, based on the discussion in this chapter, at this time we have no viable candidates for an explanation of illumination effects. Degrees of Creative Worrying: Brief Conscious Interludes I suggest that a thinker can to different degrees “worry creatively” about a problem (i.e., consciously think about a problem). The ordinary occurrence of creative worrying would entail a relatively extended bout of conscious work, which would be apparent to the thinker and reportable to others, at least for a short period of time after it occurred. It might also be possible to think very fleetingly about a problem, however, without a conscious decision or the internal “announcement” of an intention to do so. For example, there are times when I briefly anticipate my wife’s coming home at the end of the day as I am working in the kitchen. It seems to me that I do this by the merest shadow of an idea flitting across consciousness, which leaves as quickly as it comes. If this idea could be projected outward so that others could see it, they would not know what it means, but I do. I could for a very brief period of time report that fleeting thought to others, but not for very long, as it quickly fades away. On the other hand, other internal experiences of mine, like those visual or verbal images involved in explicit conscious work on a problem, would be much more likely to be understood by others if they could be projected publicly, and these would be remembered for a much longer time. In a related vein, when one is returning to a problem that one has been working on for a while, it is my experience that one does not have to say “Now I’m thinking about that problem again” and wait to see if anything happens. One simply has what one can call, for want of better words, a feeling of thinking about something; if I were asked, I could say “I’ve been thinking about a difficulty I am having in Chapter 8,” even though I said nothing of that sort to myself and did not have clear images that could simply be translated into such a sentence. Let us assume, then, that it is possible to think about a problem extremely briefly. In one of those brief conscious interludes, something relevant might come to mind or might not. If not, then one might not be able to report that one had even been thinking about the problem, except perhaps immediately after the interlude, and perhaps not even then. If the problem were solved during one of these brief conscious interludes, one might not know where the solution had come from, and one might not be able to report that one had been consciously thinking about the problem. The occurrence of the solution to a previously blocked problem would surely be attention-getting, so that the brief conscious interlude that brought it about would quickly be forgotten and become unreportable. As an example of the kind of event that might be involved in those brief conscious interludes, I just had an experience that might be analogous. I typed the word “previously” in a sentence and then went back to correct what I thought was a mistake. I thought, presumably on the basis of feedback from my fingers, that I had omitted the i. However, on looking at the word, I realized that there had not been a mistake; I had typed the word correctly. One possible explanation for my feeling that I had made a mistake is that I had typed the string of letters so quickly that I missed the awareness of typing all of them. My typing was a conscious act, and I believe that, if I had been interrupted just after typing the i and had been asked what I had just done, I could have reported it. A second or two later, however, I could not recall typing it. This might tempt one to say the letter had been typed unconsciously, but it really was not unconscious; it was just done very quickly and was easily lost from memory because of the speed at which it was done and interference from the other letters typed at about the same time. The person is conscious, but the output happens so quickly that one’s conscious postresponding report is limited. As I indicated, I have just presented complete speculation in the last several paragraphs. I have no data available to support what one could (with charity, perhaps) call hypotheses concerning how thinking might work. Obviously, speculations are not sufficient support for hypotheses. The only reason that I have even ventured to present those ideas is that we seem to have hit a dead end in our attempts to explain the phenomena reported by Poincaré and others. If we assume that, at least in outline, those reports are accurate and illuminations do occur, then we must try to develop theories that can explain them. This section is an attempt to do that, although it may fall far short of adequacy. Incubation, Illumination, and the Unconscious: Conclusions This chapter has taken us on a long journey, covering more than 100 years. We have seen that modern theories of the unconscious in creativity have not advanced very far beyond the seminal work of Poincaré (1913). His reports of the circumstances in which he made several of his mathematical discoveries, and his logical analysis of the processes that must have been involved, led him to the belief that unconscious processes led to his experiences of illumination, the sudden appearance in consciousness of the solution to a problem that had previously been abandoned at an impasse. Much modern theorizing is built directly on Poincaré’s ideas and logic, and the data called on to support modern theorizing are often Poincaré’s self-reports and / or comparable reports of others. However, analysis of theories of creativity based on unconscious processing has uncovered little strong support. Self-reports, even those from individuals of Poincaré’s renown, are not acceptable as data in support of theory, and theories of unconscious processing in creativity are supported by no hard data. It might therefore be time to move away from those theories or at least to make a concerted attempt to develop new experimental methods with the potential of providing data relevant to the question of unconscious processing. Such an endeavor is not encouraged by the consistent negative findings arising from studies that have attempted to demonstrate incubation in the laboratory (see Table 8.3D). If we do move away from the unconscious as an important factor in creative thinking, we are still faced with the phenomena of incubation and illumination, assuming that we can accept Poincaré’s (1913) reports. Unfortunately, the discussion in this chapter indicates that an alternative explanation of incubation and illumination is not yet available. We examined two memory-based explanations of incubation and illumination, the selective-forgetting view (Smith, 1995) and the failure-index or opportunistic-assimilation view (Seifert et al., 1995), but those explanations were not able to explain all the phenomena connected with incubation and illumination. At the end of the chapter, I presented some speculations concerning how one might modify Olton’s (1979) “creative worrying” hypothesis to account for Poincaré’s reports. There are no data relevant to that proposal, so at this point our conclusion is that we are still waiting for some closure on an issue that has been of interest to students of creativity for 100 years. However, such a situation should be looked upon less as a failure than as an opportunity for the students reading this book: There are opportunities for creative individuals to develop new theories and methods to test them.