Dean Keith Simonton
Many creative geniuses attain reputations of almost mythical proportions.
Apparently gifted with a special insight or intuitive power, creative ideas are supposed to pop into their heads in flashes of inspiration. In earlier times, this almost mystical process would be attributed to divine interven- tion, as is evident in the Greek doctrine of the Muses. Even during the Renaissance the greatest artist of the period would be called the “divine Michelangelo,” an artistic genius who, according to his biographer Vasari, was sent to earth by God to bless the world with his talent. Centuries later, when such religious attributions were no longer fashionable, creative genius would become linked with madness. This linkage became espe- cially prominent during the Romantic Era of19th-century Europe. A well- known illustration is to be found in the Preface toKublai Khanin which the poet Coleridge claimed to have conceived his poem in an opium-induced stupor. Eventually psychiatrists, psychoanalysts, and even psychologists were joining the chorus, associating creative genius with insanity and even criminality. The consensus seemed to be that creative geniuses were not like normal human beings. They had thought processes and personalities that set them apart from other members of the species.
Yet at the beginning of the20th century a shift was taking place. Cre- ative genius was not so special after all. This change first was apparent in attitudes toward scientific creativity. According to the emerging discipline known as the philosophy of science, scientific discovery was the product of the scientific method, a system of thinking identified with hypotheticod- eductive reasoning or some other analytical technique. All scientists who acquire this approach can engage in scientific creativity, regardless of their personal qualities or gifts. For instance, the philosopher Ortega y Gasset (1932/1957) maintained that “it is necessary to insist upon this extraor- dinary but undeniable fact: experimental science has progressed thanks in great part to the work of men astoundingly mediocre, and even less than mediocre. That is to say, modern science, the root and symbol of our 43
actual civilization, finds a place for the intellectually commonplace man and allows him to work therein with success” (pp.110–111).
This statement implies that creative genius has become largely irrelevant in modern science. This seemingly extreme claim has received endorse- ment in a whole school of research devoted to the psychology of science (Simonton,2003a). Inspired by Newell and Simon’s classic1972book on Human Problem Solving, this school argues that scientific creativity is noth- ing more than problem solving, where a problem consists
of an initial state, a goal state, and a set of permissible transformations from one state to another (called “operators”) that, when executed in a correct sequence, result in a solution path from the initial state to the goal state, via a series of intermediate states and subgoals. Operators have constraints that must be satisfied before they can be applied. The set of states, operators, goals, and constraints is called a “problem space,” and the problem-solving process can be conceptualized as a search for a path that links the initial state to the goal state. (Klahr,2000, p.23)
Creativity is thus conceived as a straightforward, logical process. To be “creative” a scientist needs only to master the knowledge and skills belonging to a chosen discipline (e.g., the “operators”) and then apply this expertise to the problems that define that discipline. To be sure, this exper- tise is not acquired overnight. Indeed, according to the10-year rule, world- class creativity is not possible without first devoting a decade to extensive study and practice (Ericsson, 1996; Hayes,1989). Nevertheless, expertise acquisition is so straightforward that even computers can be programmed to display scientific creativity of the first order. This is the lesson of so- calleddiscovery programsthat are designed to replicate the achievements of great scientists by applying logical analyses to empirical data (Kulkarni
& Simon, 1988; Langley, Simon, Bradshaw, & Zythow, 1987; Shrager &
Langley,1990). For instance, the program BACON has used inductive rea- soning to rediscover Kepler’s third law of planetary motion (Bradshaw, Langley, & Simon,1983).
Although the discussion has concentrated on scientific creativity, the same viewpoint has been extended to artistic creativity as well (Weisberg, 1992). To display exceptional creativity as a painter, sculptor, poet, nov- elist, playwright, or composer requires only the mastery of the knowl- edge and techniques appropriate to the domain. Not surprisingly, then, computer programs have also been written that purport to exhibit artis- tic creativity in its diverse forms (Boden, 1991). For example, the pro- gram EMI (Experiments in Musical Intelligence, pronounced “Emmy”) is able to write music in the style of any composer after first exposing the computer to music representative of that composer (Hofstadter,2002).
In other words, once the program assimilates the expertise implicit in the compositions, it can generate new compositions that are difficult to distinguish from those created by the original human minds. Moreover,
EMI seems to accomplish this creative imitation by the application by a straightforward set of compositional rules. The process appears to be supremely logical.
I believe that this perspective on creativity has overstated the case for both knowledge and reason. In particular, creative genius is not equiva- lent to exceptional domain-specific expertise or logic. On the contrary, the most notable scientific and artistic achievements emerge out of a far more complex process. I make the case for this increased complexity two ways.
First, I examine the role that expertise plays in the development of emi- nent creators. Second, I scrutinize the place that logic has in the creation of notable works in the arts and sciences. Both of these analyses are based on historiometrics, that is, the application of objective and quantitative methods to biographical information about creative geniuses of historic importance (Simonton,1999b,2003b).
knowledge
To understand how knowledge affects exceptional creativity, we must rec- ognize that it impinges on creativity in different ways, depending on the stage of development. The first stage entails the acquisition of the nec- essary domain-specific expertise, whereas the second stage involves the manifestation of that expertise in overt creative products.
Expertise Acquisition: Creative Potential
During childhood and adolescence an individual is supposedly acquiring the knowledge and skills necessary to manifest adulthood creativity. In other words, the future genius must develop the requisite creative poten- tial. Moreover, some have argued that the amount of creative potential is a positive function of the amount of domain-specific expertise acquired (e.g., Ericsson,1996). Creative geniuses know more than lesser creators, and the latter know more than those individuals who are not creative at all. However, historiometric research indicates that expertise acquisition has a much more ambiguous relation with creative development.
This ambiguity is first seen in the association between formal education and exceptional creative achievement (Simonton,1984c). For one thing, creative genius is not necessarily associated with extraordinary scholastic performance, and many geniuses were in fact mediocre students (Goertzel, Goertzel, & Goertzel,1978; Raskin,1936). Even more significant is the fact that creative development is not necessarily a positive monotonic function of the level of formal education attained (Simonton,1983). Specifically, the relation may be better described by an invertedU curve, meaning that there exists an optimal level of training beyond which additional educa- tion can have deleterious effects. This possibility is illustrated by a study I
conducted (Simonton,1976) of the301geniuses studied by Catharine Cox (1926), almost two thirds of whom were eminent creators, including such big names as Newton, Descartes, Goethe, Michelangelo, and Beethoven.
The achieved eminence of each genius had been previously assessed by Cattell (1903) based on the amount of space devoted to them in standard reference works. When this measure was plotted as a function the level of formal education obtained, there emerged an inverted Ucurve with a peak somewhere in the last couple of years of undergraduate educa- tion. Moreover, this result was replicated in another analysis (Simonton, 1984b) of more recent creators (collected by Goertzel, Goertzel, & Goertzel, 1978).
Admittedly, these findings have to be qualified by the fact that scholas- tic training and expertise acquisition are not necessarily equivalent pro- cesses. In some domains of creativity, such as the sciences, there is a high correspondence between the material presented in school and college and the knowledge needed to generate creative ideas. In the arts, in contrast, much of the substance of formal education may be irrelevant to creative development, if not outright detrimental. Therefore, it should come as no surprise that scientific geniuses are more likely to do better in school and to attain higher levels of formal education than are artistic geniuses (Goertzel, Goertzel, & Goertzel,1978; Raskin,1936; Simonton,1986). Creative devel- opment in the arts appears to require expertise acquisition that occurs outside the classroom and lecture hall. This extracurricular training usu- ally includes contributions of teachers, coaches, or mentors, as well as role models at large (Simonton,1977b,1984a; Walberg, Rasher, & Parkerson, 1980).
Even so, this more specialized training still fails to operate in a fashion consistent with the simple idea that creative development is equivalent to expertise acquisition. Take the10-year rule as a case in point. Supposedly it takes about a decade of intensive study and practice to acquire the capac- ity for creative genius (Hayes, 1989). In addition, it is assumed that the more extensive the training the higher the level of creative achievement (Ericsson, 1996). Yet the empirical data reveal a very different picture.
Although substantial variation exists in the amount of time devoted to domain-specific training, this variation isnegativelycorrelated with creative genius (Simonton,1996). This inverse association is illustrated in a histo- riometric study of 120classical composers (Simonton, 1991b). Although the most eminent creators in this domain tended to begin their training at younger ages than the least eminent, they also began producing first-rate compositions at a younger age. In fact, their precocity in creative output was even more pronounced than was their precocity in creative develop- ment. Stated more directly, the greatest geniuses in classical music spent less time in expertise acquisition before they began to exhibit their creativ- ity. Nor is this accelerated mastery unique to artistic creativity. Scientific
geniuses also require less time in domain-specific training before they begin to make major discoveries (Simonton,2004).
Expert Performance: Creative Productivity
Whatever may be the complexities of expertise acquisition, the mastered knowledge eventually must be converted into expert performance. More specifically, creative potential must be actualized as creative productivity.
It is the latter output, not the former ability, that earns a creator the designa- tiongenius. Moreover, if creativity entails nothing more than the exploita- tion of domain-specific expertise, then we would have certain expectations about how that productivity is manifested over the career course. In par- ticular, if “practice makes perfect,” then creators should get better and better at what they do best – generate creative products. To be sure, as they approach perfection, the “learning curve” for their expertise may level off.
Yet even allowing for such “diminishing returns” we should still predict a positive monotonic function.
That prediction is contradicted by the historiometric research on creative geniuses. In the first place, the output of creative products across the lifes- pan is not positive monotonic. Instead, creative productivity is described by an inverted-backwardsJ curve (Simonton,1988). In other words, the output rate first rapidly increases to reach a peak in the30s or40s and there- after undergoes a gradual decline. This was first demonstrated by Qu´etelet back in1835and later more fully documented in Lehman’s (1953) extensive research on the relation between age and achievement. Although various critics have questioned the validity of the age decrement (e.g., Lindauer, 2003), a large empirical and theoretical literature shows that the postpeak decline cannot be questioned (Simonton,1997,2002). Of course, one can argue that the downward slope does not necessarily reflect a decline in creative expertise. Perhaps the decline in physical health causes the rate of output to slow down without harming in any way the quality of that output (Lehman,1953; Lindauer,2003).
There are two problems with this explanation, however. For one thing, the age decrement takes place when physical health is introduced as a control variable (Simonton,1977a). Even more important, the impact of individual creative products displays the same inverted-backwardJcurve (Simonton,1980a,1980b). In other words, the best work is most likely to appear during the early career peak rather than toward the end of the career. This conclusion is reinforced by research on the longitudinal loca- tion of career landmarks (Raskin,1936; Simonton,1991a,1991b; Zusne, 1976). These landmarks are three in number: the first major work, the best work, and the last major work. If creativity were a simple matter of accu- mulated expertise, we would expect that the best work would appear at the same age as the last major work or at least that the best work would appear
close to the same age (if there took place some diminishing returns). Yet that is not what happens. Instead, the best work emerges at an age closer to the first major work than to the last major work. Most typically, the first high-impact contribution is produced in the late20s, the most influ- ential or highly acclaimed contribution in the late30s or early40s, and the last high-impact contribution in the middle or late50s. In different terms, whereas10–15years separates the first and best creative products,15–20 years separates the best and the last creative products.
Last but not least are the findings regarding overtraining and crosstrain- ing effects (Simonton,2000). For most domains of creativity it is possible to classify creative products into distinct genres. For instance, literary prod- ucts may be grouped into such genre as fiction, drama, and poetry, and each of these may be further subdivided (e.g., novels versus short stories, tragic versus comic plays, and lyric versus epic poetry). If creativity depended solely on a domain-specific knowledge base, then we would predict that expertise is genre specific as well. If a writer wishes to improve the quality of his or her poetry, it should be better to write more poetry than to write more novels or plays. Yet this expectation is not confirmed by historio- metric research. For instance, in a study of opera composers it was found that creativity was optimized by switching back and forth between genre (e.g., dramatic versus comic operas) rather than consistently producing in the same genre (Simonton, 2000). Indeed, opera composers benefited by creating works outside the opera medium altogether, including purely instrumental compositions. Similar findings emerged in an inquiry into the careers of high-impact scientists, the most influential creators going back and forth among several distinct substantive areas and methodological approaches (Root-Bernstein, Bernstein, & Garnier,1993). The less influen- tial scientists, in contrast, tend to stick to a single topic and method before switching to another (see also Simonton,2004). These effects are analogous to overtraining and crosstraining effects in sports. Creativity is nurtured by crosstraining and hindered by overtraining. It is more crucial for knowl- edge to be broad than to be deep.
This cutting back and forth between distinct domains of expertise is illustrated in the career of Charles Darwin, and especially the period in which his epochalOrigin of Species emerged. Darwin first began compil- ing a notebook on the subject of the “transmutation of species” in1837, the year after his return from his voyage on the H.M.S. Beagle. In 1859 the first edition of theOriginwas published. Between1837and1859, inclu- sively, Darwin was engaged on a great many other projects. These included several studies on the geology of South America (1837–1846), coral forma- tion (1837–1842), volcanic islands and mountain chains (1838–1844), and geological formations in Scotland and Wales (1838–1842); preparation of the volumes reporting the zoological findings of the Beaglevoyage (five volumes on fossil mammals, mammals, birds, fish, and reptiles worked
on from1837to1845); extensive monographs on both fossil and modern cirripedes (1847–1854); plus a host of miscellaneous papers, notes, and reviews on topics as diverse as earthworms, mold, glacial action, erratic boulders, volcanic rocks, a rock seen on an iceberg, dust falling on ships in the Atlantic, the effects of salt water on seeds, seed vitality, the role of bees in the fertilization of Papilionaceous flowers, Waterhouse’sNatural History of the Mammalia, and the species or generaRhea americana, Sagitta, Planaria, andArthrobalanus(1837–1858). That is an impressive range of top- ics, especially given that this period accounts for only about a quarter of his entire career as a scientist! Obviously, Darwin had many different things on his mind during the period that he conceived his theory of evolution by natural selection. Moreover, the cross-talk among these diverse projects no doubt enhanced rather than harmed the creativity of what he produced during this time.
reason
As noted, a whole school of psychologists have argued that scientific creativity entails nothing more than straightforward logical reasoning.
Given sufficient disciplinary knowledge, plus enough skill in the scien- tific “method,” discoveries become almost mundane accomplishments. For instance, Herbert Simon (1973), perhaps the most conspicuous proponent of this viewpoint, claimed “Mendeleev’s Periodic Table does not involve a notion of pattern more complex than that required to handle patterned letter sequences” (p.479). Going beyond mere speculation, Simon even conducted the following informal experiment:
On eight occasions I have sat down at lunch with colleagues who are good applied mathematicians and said to them: “I have a problem that you can perhaps help me with. I have some very nice data that can be fitted very accurately for large values of the independent variable by an exponential function, but for small values they fit a linear function accurately. Can you suggest a smooth function that will give me a good fit through the whole range?” (H. A. Simon,1986, p.7)
Of the eight colleagues, five arrived at a solution in just a few minutes. In ignorance of what Simon was up to, they had independently arrived at Max Planck’s formula for black body radiation – an achievement that earned Planck a Nobel prize for physics. What is remarkable about this example is the implicit argument that knowledge is not very important. If scientists are equipped with enough logical prowess, especially if it takes the form of mathematical reasoning, then they can be creative without knowing very much about the field. In this case, applied mathematicians could arrive at Planck’s formula in complete ignorance of the empirical and theoret- ical literature on black body radiation! No wonder it is so easy to write
“discovery programs” that duplicate the achievements of great scientists
without having to program substantial domain-specific knowledge into the systems.
Even so, great scientists themselves do not have such an exalted opinion of logical analysis as a creative force. For example, Max Planck (1949) held that creative scientists “must have a vivid intuitive imagination, for new ideas are not generated by deduction, but by an artistically creative imag- ination” (p.109). Similarly, Albert Einstein reported “to these elementary laws there leads no logical path, but only intuition” (Holton, 1971–1972, p.97). In fact, Einstein maintained that logic only came later, after the cre- ative ideas had emerged through some free, combinatorial process. “Taken from a psychological viewpoint . . . combinatory play seems to be the essen- tial feature in productive thought – before there is any connection with logical construction in words or other kinds of signs which can be com- municated to others” (Hadamard, 1945, p. 142). Consequently, Einstein continued, “conventional words or other signs have to be sought for labo- riously only in a secondary stage, when the mentioned associative play is sufficiently established and can be reproduced at will” (p.143).
A distinctive feature of this associative play is its highly chaotic nature.
According to William James (1880),
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 rarefied 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. (p.456)
This description seems a far cry from the linear, step-by-step logic that some hypothesize to underlie scientific creativity.
Others have made it more explicit that creativity is best described as a quasirandom combinatorial process in which chance plays a major role.
For instance, Jacques Hadamard (1945), the mathematician, claimed that mathematical creativity requires the discovery of unusual but fruitful com- binations of ideas. To find such combinations, it is “necessary to construct the very numerous possible combinations, among which the useful ones are to be found” (p.29). But “it cannot be avoided that this first operation take place, to a certain extent, at random, so that the role of chance is hardly doubtful in this first step of the mental process” (pp.29–30).
Similarly, the mathematician Henri Poincar´e (1921), in describing one discovery, observed how “ideas rose in crowds; I felt them collide until pairs interlocked, so to speak, making a stable combination” (p. 387).
Poincar´e compared these colliding images to “the hooked atoms of Epicurus” that jiggle and bump “like the molecules of gas in the kinematic