eral times without being married to more than one person. Only when a person lets go of the misleading (though more common) definition of the word “marry” can the person break this fixation. Fixation is a common result of the ordinary processes of problem solv- ing. When faced with any type of problem, an individual brings to the task his or her experience with similar problems, such as the knowledge about the domain and the individual’s expectations or intuitions about how to approach the problem. Often schemas provide useful short-cuts in the solu- tion of well-defined problems. For example, most people have schemas for solving word problems in math, based on their previous experience with such problems (Newell & Simon, 1972 ). Usually, we examine the question and use the given numeric values to set up a familiar formula for solution. However, when a problem’s definition and its goals are ill structured, our expectations about how the problem should be approached may be more detrimental than helpful. In fact, the key difficulty in many insight prob- lems is that they are based on the premise that the problem solver will build on an incorrect or misleading expectation that must be overcome in order to solve the problem. For example, consider the following problem: If you have black socks and brown socks in your drawer, mixed in a ratio of 4 to 5 , how many socks will you have to take out to make sure that you have a pair of the same color? When people see a problem involving numbers, they usually assume, correctly, that there are some calculations to be done. Therefore, they con- centrate on the numerical information, in this case the ratio information, in pursuing a solution (Davidson, 1995 ). However, this assumption is an example of negative transfer, a misleading expectation. If the problem is represented without this numerical information, we notice that it can be solved in a straightforward manner without considering the ratio infor- mation at all: Pull out two socks and they may match. If not, the third sock will definitely match one of the other two, given that there are only two colors of socks in the drawer. Based on the limited amount of research that has been done on the information-processing components of problem recognition, definition, and representation, it appears that these aspects of problem solving may not require special kinds of thinking. However, attention and openness are likely to be crucial to the discovery and creation of problems and to the selection of a problem representation. The metacognitive processes in- volved in these early stages of problem formulation are both divergent and convergent and appear to rely on analogical thinking as well as incubation and insight. Recognizing, Defining and Representing Problems 21 individual differences: abilities and dispositions Traditionally, problem solving research has not focused on the role of in- dividual differences beyond a consideration of general cognitive ability. However, psychologists who have examined the early stages of prob- lem solving have found that there are important sources of individual variation that affect the processes of problem recognition, definition, and representation. Individual differences have been found to play a role in the early stages of well-defined problem solving (MacLeod, Hunt, & Mathews, 1978 ; Sternberg & Weil, 1980 ). For example, MacLeod et al. ( 1978 ) found that in- dividual differences in ability influence problem representation. In their study, participants were presented with simple declarative sentences such as “Plus is above star.” Their results showed that most participants repre- sented the sentence linguistically. In contrast, participants who had high spatial abilities were more likely to represent the content of the sentence pictorially. The authors concluded that the processes in sentence compre- hension are not universally generalizable, but rather depend on the abil- ities of the individual. Similarly, mental representations of problems are also affected by individual differences in ability. Getzels and Csikszentmihalyi ( 1976 ) found that individuals who were successfully creative exhibited a concern for problem finding throughout the creative process. This “concern” can be characterized as a disposition or a mental set that attends to the nature of the problem definition and representation throughout the process of solving the problem at hand. Their study found that the most products were produced by individuals who reevaluated the way they had initially defined and represented the problem during all stages of the problem-solving process. One source of information about the abilities and dispositions that may be influential factors in the processes of problem recognition, defi- nition, and representation in ill-defined problem solving is the literature on creativity. As discussed earlier, the processes of recognizing a prob- lem, redefining problems, and representing them in various ways are es- sentially creative processes. The creativity literature has identified several individual-difference variables that appear to influence creative problem solving, including divergent thinking, openness, tolerance of ambiguity, and intrinsic motivation. Do some people have the ability to think more divergently or flexibly than others? Individual differences in intelligence and personality have been linked to differences in creative performance in various studies (see Sternberg & Lubart, 1995 ). Psychologists have often pointed out the im- portance of divergent-thinking abilities in creative problem solving. One way divergent thinking has been measured is using the Torrance Tests of Creative Thinking (Torrance & Ball, 1984 ), which include several measures 22 Pretz, Naples, and Sternberg of an individual’s ability to think divergently and flexibly. For example, the Alternate Uses Task asks participants to name as many uses as they can for an everyday object, such as a paper clip or an eraser. Responses of great diversity and number allegedly indicate greater divergent think- ing ability and cognitive flexibility. Scores on the Torrance Tests have been associated with greater creative performance. This association between di- vergent thinking and creativity may be due to the ability to think of many and diverse ways of defining and representing a problem. Thinking diver- gently and flexibly may not help in the latter stages of problem solving, when a solution must be evaluated for accuracy; evaluation relies on an- alytical and convergent thinking abilities. However, divergent thinking ability is more likely to be critical in the early stages of solving, when the problem remains open-ended and various definitions and representations of the problem must be considered. As mentioned earlier, one of the critical processes associated with prob- lem finding is attention to and perception of the environment in which a problem is discovered (Mumford et al., 1994 ). Research on creativity has demonstrated that highly creative individuals are those who have a broad range of attention relative to less creative people. When experiencing the world, creative people tend to filter out fewer distracters in the environ- ment (Eysenck, 1997 ). Because creative individuals take in information that other people would consider irrelevant, a highly creative person’s chances of detecting subtle patterns and hidden anomalies are greater than the chances of a less creative person doing so. Besides abilities, are there dispositional traits, such as personality at- tributes or cognitive style, that predispose people to being able to identify problems and realize creative ways to define and represent them? Many psychologists have argued that dispositions are a key factor in problem finding (e.g., Ennis, 1987 ; Jay & Perkins, 1997 ). Jay and Perkins ( 1997 ) have claimed: “Abilities, knowledge, and strategies enable a person to problem find, and contexts provide the stimulus, but it is dispositions that actu- ally promote the initiation of problem finding” (p. 286 ). Jay ( 1996 ) found that problem-finding behavior was enhanced when it was encouraged and guided. Given the fact that real-world problem-solving situations often do not include such guidance and prompts, it appears that the disposition spontaneously to engage in problem-finding behavior is very important. Perhaps individuals who are prompted to take a lot of time during the identification, definition, and representation phases of problem solving will eventually internalize these strategies and spontaneously engage in problem-finding behavior, even in the absence of prompts and encourage- ment to do so. Are there personality traits associated with creative problem solving? Quite a bit of research has sought to find a link between personality and creativity. In a meta-analysis of the relationship between personality traits and creativity, Feist ( 1998 ) found that creative individuals tended to be Recognizing, Defining and Representing Problems 23 “autonomous, introverted, open to new experiences, norm-doubting, self- confident, self-accepting, driven, ambitious, dominant, hostile, and impul- sive” (p. 299 ). Other traits associated with creativity include tolerance of ambiguity (MacKinnon, 1978 ; Sternberg & Lubart, 1995 ) and intuitiveness (Bastick, 1982 ). Eysenck ( 1997 ) has discussed the creative individual in terms of ego strength and psychopathology. Ego strength is a term used by Barron ( 1969 ) and others to refer to a strong, self-determined, dominant, self-reliant, and independent person. Eysenck has found a link between creativity and sub- clinical levels of psychoticism as measured by the Eysenck Personality Questionnaire (Eysenck & Eysenck, 1975 ). Eysenck conceived of psychoti- cism as a continuum ranging from conventional, socialized, and altruistic traits to aggressive, impulsive, and psychotic traits. Creative individuals were found to be slightly more psychotic than average with respect to this continuum. This observation has been supported by various reports of heightened levels of actual psychopathology among creative populations (e.g., Kaufman, 2001 ). Most research that has attempted to identify the personality charac- teristics associated with creativity has found a great deal of variability among creative individuals, suggesting that the ability to create problems and solve them in a way that is considered useful and original may vary greatly from domain to domain. For example, the traits that are associated with being a creative visual artist may be very dissimilar from the traits associated with being a creative business manager. For a creative visual artist to transform his or her creative idea into a reality, he or she often must spend long hours in the studio. But a creative business manager will probably need to interact intensely with many different types of people in order to carry out her creative vision for her organization. Another important factor that has been identified as critical to the cre- ative process (e.g., Amabile, 1996 ), as well as to the early stages of problem solving, is motivation. It is logical that you will not recognize problems that you are not motivated to find. For example, recall the problem of lack of parking mentioned at the beginning of this chapter. If you walk or take public transportation to work every day, you may not even notice, let alone be concerned with, the problems facing automobile commuters. If you lack intrinsic motivation, you are less likely to pursue a difficult problem such as this one. Extrinsic motivation can also encourage creative problem solv- ing if it provides more information or somehow makes it easier to solve the problem; however, extrinsic motivation that simply offers a reward but does not aid the problem-solving process (such as being paid to work on the downtown parking problem despite your lack of interest in the is- sue) will not lead to more creative solutions (Collins & Amabile, 1999 ). Amabile ( 1996 ) has also noted the importance of curiosity and a playful attitude in the facilitation of creative problem solving. People who enjoy experimenting with unusual ideas are more likely to recognize novel ways 24 Pretz, Naples, and Sternberg of defining and representing problems, in the same way that curious people are more likely to discover or create problems that escape the awareness of others. These abilities and dispositions have been associated with creativity. However, the relationship of these with problem recognition, definition, and representation remains to be investigated carefully. Individual differ- ence variables that are associated with creativity may prove to be a fruitful starting point for further research on the factors that influence the early stages of problem solving. social context Any discussion of problem-solving abilities must survey the environment in which an individual encounters a problem. Peers, culture, and even lan- guage structure play a role in the recognition, definition, and representation of a problem. Social forces can influence substantially an individual’s efforts in cre- atively defining, recognizing, or representing a problem (e.g., Sternberg, Kaufman, & Pretz, 2001 ; Sternberg, Kaufman, & Pretz, 2002 ). When an in- dividual recognizes a problem in his or her field, this recognition may be viewed as “rocking the boat.” The existence of a new problem may sug- gest an overlooked or ignored shortcoming in a field or situation. The social context affects problem recognition and definition through the field’s ad- herence to current paradigms. For example, problems studied in the field of social cognition previously employed social-psychological methodol- ogy to examine the effect of beliefs about social groups on behavior. How- ever, recent attraction to the use of functional magnetic resonance imaging techniques in research by neuroscientists and cognitive psychologists has become a tool of interest to some social psychologists who are interested in social cognition (e.g., Cunningham, Johnson, Gatenby, Gore, & Banaji, 2001 ). The availability of such resources, the field’s acceptance of the valid- ity of the methodology, as well as the neuroscience community’s acceptance of social psychologists will affect the way that social psychologists discover and define problems in their field, especially among researchers interested in embarking on the new subdomain of “social cognitive neuroscience” (Ochsner & Lieberman, 2001 ). Problem definition is affected by social context in any domain. Indi- viduals can become unable to redefine problems or evaluate progress on current problems due to the attitudes of the group. For example, in an office environment, individuals may be familiar with a particular com- puter application for word processing. However, the program eventually may become outdated or unsupported. Initially, the group may simply go through the process of converting files or rewriting documents, rather than abandoning the program for one that is more appropriate. Here the Recognizing, Defining and Representing Problems 25 problem has become not word processing, but rather the word processing program itself. The problem is not particularly difficult to spot, but the ways of the group may be so entrenched that changing programs becomes an unacceptable option. In other words, the attitudes of a group can be pervasive in the decision process of the individual. The influence of the social context on problem recognition can be il- lustrated by an example from the field of psychology. In the late 1950 s, Rosenblatt ( 1958 ) developed neural networks using elements that were designed to model human cognition, which he called perceptrons. Fol- lowing this early work, other researchers in the field pointed out limita- tions of Rosenblatt’s networks (Minsky & Papert, 1969 ). Minsky and Papert claimed that these early networks were unable to solve classification prob- lems whose solutions were nonlinear (Beale & Jackson, 1990 ). Based on the argument that most interesting problems attempted by humans of- ten require a nonlinear solution, this weakness was regarded as a fatal flaw in Rosenblatt’s network design. As a result of the field’s influence, there was little research in the field of neural networks for almost three decades; networks had been deemed inappropriate for modeling cogni- tion. It was not until much later that the field gave neural networks another chance (Rumelhart, McClelland, & University of California San Diego, 1986 ). Rumelhart and McClelland’s new vision of neural networks illus- trated that such models did have the power to model more complex human cognition, and resulted in a rush of research interest in this area. Despite the fact that there was not a tremendous amount of evidence against the viability of neural networks at the time of Minsky and Papert’s critique, the social context of the field hindered the progress of research in this vein for quite some time. The social context has a strong, sometimes unnoticed, effect on problem solving, beginning with the very early stages. Immediate clues from the en- vironment can affect the type of definition or representation used to solve a problem (Gick & Holyoak, 1980 , 1983 ). Even the traditions and attitudes of a group will affect the types of problems recognized by its members, the terms in which they define those problems, and the ways they repre- sent the problems as they prepare to solve them. Often, the most difficult part of problem formulation requires an individual to call into question these norms and expectations in order to most appropriately examine the phenomenon of interest. summary and conclusions What We Know The earliest stages of problem solving involve recognizing that a prob- lem exists, defining the scope and goals of the problem, and representing 26 Pretz, Naples, and Sternberg information about that problem in a way that helps establish a viable path to solution. For the most part, research on problem solving has focused on explaining the solution of well-defined problems that are already recog- nized and presented directly to the problem solver. When we approach a new situation, our knowledge based on prior experiences will influence our ability to define and represent a problem correctly. In fact, we may fail to notice the existence of a problem if it runs counter to our strongly held expectations. To the extent that an individual has misleading expectations or schemas about a problem, due either to crystallized expertise or to the effects of misleading context, that person may have difficulty thinking flexibly about how to approach the dilemma. Recall the lemonade and iced tea example. Our assumption that lemonade and iced tea are beverages in liquid form impedes our ability to think of them in any other form. The processes involved in problem recognition, definition, and represen- tation are quite varied. To notice a problem, a person must attend broadly to all pieces of relevant information in a situation. Additional knowledge from past experience with similar problems must also be accessed. How- ever, the likelihood that an individual will spontaneously notice analogies between problems in disparate domains is rather small (Gick & Holyoak, 1980 ). Individual differences in cognitive abilities and personality may ex- plain why some people are better at solving ill-defined problems than are others. The ability to think divergently and flexibly is valuable in the process of problem formulation, as is an open and intrinsically motivated dispo- sition. Perhaps the most critical variable in determining whether a person discovers or creates a novel problem is that individual’s motivation to find it and work on developing an appropriate definition and representation of the issue. This disposition characterized by openness and curiosity may be regarded as a trait version of a mental set, a constant metacognitive attentiveness to the environment and the process of problem solving. In- dividuals with this disposition are always thinking of different ways to regard the information in their environment and the information they pos- sess in long-term memory. When they are working on a problem, they naturally attempt to redefine and re-represent the problem, thus increas- ing their chances of finding a definition and representation that will yield a creative solution. Finally, the social context may also facilitate the likelihood of noticing problems and thinking divergently about their solutions. If an environment does not encourage potentially creative individuals to seek and explore, they will not discover gaps in their understanding, and they will not learn to play with ideas nor practice taking different perspectives on problems with which they are confronted. Recognizing, Defining and Representing Problems 27 What We Need to Know In contrast to the later stages of problem solving, the stage of problem formulation appears to rely more heavily on disposition and social con- text. Unfortunately, relatively little empirical research has addressed these topics. We need to understand what makes a person more likely to engage him- or herself in seeking out ill-defined problems and experimenting with various ways to represent them. We need to know how people who are con- strained by misleading expectations and schemas break out of their mental sets in order to gain new perspectives on problems. Can we teach children to think with this kind of mindful curiosity? We hope that teachers will allow children to practice suspending their judgment when necessary, to be playful in their search for a variety of solutions to problems. If our ultimate goal is to help people become better able to solve prob- lems that confront them in their personal and professional lives and in the concerns of the world, we must be prepared to examine the fuzzy issues surrounding problem recognition, definition, and representation. Because most of life’s problems are not cleanly packaged with one correct path to so- lution, it is important that we take on the ill-defined challenge of studying these early phases of problem solving in an effort to understand how prob- lem solving can be enhanced in these initial stages. Rather than educate others to become followers, it is in our best interest to encourage problem solvers to become active problem finders, to stay curious so that they dis- cover and create novel problems, and to think flexibly in the process of solving those problems. references Amabile, T. M. ( 1996 ). Creativity in context: Update to “The social psychology of creativ- ity.” Boulder, CO: Westview Press. Barron, F. ( 1969 ). Creative person and creative process . New York: Holt, Rinehart & Winston. Barsalou, L. W. ( 1999 ). Perceptual symbol systems. Behavioral & Brain Sciences, 22, 577 – 660 . Bastick, T. ( 1982 ). Intuition: How we think and act . New York: Wiley. Beale, R. J., & Jackson, T. ( 1990 ). Neural computing: An introduction . Bristol: Institute of Physics Publishing. Boden, M. ( 1999 ). Computer models of creativity. In R. J. Sternberg (Ed.), Handbook of Creativity (pp. 351 – 372 ). New York: Cambridge University Press. Bransford, J. D., & Stein, B. S. ( 1993 ). The ideal problem solver: A guide for improving thinking, learning, and creativity ( 2 nd ed.). New York: W. H. Freeman. Brophy, D. R. ( 1998 ). Understanding, measuring and enhancing individual creative problem solving efforts. Creativity Research Journal, 11, 123 – 150 . Chase, W. G., & Simon, H. A. ( 1973 ). Perception in chess. Cognitive Psychology, 4, 55 – 81 . 28 Pretz, Naples, and Sternberg Chi, M., Feltovich, P. J., & Glaser, R. ( 1981 ). Categorization and representation of physics problems by experts and novices. Cognitive Science, 5, 121 – 152 . Chi, M. T. H., Glaser, R., & Farr, M. J. (Eds.). ( 1988 ). The nature of expertise . Hillsdale, NJ: Erlbaum. Clement, E. & Richard, J. ( 1997 ). Knowledge of domain effects in problem rep- resentation: The case of Tower of Hanoi isomorphs. Thinking and Reasoning, 3 , 133 – 157 . Collins, M. A., & Amabile, T. M. ( 1999 ). Motivation and creativity. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 297 – 312 ). New York: Cambridge University Press. Csikszentmihalyi, M. ( 1996 ). Creativity . New York: Harper Collins. Csikszentmihalyi, M., & Sawyer, K. ( 1995 ). Creative insight: The social dimension of a solitary moment. In R. J. Sternberg & J. E. Davidson (Eds.), The nature of insight (pp. 329 – 363 ). Cambridge, MA: MIT Press. Cunningham, W. A., Johnson, M. K., Gatenby, J. C., Gore, J. C., & Banaji, M. R. ( 2001 , April). An fMRI study on the conscious and unconscious evaluations of social groups . Paper presented at the UCLA Conference on Social Cognitive Neuroscience, Los Angeles, CA. Davidson, J. E. ( 1995 ). The suddenness of insight. In R. J. Sternberg & J. E. Davidson (Eds.), The nature of insight (pp. 125 – 155 ). Cambridge, MA: MIT Press. Ellis, S., & Siegler, R. S. ( 1994 ). Development of problem solving. In R. J. Sternberg (Ed.), Thinking and problem solving. Handbook of perception and cognition ( 2 nd ed.) (pp. 333 – 367 ). San Diego, CA: Academic Press. Ennis, R. H. ( 1987 ). A taxonomy of critical thinking dispositions and abilities. In J. B. Baron & R. J. Sternberg (Eds.), Teaching thinking skills: Theory and practice . Series of books in psychology (pp. 9 – 26 ). New York: Freeman. Eysenck, H. J. ( 1997 ). Creativity and personality. In M. A. Runco (Ed.), The creativity research handbook (pp. 41 – 66 ). Cresskill, NJ: Hampton. Eysenck, H. J. & Eysenck, S. B. G. ( 1975 ). Manual of the Eysenck personality question- naire . London: Hodder & Stoughton. Feist, G. J. ( 1998 ). A meta-analysis of personality in scientific and artistic creativity. Personality and Social Psychology Review, 2, 290 – 309 . Frensch, P. A., & Sternberg, R. J. ( 1989 ). Expertise and intelligent thinking: When is it worse to know better? In R. J. Sternberg (Ed.), Advances in the psychology of human intelligence (Vol. 5 , pp. 157 – 188 ). Hillsdale, NJ: Erlbaum. Getzels, J. W. ( 1982 ). The problem of the problem. In R. Hogarth (Ed.), New direc- tions for methodology of social and behavioral science: Question framing and response consistency (No. 11 ). San Francisco: Jossey-Bass. Getzels, J. W., & Csikszentmihalyi, M. ( 1976 ). The creative vision: A longitudinal study of problem finding in art. New York: Wiley. Gick, M. L., & Holyoak, K. J. ( 1980 ). Analogical problem solving. Cognitive Psychol- ogy, 12, 306 – 355 . Gick, M. L., & Holyoak, K. J. ( 1983 ). Schema induction and analogical transfer. Cognitive Psychology, 15, 1 – 38 . Goldstone, R. L., & Barsalou, L. ( 1998 ). Reuniting perceptionand conception. In S. A. Sloman and L. J. Rips (Eds.), Similarity and symbols in human thinking (pp. 145 – 176 ). Cambridge, MA: MIT Press. Hayes, J. R. ( 1989 ). The complete problem solver ( 2 nd ed.). Hillsdale, NJ: Erlbaum. Recognizing, Defining and Representing Problems 29 Jay,E.S.( 1996 ). The nature of problem finding in students’ scientific inquiry . Unpub- lished doctoral dissertation, Harvard University, Graduate School of Education, Cambridge, MA. Jay, E. S., & Perkins, D. N. ( 1997 ). Problem finding: The search for mechanism. In M. A. Runco (Ed.), The creativity research handbook (pp. 257 – 293 ). Cresskill, NJ: Hampton. Kaufman, J. C. ( 2001 ). The Sylvia Plath effect: Mental illness in eminent creative writers. Journal of Creative Behavior, 35 ( 1 ), 37 – 50 . Kotovsky, K., Hayes, J. R., & Simon, H. A. ( 1985 ). Why are some problems hard? Evidence from the Tower of Hanoi. Cognitive Psychology, 17, 248 – 294 . Lesgold, A. M. ( 1988 ). Problem solving. In R. J. Sternberg & E. E. Smith (Eds.), The psychology of human thought (pp. 188 – 213 ). New York: Cambridge University Press. MacKinnon, D. W. ( 1978 ). In search of human effectiveness: Identifying and developing creativity. Buffalo, NY: Creative Education Foundation. MacLeod, C. M., Hunt, E. B., & Mathews, N. N. ( 1978 ). Individual differences in the verification of sentence-picture relationships. Journal of Verbal Learning and Verbal Behavior, 17, 493 – 507 . Metcalfe, J. A., & Wiebe, D. ( 1987 ). Intuition in insight and noninsight problem solving. Memory & Cognition, 15, 238 – 246 . Minsky, M., & Papert, S. ( 1969 ). Perceptrons:An introduction to computational geometry . Cambridge, MA: MIT Press. Mumford,M. D., Reiter-Palmon, R., & Redmond, M. R. ( 1994 ). Problem construction and cognition: Applying problem representations in ill-defined domains. In M. A. Runco (Ed.), Problem finding, problem solving, and creativity (pp. 1 – 39 ). Norwood, NJ: Ablex. Newell, A., & Simon, H. ( 1972 ). Human problem solving . Englewood Cliffs, NJ: Prentice-Hall. Nisbett, R. E., & Wilson, T. D. ( 1977 ). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231 – 259 . Ochsner, K. N., & Lieberman, M. D. ( 2001 ). The emergence of social cognitive neuroscience. American Psychologist, 56, 717 – 734 . Reed, S. K. ( 1987 ). A structure-mapping model for word problems. Journal of Experimental Psychology: Learning, Memory, & Cognition, 13, 125 – 139 . Rosenblatt, F. ( 1958 ). The perceptron: A probabilistic model for information storage and organization in the brain. Psychological Review, 65, 386 – 408 . Rumelhart, D. E., McClelland, J. L., & University of California San Diego, PDP Re- search Group. ( 1986 ). Parallel distributed processing: Explorations in the microstruc- ture of cognition . Cambridge, MA: MIT Press. Seifert, C. M., Meyer, D. E., Davidson, N., Patalano, A. L., & Yaniv, I. ( 1995 ). Demystification of cognitive insight: Opportunistic assimilation and the prepared-mind perspective. In R. J. Sternberg & J. E. Davidson (Eds.), The nature of insight (pp. 125 – 155 ). Cambridge, MA: MIT Press. Siegler, R. S. ( 1978 ). The origins of scientific reasoning. In R. S. Siegler (Ed.), Chil- dren’s thinking: What develops? (pp. 109 – 49 ). Hillsdale, NJ: Erlbaum. Simon, H. A., & Newell, A. ( 1971 ). Human problem solving: The state of the theory in 1970 . American Psychologist, 26, 145 – 159 . 30 Pretz, Naples, and Sternberg Simonton, D. ( 1999 ). Creativity from a historiometric perspective. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 116 – 136 ). New York: Cambridge University Press. Smith, S. M. ( 1995 ). Fixation, incubation, and insight in memory and creative think- ing. In S. M. Smith, T. B. Ward, & R. A. Finke (Eds.), The creative cognition approach (pp. 135 – 156 ). Cambridge, MA: MIT Press. Smith, S. M., & Blankenship, S. E. ( 1989 ). Incubation effects. Bulletin of the Psycho- nomic Society, 27, 311 – 314 . Sternberg, R. J. ( 1985 ). Beyond IQ: A triarchic theory of human intelligence. New York: Cambridge University Press. Sternberg, R. J. ( 1986 ). Intelligence applied? Understanding and increasing your intel- lectual skills . San Diego, CA: Harcourt Brace Jovanovich. Sternberg, R. J. ( 1999 ). Cognitive psychology ( 2 nd ed.). New York: Harcourt Brace. Sternberg, R. J., & Davidson, J. E. (Eds.) ( 1995 ). The nature of insight . Cambridge, MA: MIT Press. Sternberg, R. J., & Weil, E. M. ( 1980 ). An aptitude-strategy interaction in linear syllogistic reasoning. Journal of Educational Psychology, 72, 226 – 234 . Sternberg, R. J., Kaufman, J. C., & Pretz, J. E. ( 2001 ). The propulsion model of creative contributions applied to the arts and letters. Journal of Creative Behavior, 35, 75 – 101 . Sternberg, R. J., Kaufman, J. C., & Pretz, J. E. ( 2002 ). The creativity conundrum: A propulsion model of kinds of creative contributions . Philadelphia: Psychology Press. Sternberg, R. J., & Lubart, T. I. ( 1995 ). Defying the crowd: Cultivating creativity in a culture of conformity. New York: Free Press. Terman, L. M. ( 1950 ). Concept Mastery Test . New York: Psychological Corporation. Torrance, E. P., & Ball, O. E. ( 1984 ). Torrance Tests of Creative Thinking: Revised manual. Bensenville, IL: Scholastic Test Service. Wallas, G. ( 1926 ). The art of thought . New York: Harcourt, Brace. Weisberg, R. W., & Alba, J. W. ( 1981 ). An examination of the alleged role of “fixation” in the solution of several “insight” problems. Journal of Experimental Psychology: General, 110, 169 – 192 . Wertheimer, M. ( 1945 ). Productive thinking . Westport, CT: Greenwood Press. author note Preparation of this book chapter was supported by Grant REC- 9979843 from the National Science Foundation and by a government grant un- der the Javits Act Program (Grant No. R 206 R 000001 ) as administered by the Office of Educational Research and Improvement, U.S. Department of Education. Grantees undertaking such projects are encouraged to express freely their professional judgment. This article, therefore, does not neces- sarily represent the positions or the policies of the U.S. government, and no official endorsement should be inferred. Requests for reprints should be sent to Robert J. Sternberg, Yale Univer- sity, The Yale Center for the Psychology of Abilities, Competencies, and Expertise, P.O. Box 208358 , New Haven, CT 06520 – 8358 . 2 The Acquisition of Expert Performance as Problem Solving Construction and Modification of Mediating Mechanisms Through Deliberate Practice K. Anders Ericsson How do experts reach their high level of performance? Recent reviews (Ericsson, 1996 , 1998 b, 2001 ; Ericsson & Lehmann, 1996 ) dispel the com- mon belief that “talented” expert performers attain very high levels of performance virtually automatically through cumulative domain-related experience. Instead, empirical evidence strongly implies that even the most “talented” individuals in a domain must spend over ten years actively en- gaging in particular practice activities (deliberate practice) that lead to gradual improvements in skill and adaptations that increase performance. In this chapter I argue that the acquisition of expert performance can be described as a sequence of mastered challenges with increasing lev- els of difficulty, such as playing pieces of music, performing challenging gymnastic routines, and solving complex mathematical problems. Differ- ent levels of mastery present the learner with different kinds of problems that must be solved for the skill to develop further. And each individual’s path toward skilled performance is distinct; it depends on when technical challenges were encountered and the specific methods used to help the individuals continue their development. When beginners are first introduced to a domain of expertise they can successfully perform only the most simple tasks and activities. With the aid of instruction and training many individuals are able to master increas- ingly difficult tasks, thus gradually improving and slowly approaching the level of expert performers. The incremental nature of gaining mastery means that tasks that were initially impossible to perform can be executed effortlessly as increased skill is attained. When an individual attempts to perform a task that is too difficult, his or her available repertoire of methods and skills is insufficient to perform the task successfully. In this chapter I argue that when motivated individuals strive to overcome obstacles and master prerequisite aspects of a given task, they must engage in problem solving. Studies of how individuals eventu- ally master various types of problems should provide unique insights into 31 32 Ericsson how their cognitive mechanisms, representations, and knowledge change to solve those problems. Mastery of very difficult problems, such as an un- familiar technique or developing a better scientific theory, might require days, weeks, months, or even years, rather than minutes or hours available in most studies of problem solving in psychological laboratory. Some people will probably object that the label problem solving is in- appropriate when applied to the process of mastering tasks in a domain of expertise. They will propose that these phenomena should be classi- fied as skill acquisition or complex learning. To be sure, when most text- books in psychology discuss problem solving, they rarely refer to highly skilled performance, focusing instead on studies of classic problems, such as the Tower of Hanoi (Gagne & Smith, 1962 ; Simon, 1975 ), the tumor radiation problem (Duncker, 1945 ), and the pendulum problem (Maier, 1931 ). The large differences in tasks seem to support the argument that the two types of phenomena have always been viewed as distinct. However, a historical survey suggests that at one time a much closer relationship was assumed to exist between problem solving with puzzles and expert performance. In the foreword to his classic paper, Duncker ( 1945 , p. v), one of the foremost pioneers in the study of problem solving, claimed that he was tempted “to study productive thinking where it is most conspicuous in great achievements” and that “important information about the genesis of productive thought could be found in biographical material.” However, he concluded that “although a thunderstorm is the most striking example of electrical discharge, its laws are better investigated in little sparks within the laboratory. To study in simple, convenient forms what is complicated and difficult to access is the method of experimental science; to lose in this simplification just the essential aspects, is its notorious danger. Experimen- tal psychology, more than all other sciences, is continually faced with this danger.” Drawing on the analogy with the study of lightning, Duncker ( 1945 , p. v) intentionally restricted his study of problem solving to “practical and mathematical problems,” “because such material is more accessible, more suitable for experimentation.” Duncker assumed that the process of solving mathematical problems in the laboratory induced the same phenomenon of problem solving as the one observed in generation of great achievements and expert performance; the differences were primarily a matter of degree. The hypothesis that it is possible to use the laboratory to capture pure manifestations of mental functions observed in everyday life, such as prob- lem solving and memory, in simpler form, is one of the cornerstones of modern experimental psychology. Let us first use the example of lightning to show convincing evidence for the parallels between laboratory and ev- eryday phenomena, and then examine the evidence for parallels between problem solving in the laboratory and great achievements. The Acquisition of Expert Performance 33 How did 18 th-century scientists establish that sparks of static electricity encountered in everyday life were miniature versions of the lightning ob- served during thunderstorms? Since the earliest times people have recog- nized the strange effects of static electricity that cause invisible effects and sparks of discharge. In the 17 th and early 18 th century, scientists designed and refined machines to produce and store very high levels of static electric- ity and discharge bright sparks. Several scientists, among them Benjamin Franklin (Clark, 1983 ), proposed that lightning was a more intense ver- sion of the same phenomenon. Franklin listed many similarities, ‘They both gave out light of the same color and had crooked direction and swift motion … both were conducted by metals, made noise, ‘subsisting’ in water and ice, and could tear apart materials that they went through. In addition, both could kill animals, melt metals, set fire to inflammable sub- stances and produce sulfurous smell” (Clark, 1983,p. 80 ). Franklin then demonstrated that he could tap the static electricity in a thunderstorm by sending up a kite and conducting the static electricity down the wet string from the cloud. He was able to use the naturally occurring phenomenon of the storm to reproduce characteristics of static electrical discharge induced under laboratory conditions. Do we know enough about the characteristics of productive thinking used to solve practical and mathematical problems and the processes me- diating great achievements to draw the conclusion that productive think- ing captured in the laboratory will automatically apply to great achieve- ments? The answer must be “No.” The progress on understanding puzzle problem solving during the 20 th century was remarkable, and was largely attributable to Duncker’s ( 1945 ) research. However, as I show in the next section, the concept of problem solving does not correspond to a single well-defined phenomenon. It has changed considerably during the last century in tight connection with the changes in the dominant theoretical framework of general psychology. In fact, as more knowledge has been accumulated about various forms of thinking, such as decision making, comprehension, reasoning, planning, and creative thinking, the harder it has become to distinguish problem solving as a separate phenomenon with its unique processes and mechanisms. The research efforts to identify the structure of thinking led to great advances in the design of experiments and the methodology for trac- ing complex cognitive processes in the laboratory with recordings of eye movements and concurrent verbalization. These developments revealed that performance on simple laboratory tasks is often mediated by complex knowledge and semantic memory, and they provided tools for studying complex performance. Ericsson and Smith ( 1991 ) showed that the same methodology can be adapted to study expert performance and its mediating cognitive pro- cesses as well as the learning processes that led to this superior level of 34 Ericsson performance. For example, if we are interested in how world-class chess players are able to play better than other less accomplished players, we should study the cognitive processes involved in playing at the world-class level. If we are interested in how scientists are able to produce consistently superior pieces of research and how musicians are able to produce rich musical experiences for their audiences, we should study the processes involved in producing these achievements. Once we are able to reproduce expert performance with representative tasks in the laboratory (de Groot, 1946 / 1978 ; Ericsson & Smith, 1991 ), it is possible to submit the mediating processes to analysis and experimental variation. Consider how Benjamin Franklin waited for a thunderstorm to send up his kite to siphon off static electricity from the clouds. In much the same way, it is possible to reproduce the necessary representative conditions for expert performance and then merely request that willing expert performers exhibit their superior performance under those conditions. When the ev- eryday phenomenon of expertise can be reproduced in the laboratory, then the difficult problem of establishing the equivalence between phenomenon in everyday life and the laboratory can be avoided. outline of the chapter Until the 19 th century, most scientists and philosophers believed that it would be impossible to use scientific methods in the rigorous study of thinking and problem solving. I therefore briefly discuss how the study of problem solving evolved within the scientific discipline of psychology and how it led to studies of problem solving primarily using puzzles and traditional laboratory tasks. Then I focus on the methodological advances in studying thinking and how they allowed scientists to describe the struc- ture of problem solving with puzzles. In the main body of the chapter I discuss how the same methods can be used to study problem solving and thinking within the context of representative tasks that capture expert performance and its acquisition. I conclude with a brief discussion of our emerging knowledge of problem solving in highly skilled performance and its relation to problem solving with traditional puzzles and discuss some future directions of problem solving research. Approaches to the Study of Problem Solving and Thinking: Historical Background Conceptions of the structure of the human mind have gone through dra- matic changes during the history of our civilization. Humans have al- ways reflected on experiences and feelings. With Aristotle and other Greek philosophers, the search for the structure of consciousness and its basic The Acquisition of Expert Performance 35 elements became more systematic, based on observation and analysis of one’s own thinking. One of the central problems of these introspective efforts was the private nature of consciousness; one person’s conscious awareness could not be directly experienced by others. Only a few cen- turies ago prominent philosophers such as Immanuel Kant denied the possibility of even studying complex mental phenomena and subjective experience with scientific methods. When the first psychological laboratory was established in Germany to- ward the end of the 19 th century, Wilhelm Wundt, the founding father of experimental psychology, deliberately focused his research on the most ba- sic phenomenon, namely, sensory perception. Other pioneering researchers in psychology, such as Hermann Ebbinghaus, also designed techniques to measure basic forms of memory processes in which thinking and meaning- ful associations were minimized, such as Ebbinghaus’s famous nonsense syllables. Ebbinghaus showed that it was possible to design experimen- tal situations in which the results of complex mental capacities, such as memory, could be directly observed and measured. His empirical studies focused on general processes and capacities so basic that they were not me- diated by thinking and therefore could not be analyzed by introspective methods. These empirical paradigms allowed measurement with quanti- tative methods and the mathematical expression of general laws that were consistent with other natural sciences. The scientific approach of identifying general laws and simple basic mechanisms to account for the complex natural phenomena had been very successful in the natural sciences. The application of this analytical ap- proach to psychological phenomena is nicely summarized in Morgan’s canon: “In no case may we interpret an action as the outcome of the exer- cise of a higher faculty, if it can be interpreted as the outcome of the exer- cise of one which stands lower in the psychological scale” (Morgan, 1894 , p. 53 ). This general commitment to reductionism has had major impact on the study of higher level cognitive processes, such as problem solving. First, it led to prioritization of research that would extend our knowledge about the basic processes of sensation, perception, memory, and action. Second, the study of thinking and other complex mental phenomena was primarily motivated by the question of whether these complex phenomena could be accounted for within the current theoretical framework, and thus reduced to explanations based on the existing set of basic processes and capacities. As new knowledge emerged and new theoretical frameworks were de- veloped, the boundary between phenomena that could and could not be accounted for within the dominant theoretical framework kept changing. These theoretical transitions had far greater effect on the study of complex phenomena such as problem solving than they did on phenomena that could be explained by less complex mechanisms. 36 Ericsson The central challenge when studying thinking and problem solving in everyday life is their covert nature and complexity. Scientists must either find methods that allow them to monitor the complexity of thought in a task or find unfamiliar tasks in which participants’ relevant knowledge is min- imized. I sketch some important methodological developments that were significant stepping stones in the modern study of thinking and problem solving. I also briefly point out how the dominant theoretical frameworks in psychology changed during the 20 th century, how new frameworks con- ceived of problem solving differently and favored different empirical tasks for its study. Initial Attempts to Study Thinking The pioneering researchers in the 19 th century, such as Wilhelm Wundt, were explicit about the limitations of their methods and rejected the possi- bility of extending them to the study of complex experience and thinking in everyday life. Their studies were restricted to simple sensations, such as pure tones and points of light. The primary focus of the research was establishing how variations in the physical aspects of a stimulus are reg- istered by neural receptors and processed by the nervous system. It is important to remember that the very existence of many of these recep- tors was controversial at the start of this research. Wundt also studied the speed of neural transmission of information and recorded the time to re- act and make other judgments about simple stimuli (cf. Donders, 1868 / 1969 ). It is not typically recognized that Wundt deliberately limited his re- search to simple sensory stimuli to map out their encoding by receptors and the nervous system. Wundt argued that complex experience of the type analyzed by contemporary philosophers was qualitatively different from sensory stimulation. According to Wundt, an individual’s experience in everyday life is a complex mixture of sensory information that is merged with the vast amount of accumulated prior experience. Wundt believed that his analytic methods could not be used to uncover the fluid and com- plex structure of experience and thinking. Around the beginning of the 20 th century, many psychologists became increasingly interested in going beyond simple sensory stimuli and study- ing complex thought. They tried to develop Wundt’s rigorous introspective methods and the old philosophical method of self-observation into a sci- entifically acceptable method for describing the detailed elements of com- plex thought. The most famous developments occurred at the University of W ¨urzburg where highly trained observers (often professors of psychol- ogy) were asked to perform various tasks involving reasoning and deci- sion making and afterward give introspective analyses of their thoughts. To induce original thinking the investigators designed tasks that observers were unlikely to have encountered previously. Some of the tasks involved The Acquisition of Expert Performance 37 answering questions, such as “Do you understand the meaning of the fol- lowing saying ‘We depreciate everything that can be explained’?” (B ¨uhler in Rapaport, 1951,p. 41 ) These introspective analyses revealed many different types of thoughts and mental elements, even thoughts without associated sensory images – imageless thoughts. Karl B ¨uhler argued in a series of papers that because Wundt’s theory could not account for the existence of imageless thought, it must be incorrect. Wundt countered that it was impossible to simultane- ously perform the assigned task and observe neural activity in the brain. Wundt emerged victorious from his famous exchange with B ¨uhler, and some of Karl B ¨uhler’s original observers even conceded that their reports of imageless thought must have been flawed. It is important to note that Wundt forcefully argued all along that con- current introspective analysis of thinking is not possible because it would disturb the flow of thinking and produce potentially biasing inference. Subsequently, all of the main psychological theories for studying think- ing, such as behaviorism, gestalt psychology, and information process- ing, developed verbal reports methodologies to externalize thinking that would not disturb the associated thought processes and bias the verbalized information. Behaviorism and Studies of Thinking When John B. Watson ( 1913 ) introduced behaviorism in his seminal pa- per he criticized the introspective analysis of experience and proposed an alternative approach based on observable behavior and performance. A careful reading of this paper shows that there was considerable agreement between Wundt and Watson. Both rejected introspective analysis of com- plex mental phenomena such as thinking, although their theoretical ratio- nales differed. Furthermore, Watson accepted the research contributions on sensory perception and psychophysics by Wundt and his colleagues because these findings were supported by observable performance in the form of accurate perceptual judgments of presented stimuli. A fundamental assumption of behaviorism is that behavior can be de- scribed by sequences of responses elicited by selected aspects of environ- mental stimuli. This assumption closely matches the consensus view of thinking as a sequence of thoughts (Ericsson & Crutcher, 1991 ), illustrated in Figure 2 . 1 . Watson ( 1924 ) proposed that thinking could be described as covert, internalized behavior, especially in the form of inner subvocal speech. If one wished to study thinking, he or she should instruct sub- jects to “think aloud.” Watson ( 1920 ) was the first investigator to publish a case study of thinking aloud. These think-aloud verbalizations provided a direct externalization of the subject’s inner speech. They provided an un- biased trace of thinking that could be readily elicited by untrained adults, thus alleviating the problems of the extensive training and retrospective analysis associated with introspection.