The tendency to rely on habits and procedures that worked to solve a problem in the past is called

Blockbusting

Many authors have argued that invention requires divergent or lateral thinking to break free from conventional and habitual ways of thinking. People have a natural tendency to look at problems in a fixed way and to become trapped in a conservative view. Blockbusting is the conscious strategy to identify and overcome this functional fixedness. Key blocks are perceptional, emotional, cultural, and intellectual blocks. A typical perceptional block is stereotyping, where one can only look at something in a preconceived way. Emotional blocks can be found in the fear of failure or risk taking, and intolerance for insecurity. Intellectual blocks occur when one is too fixated on one's specialty, one type of solution, or not wanting to abandon certain dogmas or assumptions. Einstein demonstrated an intellectual block when he rejected quantum mechanics for esthetic reasons. Cultural blocks occur when an inventor does not consider solutions that are not currently fashionable or are politically incorrect. Many of the aforementioned heuristics, in one way or another, circumvent this type of block. Thus the inventor may analyze explicitly what limits he has put on the type of representation, materials, size, perceptional modality, modeling, etc.

Rigidity in Thinking

A major obstacle to effective problem solving is rigidity in thinking. For example, in some problem-solving situations the problem solver must use an object in a new way, such as using a brick as a doorstop or using a pencil as a bookmark. When a problem solver can only conceive of using an object in its most common function, then the problem cannot be solved. Duncker (1945) used the term functional fixedness to refer to a situation in which a problem solver cannot think of using an object in a new function that is required to solve the problem. Another example of rigidity occurs when a problem solver uses a well-learned procedure on a problem for which the procedure is inappropriate. For example, if a student solved a long series of arithmetic story problems that all contain the word ‘more’ and require adding the numbers together, the student may incorrectly carry out this same addition procedure for a new problem that actually requires subtracting the numbers from one another. Luchins (1942) used the term einstellung (or problem-solving set) to refer to this phenomenon. A goal of instruction in problem solving is to help students avoid rigid thinking.

Perceptual Barriers

From a lifetime of learning we are accustomed to perceiving things in familiar ways, often making it difficult to see new meanings, relationships, and ideas. Psychologists refer to our predisposition to perceive things in certain ways as a perceptual set, a mental set, or functional fixedness. It is the reverse of flexible, innovative thinking. Perceptual sets are different for different people, rooted in our unique experiences, interests, biases, and values. Perceptual sets are tied to our tendency to make quick decisions and jump to conclusions, rather than flexibly see alternatives.

Perceptual barriers make us ‘kick ourselves’ for not seeing a solution sooner. One classic demonstration showed that when a piece of string was needed to solve a problem, the string would quickly be perceived and used if it were dangling from a nail on the wall, but not if it were hanging a ‘No Smoking’ sign, a mirror, or a calendar. Perceptual blocks also prevent us from getting a complete and accurate picture of our world; the ‘real problem’ or ‘truth’ may be missed. For example, based on symptoms that seem familiar, a physician, auto mechanic, or computer expert may persist in misclassifying a problem and will recommend incorrect treatments. A school teacher who fixates on IQ scores will fail to perceive students who are highly creative, highly artistic, or gifted in a single area such as theater studies or computer programming. An instructor who successfully uses a particular teaching method for many years will not recognize another technique as being even more effective. Old habits interfere with new perceptions.

One creativity recommendation, “make the familiar strange,” encourages us to see common objects and situations in new ways, to overcome too-familiar perceptual features, and to look for new and different ideas and perceptions. Much creativity involves a mental transformation – the perception of new meanings, combinations, and relationships that depend upon overcoming perceptual barriers.

Rigid Attachment to the Experiential Self

The weight of experience can be framed within the evolutionary perspective discussed previously. The relative stability of the environment has favored stability in learning outcomes that can sometimes persist for longer than is adaptive in the light of changing circumstances. The body and brain, which hold these patterns, rely on perception and they are contributors to perception. Groopman and Prichard (2007) have observed that doctors who see a particular ailment many times in a row can misdiagnose a new patient because they more easily perceive the new syndrome as the pattern that has become familiar. Similarly, Duncker (1945) showed that individuals’ familiarity with the way that objects are traditionally used prevents them from solving problems that would require a novel use. He called it “functional fixedness.” A person sent to a company to solve a challenging situation may be able to behave most creatively only if he or she sees ways in which that company’s problem is different from problems he or she has encountered in other companies, and act on what he or she perceives, rather than acting automatically on problems he or she knows how to solve. Prior experience can move the Self system automatically, often with minimal processing by the mind (Bargh & Chartrand, 1999; Bargh & Ferguson, 2000). Reenactment of old behavioral patterns without regard for changed circumstances is at the root of what Langer calls “mindlessness” (1997), which is antithetical to creativity.

Seeing the past as the present, with the inability to see current circumstances as they are, is particularly difficult for those who suffer from posttraumatic stress disorder (PTSD). Van Der Kolk (2014) suggests that people with this disorder are often suspended in time: the threat of the past is still in the present. Their plight is an accentuated version of what many of us experience, as we carry out action based on what we perceive in our minds rather than in the world. Lightening the weight of experience can allow nonmindless, noncompulsive action on what is there, rather than what is in the past, or what we already know how to act upon.

3 Problem-Solving Set

The phenomenon of set occurs in many situations, reflecting a state of preparation for a particular input and response. In many circumstances where a speedy but accurate reaction is desired, sets serve positive purposes. To be prepared is to detect and react more quickly than would otherwise be the case. Contrast the reactions of runners who are given a “ready” signal to what would happen if the starting gun were fired unannounced. In a variety of experimental situations, warning signals have been found to facilitate responding. It is therefore somewhat curious that sets have been discussed largely in terms of negative effects in the context of problem solving.

The positive effects of set mentioned above are best understood in terms of general preparedness. A different aspect of set is to prepare the organism for a particular type of input or response, in effect, to bias the organism. For example, advance information that the next word to be presented is a flower name would speed reactions to rose but retard reactions to nonflower names. In the usual problem situation, initial perceptions or expectations are wrong, hence sets would be seen as impairing discovery of the correct response.

According to Woodworth and Schlosberg (1954), what Maier (1930) called “direction” was a kind of set. Maier (1940) distinguished between habitual directions and new directions; habitual directions would correspond to sets. Along with others, Maier believed that problem–solving efforts were guided by a direction or orientation that was not itself a habit but reflected (also) the influence of the problem situation. That is, the structure of the problem situation was held to affect the direction that would be followed. When problem solving requires doing something new, habitual directions must be overcome and new directions created. Indeed, Ruger (1910) had noted that some subjects stubbornly adhered to a particular line of attack despite repeated failure. From this perspective, any instance of insightful problem solving can be seen as breaking or disrupting a set. The usual meaning of problem–solving set as a predisposition or bias that exists when a problem is initially presented corresponds to Maier's concept of habitual direction. Most research on sets focused on negative transfer from prior experience to a problem situation.

One example of a negative set is functional fixedness, a term coined by Duncker (1945) to refer to the fact that an object with a strong customary function will not easily be seen as serving a different function. More specifically, problem presentations or prior experiences that emphasize the usual function were predicted to inhibit the use of an object in a more novel fashion.

For example, in the box–candle problem, the solution requires using a small box as a platform for a candle. In the solution, the box must be tacked to a wall, with the candle then mounted on the box. The critical manipulation concerned the function exhibited by the box when the problem materials were presented. Compared to presenting an empty box with tacks spread out on a table, presenting the box filled with tacks significantly impaired solution (Adamson, 1952). The essential idea was that the filled box emphasized the usual function of a box, thus blocking the idea of using the box as a platform. Although later research led to a subtle modification of the explanation, this version of functional fixedness has proven to be a reliable phenomenon.

Another variation of functional fixedness employed a transfer design. In a first phase, some subjects would employ an object in its usual fashion, for example, using a switch to complete a small electrical circuit. When subsequently given a problem requiring a novel use of an object, and having two suitable objects available, those subjects who had just used a switch in its usual, electrical function would avoid using it in the new, more novel function (e.g., Birch & Rabinowitz, 1951). The test problem was the two–string problem, arranged to require that a small object of suitable size and weight be tied to the end of one string. Doing this would allow that string to be swung back and forth (the “pendulum solution”) so that the problem solver, grasping the second string, could catch the swinging string and tie the two strings together. Two suitable objects were made available, a small electrical switch, and a small electrical relay. If people had just used the switch to complete an electrical circuit, they almost always used the relay to solve the two–string problem (and vice versa). The essential point was that using an object in its ordinary function blocked an unusual use of the object. An important additional finding was that the impairment was unidirectional, that is, unusual use of an object had no effect on its likelihood of being used for its ordinary function (e.g., van de Geer, 1957).

Research on functional fixedness focused on uses for particular objects. A quite different approach to problem–solving set concerned the dominance of a particular type of response to situations with a high degree of surface similarity. The typical procedure was to give subjects experience with solving a series of similar problems in “the same way,” then to test them with apparently similar problems that could (or must) be solved in a different way. The research thus focused on the acquisition of problem–solving sets and their subsequent (usually negative) effects on later problem solving.

Luchins's (1942) work provides an excellent example (later research showed his essential findings to be quite replicable). The basic procedure was to give people a series of “water jar” problems to solve. These problems involve giving subjects three jars with specified capacities (and, important, no gradations), and requiring them to indicate how they would measure (obtain) a specified amount of water. For example, given jars holding 21, 127, and 3 quarts, respectively, the subject would be asked to indicate how to measure out exactly 100 quarts. The solution was to fill the largest jar (B), and then to fill the second largest jar (A) once from B, then to fill the smallest jar (C) twice from B, which would leave the desired amount (100 quarts) in B. The “shorthand” description of the solution was thus B – A – 2C.

The basic procedure was to give subjects a number of “set–inducing problems,” all of which could be solved (only) by the B – A – 2C method. Following this experience, one or more test problems would be presented; these could be solved by either the B – A – 2C method or by a simpler procedure. For example, given “A = 14, B = 36, C = 8, obtain 6,” one solution is “36 – 14 – 8 – 8 = 6,” whereas a new and simpler solution is “14 – 6 = 8.” The basic finding was that people who had been exposed to the set–inducing series of problems were far more likely to use the (unnecessarily) long solution method. In other research, the test problems included “extinction problems” that could not be solved by the “set” method, in which case it was found that set–inducing experience delayed finding a solution (e.g., Gardner & Runquist, 1958).

Problem–solving sets, up to a point, were amenable to study from an associative–learning (behavioristic) perspective. That is, sets could be conceived (and experimentally investigated) as the product of a variable number of learning (conditioning) trials, as subject to extinction (from counter–set problems) and “spontaneous recovery.” A number of results fit this scenario, but there were also discordant findings that may well have contributed to the emergence of the “cognitive revolution” of the 1960s. The attempt to extend associative–learning principles (à la Hull, and Spence, primarily) to “higher processes,” required adherence to certain conditioning principles. Therefore, Kendler, Greenberg, and Richman's (1952) finding that problem–solving sets were stronger after massed training, compared to distributed training, posed a problem because the typical learning result was that distributed practice was superior. Partial reinforcement (of set solutions) should have produced greater resistance to extinction, but contradictory findings were obtained (e.g., Mayzner & Tresselt, 1956). To maintain the distinction between habit strength and drive, Maltzman (e.g., 1955) argued that sets established by instructions (akin to drive) should be more transient than those established by training (habit strength), but results were at best inconclusive (e.g., Maltzman & Morrisett, 1953). The difficulties encountered in attempts to extend associative ideas to more complex behavior might have smoothed the way for a new theoretical orientation.

DMs as Cognitive Strategies

Progress in the understanding of DMs requires a conception of how DMs accomplish their ego-protective function. An important observation that provides a clue for an answer is that beyond their application for resolving intrapsychic conflicts, DMs are widely used in everyday life as strategies for the successful performance of simple or complex cognitive tasks. Most DMs are likely to remind us of highly similar procedures occurring in contexts that are unrelated to ego defense. Thus, some DMs resemble strategies applied by children before they are assumed to have an ego that may stand in need of defense, for example, ignoring and immediately forgetting a parental restriction (i.e., repression), or assuming that a beloved doll is hungry and trying to feed it (i.e., projection). Indeed, projection remains for a long time an important means for understanding of engines and other appliances (e.g, “Look, the fireplace must be very hungry, it eats the wood so fast”). It is however complemented by introjection that also figures often in children’s attempts to comprehend how things work. A famous example was provided by Piaget (1951) who described how a child understood the mechanism of the matchbox only after it opened and closed its mouth a few times.

Most impressive are examples demonstrating the frequent use of defense-like strategies in different domains, notably problem solving, geometry, logic and regular daily functioning. For example, methods for solving mathematical problems include decomposing the complex problem into its constituent parts and working on isolated parts one at a time (i.e., isolation), or focusing first on a simpler problem (i.e., regression or displacement) (Polya, 1954, 1957; Wickelgren, 1974). Also in working on non-mathematical problems it is often necessary “to repress” habitual strategies so as to overcome functional fixedness; “to isolate” strategies previously applied for solving problems of another kind and “to project” them onto the presently given situation; “to regress” to an earlier stage, perhaps “undoing” something of what has already been done; raise the problem onto a more theoretical, hence higher level, thus exercising sublimation; and even to engage in reversal by trying the contrary of that which has previously been attempted. In the domain of logic we encounter the procedures of disproving a thesis by driving it ad absurdum (i.e., exaggeration) or by showing that reversing it ends up in a contradiction (i.e., reversal). In the domain of geometry we deal with projection into other dimensions (i.e. projection), reversals (e.g., of symmetry) and rotations (i.e., displacement).

More familiar examples concern everyday life. Thus, we naturally expect others to know what we know (Nickerson, 2001) and behave as we do (i.e., projection). Or, we may have become so proficient at ignoring disturbing stimuli that we no longer hear the usual traffic noise (i.e., repression). Again, if irritated by the recurrent misunderstanding of journal editors, we may decide to write a book (i.e., reaction formation). Or, when Beethoven lost some coins he wrote the piece “Die Wut ueber den verlorenen Groschen” (The anger over the lost nickel) that exemplifies sublimation, and perhaps acting out too.

These examples indicate that DMs are essentially basic and often applied cognitive strategies. Yet there are evident differences between the use of these strategies as DMs and in general. The most important difference is that as DMs they are much less amenable to conscious control, and hence are also far less flexible and less situation-relevant. Accordingly, we suggest that DMs are cognitive strategies that have undergone a transformation as a result of their specialized use. Thus, they become DMs under three major interrelated conditions: (a) when their function is limited primarily to serving intrapsychic needs in general or to resolving conflicts, such as those between the superego or ego and the id in particular; (b) when they are used without conscious control; and (c) when they have undergone schematization resulting in reduced flexibility and variation.

Thus, DMs may be defined as cognitively-based schemata functioning mainly in the intrapsychic sphere, especially in conflict resolution. Yet, the definition of DMs as cognitive strategies does not limit their manifestations only to the cognitive domain. DMs may affect perceptions (Shervin, 1995), emotions (Kwon, 2002), verbal behavior (Barett, Williams, & Fong, 2002), interpersonal behaviors (Cramer, 2002) and physiological responses (Derakshan & Eysenck, 2001) no less than cognitions (Borton, 2002). Any theory of DMs would need to account for the broad psychological impact of DMs.

A Limited, Common Set of Gestures and Controls Can Be Emotionally Expressive

Each player in Way shared a limited set of gestures, movements, and articulations, including moving right and left, moving arms up and down, giving a “shout” noise, and jumping. There was no verbal or written communication allowed. For example, players could lift their avatar’s arm to signal pointing to express where the other player should go, but they could not speak or write further instructions. This sometimes led to misinterpretation—for example, was the other player’s avatar pointing upward because the player needed to look upward, move their arm upward, jump upward, or something else? However, most of the time, after using trial and error, as well as the game’s context and their previous knowledge, the participants were able to understand what each other was expressing. Limiting the movements enabled the players to focus on quickly signaling and observing each other, and then move their own avatar accordingly. Players repeated their signal when necessary, or sometimes tried other movements. Because the gestures were based on universal human body movement, anyone, regardless of their language or literacy ability, could play the game and communicate with each other.

Limited gestures also affected the ways a player could express their emotional state—and could potentially affect one’s interpretation of the other’s state. For example, repeating gestures rapidly and the concomitant “shout” done by the partner suggested to Participant Three that the other player was frustrated and angry. Instead of looking for alternate solutions, Participant Three seemed overwhelmed by the feeling of letting the other person down and their own frustration, and they were unable to find the correct solution. The partner shut down the game, further confirming to Participant Three that the partner was angry. However, since we do not ever meet or know who the partner was, there is no way to confirm that the emotions Participant Three interpreted are accurate. It is possible that Participant Three overestimated the intensity of the partner’s negative emotions, which contributed to his functional fixedness and inability to find a new solution to the problem he was facing. Including only a “shout” as an audio version of a negative emotional expression limited the emotional interactions—there were no sound files in Way that expressed happiness or confusion, for example. Despite the limited interactions, participants did express aloud a variety of emotions that they were feeling, or that they believed their partner was feeling, though with this limited sample, they seemed to focus more on expressing the negative emotions.

That said, there was an opportunity for participants to express happiness, sadness, or confusion by clicking the relevant emoticon buttons on the game’s interface, which would have signaled their avatar to gesture. However, the researchers are not aware that any of the participants ever pressed these buttons. In addition, the researchers are not aware that any of the participants identified that the partner had hit any of those emoticon buttons. There are a few possible reasons for this. First, the participants were too focused on playing the game and trying to “listen” to what the partner’s avatars were communicating. Second, the emoticons were not emphasized during the tutorial. Third, the emoticon buttons were not involved in the actual gameplay. Pressing or not pressing them was not relevant to the actual problem-solving or tasks in the game, whereas the other gestures (pointing, jumping, moving) were necessary to complete the game. The “shout” would automatically happen if a player pressed down, which was necessary for some of the movements. Fourth, it may not have been clear that the other player had pressed an emoticon, as the only indication that it was pressed seemed to be based on how the partner’s avatar responded. The avatars did not have facial expressions, and were only able to gesture with their arms, legs, and head. Also, the “happy face” gesture was not very distinct from the “resting gesture,” in that the arms are just outward a little more (see Figure 12.6).

The confusion and sadness postures may have also not been obvious if the player’s avatar was moving around, jumping or otherwise mid-activity. Finally, the act of actually using the mouse to hit the emoticon buttons was disruptive to the game play, in that it was harder to hit it if you were in the middle of jumping, gesturing, or doing another movement as part of the game.

Designers who are creating environments and activities that will be used by those who do not speak the same language should consider how they could encourage participants to use a shared set of universal expressions, and to ensure that misinterpretations of emotional states are minimized. The designers of Way clearly thought that emotion communication was important, but they did not integrate it enough into the gameplay. It also should have been more obvious that an emotion was being communicated. Having the avatar make bodily gestures may not be enough in an online game environment to clearly express one’s emotions, particularly because the avatars did not have facial expressions as well.

V Discussion

Semantic knowledge is organized such that it affords meaningful and adaptive inferences (e.g., apples and oranges are fruit and therefore can play similar functional roles). The studies reported in this chapter show that such object-based inferences affect how people solve problems, transfer previously learned solutions to novel problems, or judge similarity. Put in this way, the present findings are not very surprising—it is not very surprising that people prefer to compare or combine apples and oranges rather than apples and baskets. What is quite surprising, however, is that such effects have been overlooked by researchers who study higher order cognition.

It is not that researchers failed to notice that object-based inferences affect reasoning. A classic example of such effects would be Duncker’s (1945) work on “functional fixedness,” whereby the functional role of a box as a container pevented people from using the box as a platform on which they could mount a candle. In a more recent example, Kotovsky, Hayes, and Simon’s (1985) study compared people’s solutions to two versions of the Tower of Hanoi problem. They found that subjects had little difficulty placing a large disk on top of a small disk, but when the disks were labeled “acrobats,” subjects refrained from letting a large acrobat jump on the shoulders of a small acrobat. Unfortunately, such demonstrations were merely added to the list of studies showing that superficial aspects of content and phrasing lead to errors, or make some problem isomorphs more difficult than others. They did not initiate research that looks for regularities in the way people select, or adjust, their reasoning tools (e.g., comparison vs integration) to semantic distinctions they deem important (e.g., functional symmetry vs asymmetry).

A line of research that, in its gist, is probably the closest to the studies reported here looks for regularities in the way semantic knowledge affects reasoning about conditional syllogisms (e.g., Cheng & Holyoak, 1985,1989; Cheng, Holyoak, Nisbett, & Oliver, 1986; Cosmides, 1989; Cosmides & Tooby, 1994; Cummins, 1995; Gigerenzer & Hug, 1992). This work has shown that different content instantiations of the material implication (“if p then q”) induce reasoning rules of distinct pragmatic and/or social schemas. According to Cheng and Holyoak (1985), who initiated this line of research, people apply different reasoning rules to formally isomorphic statements such as “if there are clouds, then it rains” or “if you drink beer, then you must be at least 21 years old.” In the first case, people are guided by knowledge that clouds are a necessary albeit insufficient cause for rain (e.g., causation schema), whereas in the second they are guided by knowledge that drinking age is established by a law that might be disobeyed (i.e., permission schema).

As in the case of arithmetic operations (Bassok et al., 1997), the rules of formal logic are semantically compatible with the rules of some schemas but not others. For example, Modus Tollens (“if not q then not p”) is compatible with the rules of the permission schema (e.g., if you are under 21 [not q], then you are not permitted to drink alcohol [not p]). However, this rule sometimes conflicts with the rules of the causation schema (e.g., if it doesn’t rain [not q], it does not necessarily follow that there are no clouds [not p]).Cheng et al. (1986) found that, much as the mathematically sophisticated subjects in Bassok et al. who sometimes “failed” to construct addition or division word problems for semantically incompatible object sets, subjects who were trained in evaluating the validity of conditional syllogisms committed “logical errors” on test problems that induced the rules of incompatible pragmatic schemas.

Cheng et al. (1996) argued convincingly that adherence to semantic and pragmatic constraints (i.e., content effects) protects people from arbitrary and anomalous conclusions. In fact, the results of Bassok et al. (1997) strongly suggest that, when application of formal rules conflicts with people’s semantic and pragmatic knowledge, they may prefer arriving at reasonable and logically invalid conclusions to arriving at logically valid but anomalous conclusions. Unfortunately, the notion that abstraction of structure from content is a mark of intellectual achievement appears to be so appealing that even sensible responses to semantic constraints may be classified by researchers as reasoning errors. The explanatory parsimony that is implied by this notion may also explain why researchers prefer to construct and test content-independent accounts of reasoning. While testing such accounts, they typically average people’s responses to stimuli that differ in content, viewing such responses as measurement errors that obscure basic processing regularities (see Goldstein & Weber, 1995, for an insightful discussion and historical analysis of this view). This practice would most probably average out the very effects that were the focus of the studies described in this chapter.

Object-based inferences that reflect adherence to semantic and pragmatic distinctions implied by content are inherent to rather than a deviation from normal processing. Also, such inferences do not indicate poor understanding, lack of maturity, or insufficient cognitive resources. Hence, effects of object-based inferences on reasoning do not fit well the “hidden-treasure” or any other variant of the faiure-of-abstraction-from-content metaphor. Instead, such effects seem to suggest a “toolbox” metaphor, whereby people attempt to find the best fit between their processing tools and the constraints implied by the stimuli they encounter. I strongly believe that, by adopting the toolbox metaphor, researchers are likely to discover interesting regularities in the way people adjust processing to their highly organized semantic and pragmatic knowledge.

Team Conflict

Research on team conflict suggests that it can be distinguished by whether it is task (cognitive conflict related to the task) or relationship (socioemotional conflict as a result of interpersonal disagreement) based. However, the two are often related and conflict in one area can spill into another (Jehn, 1997). It has been suggested that task-related conflict can be particularly important for enhancing creativity and innovation, as different views are discussed, differences in approaches can be resolved, and multiple points of view can be incorporated (Kurtzberg & Amabile, 2001; Mannix & Neale, 2005). A meta-analysis by Hulsheger et al. (2009) found non-significant results for the relationship between task or relationship conflict and team creativity and innovation, however, the number of studies used was relatively small (6 for relationship conflict and 13 for task conflict).

Carnevale and Probst (1998) in a series of studies using Duncker’s functional fixedness (candle) task and Rosch’s categorization task manipulated participant expectations regarding the team. Some participants were led to believe that teams would be cooperative whereas others were informed that teams would be in conflict. Results indicated that when participants expected conflict they were less likely to develop novel solutions and developed more restricted categorizations, supporting the notion that conflict can hinder creativity. Similarly, Mortensen and Hinds (2001), using product development teams, found that both relationship and task conflict were negatively related to creative performance.

Furthermore, a study by De Dreu (2006) examining the impact of task conflict on creativity indicated that task conflict could facilitate as well as inhibit team creativity. Specifically, De Dreu found a curvilinear, inverted U relationship between task conflict and creativity in postal service teams. A moderate level of task conflict facilitated creative performance, but too much or too little task conflict stifled creativity. Such findings could also be partially explained by the work of Jehn and Mannix (2001), which indicate that task conflict breeds interpersonal conflict. In other words, high levels of task conflict may increase interpersonal conflict, which in turn, hinders creative performance.

Jehn (1995) evaluated the effect of relationship and task conflict on team performance and satisfaction. In addition, Jehn investigated whether the effect of conflict was moderated by the type of task the team performed, routine vs. non-routine tasks. The findings suggested that relationship conflict was detrimental, regardless of task. However, the findings regarding task conflict were more complex. When tasks were routine, task conflict was detrimental to group functioning. In contrast, when teams performed non-routine tasks, task conflict either had no effect or was beneficial for group functioning. Interviews conducted as part of the study revealed that when task conflicts in groups performing non-routine tasks were discussed openly, it resulted in re-evaluation of the tasks, ways to achieve team goals, and more creative ideas. Jehn also found a curvilinear relationship between task conflict and team performance for non-routine tasks. While this study focuses on performance in general, it is important for understanding the effect of conflict on team creativity and innovation as those tasks are non-routine. It is therefore possible that Jehn’s (1995) finding regarding non-routine tasks will extend to teams involved in creative tasks. The inverted U relationship found may also explain the fact that the meta-analysis by Hulsheger et al. (2009) did not find a significant relationship between task conflict and team creativity, as curvilinear relationships were not evaluated.

Chen (2006) provided more direct support for the role of task type in moderating the relationship between conflict and team creativity and innovation in a study that investigated the relationship between task conflict, relationship conflict, and creativity in R&D teams. Chen evaluated whether the relationship between conflict and creativity is dependent on the type of project the team was working on (service vs. technology). For service oriented teams, task conflict had no effect on team creativity, but relationship conflict was significantly and negatively related to team creativity. On the other hand, for teams focusing on technology projects, no effects were found for relationship conflict, but task conflict was strongly associated with increased creativity.

Similarly, the effect of conflict and group polarity has been found to depend on the task and the timing in the project (Kratzer, Leenders, & van Engelen, 2006). In a study using R&D teams, Kratzer et al. found that team conflict had an overall negative relationship with team creativity. However, further investigation found that the relationship between team conflict and team creativity differed based on the phase of the project. During project conceptualization, an inverted U shaped relationship was found, suggesting that moderate levels of conflict are beneficial for creativity. However, during later phases of the team creative process, conflict had a negative effect on creativity.

Research that has focused on strategies that teams use to manage conflict found that conflict management is important and can minimize the negative effects of conflict. Song, Dyer, and Thieme (2006) found that techniques viewed as more positive and effective, such as integration or accommodation were positively related to constructive conflict and negatively related to destructive conflict. Similarly, more negative and less effective approaches, such as avoidance, were related to conflict in the reverse. Conflict management techniques were found to be directly and indirectly related to team innovation. Further support for the role of conflict management techniques was found by Lovelace et al. (2001). In a study of 43 cross functional teams, task disagreement was found to be beneficial for team innovation when team members felt free to express their concerns and openly discuss these concerns.

The results summarized here paint a complex picture regarding the relationship between conflict and creativity and innovation. Whether task and relationship conflict have different effects on creativity and innovation is still open to debate. It is possible that the relationship between task conflict and creativity is curvilinear. It seems likely that the type of task also influences the effect of conflict, with more complex and technical tasks benefiting from task conflict, whereas interpersonal tasks are hindered by interpersonal conflict. Finally, it seems that when there are positive effects for task conflict, these effects are possible because teams have mechanisms to handle these conflicts such as high cohesion, trust, and open communication (Lovelace et al., 2001; Mortensen & Hinds, 2001).

Future research should specifically evaluate the possibility of a curvilinear relationship between task conflict and team creativity and innovation. The studies reviewed here also suggest a possible moderated relationship between conflict, especially task conflict, and project phase or time, which needs to be evaluated further. Additional research needs to evaluate the direct effect of conflict and possible mediators of other social processes such as trust and communication, as ways to ameliorate the detrimental effects of conflict. Further, the relationship between task and relationship conflict as well as their relation to team creativity and innovation must be evaluated. It is possible that a high degree of task conflict can directly create or contribute to relationship conflict, resulting in negative effects of task conflict. Finally, the role of conflict may be particularly important when teams are diverse, as team diversity may lead to increased conflict (Mathieu et al., 2008). It is important to identify whether different types of diversity lead to different types of conflict or may be mitigated by different social processes.

Psycholinguistics: Some Traditional Approaches

If indeed language is the prototypical form on everyday creativity then a central question turns to understanding the mechanisms whereby humans comprehend and produce language. The first half of the twentieth century was dominated by behaviorist theoretical approaches to language, exemplified by the linguist Leonard Bloomfield in the 1930s and the psychologist B. F. Skinner in the 1950s. This approach eschewed mental concepts and argued that language can be understood as stimulus-response patterns. Skinner, for instance, argued that syntax could be understood as a form of stimulus–response chains similar conceptually to the behavior chains one could train an animal to perform. The dominance of behaviorist approach led to almost half a century in which the representational and mental processes of concern to modern psycholinguistic inquiry were virtually nonexistent. One notable exception to the behaviorist trend during this time span is found in the mentalist work of Edward Sapir and Benjamin Lee Whorf (the so-called Sapir–Whorf, or linguistic relativity, theory). The basic claim is that a person's native language sets up a series of ‘lenses’ through which reality is understood and interpreted. A corollary of this approach is that the language used by a person (in a given cultural context) provides insights into the minds of a given linguistic community or culture. Initial psychological tests of the hypothesis did not support a strong version of the hypothesis, which claims that the structure of a language determines the contents of thought. However there is support that has emerged in recent years from the laboratory of the psychologist Lera Boroditsky and others for a weaker version, namely that different languages shape or predisposes one to think of space and time in different ways. The emphasis on the representational aspects of language, both as seen in the recent work on linguistic relativity and with Bickerton's approach to the evolution of language, have direct implications for our understanding of creativity. For instance, a classic block to creativity is functional fixedness, that is, the difficulty or inability to conceptualize events or objects in other than canonical ways. Psycholinguistic experiments have demonstrated that how one labels the objects presented for use in a productive problem-solving situation influences whether or not the problem is solved. Labeling objects as ‘box and tacks’ permits invites thinking of these as two objects, each of which can be used to solve the problem whereas labeling the same objects as ‘box of tacks’ invites participants to represent the objects as unified (and hence that the box cannot be used as an aid to creative problem solving). The theory on the one hand makes contact with the geneplore model of creativity and empirically-based computational models of problem solving in which the representation of a problem determines the likelihood that a solution will be found (and that creative leaps involve restructuring how one represents a problem) and the theoretical extension, not proven, that because different languages (of which there are about 5000 in use world-wide today) represent reality differently, some languages are better equipped to perform certain type of creative activities.

By the 1950s there was a paradigm shift away from behaviorism toward mentalist approaches in psychology and language alike. Prominent in this movement was the work of the linguist Noam Chomsky, who, in the late 1950s, influenced psychologists initially with a critique of Skinner's approach to language, especially with regard to Skinners attempt at explaining syntax as a chaining mechanism. Although Chomsky's specific theoretical arguments have undergone many changes since the publication in 1957 of his book Syntactic Structures, he has maintained an emphasis on grammar and on everyday linguistic creativity. Any model of language he argued must be capable of predicting which combination of words is felt to be acceptable and which unacceptable to an adult native speaker (labeled a ‘descriptive grammar’). Language is a paradigmatic case of human creativity given our ability to produce and understand novelty on a daily basis. Because natural languages are infinitely large, Chomsky argued that any acceptable descriptive grammar must be generative in the sense that from a finite set of rules one can describe the structure of the infinite set of sentences that are grammatically acceptable. In essence the goodness of a descriptive grammar is its ability to generate the sentences of a language and to associate each sentence with an appropriate grammatical description, a fatal flaw with the chaining mechanism proposed by Skinner (which Chomsky argued was a variant of a finite state grammar). An important evolution in his theorizing was the development of transformational-generative grammar, based on the insight that one cannot understand linguistic creativity or language competence by merely describing the structure of the utterances we produce or sentences we read (the surface or s-level); rather we have to consider also the thematic relations being described (a more abstract deep or d-structure). Given these two levels of representation we would then need some transformational process between these structures (thus, for instance, conserving the intuition that the same meaning can be expressed in multiple ways).

Although many aspects of Chomsky's approach has been modified or rejected by others over the years, the influence that he has had on the field has been monumental and many of his approaches to the nature of the human mind have direct relevance for discussions about human creativity. One such claim is for the biological basis for language in humans. Because humans become proficient in the comprehension and use of their native language surprisingly rapidly, Chomsky (and others) argue that linguistic competence is biologically based, and as such there must be universal grammar – the innate schema of initial assumptions which all humans bring to language learning, and on the basis of which they construct the grammar of their linguistic community (from among all the possible grammars that one could have learned). One possibility for this that Chomsky proposed in some of his later writings is that humans are born with a common set of principles each of which can be set to one or another parameters. The specific parameter settings are based on exposure to a given language, even the impoverished experiences that one faces in early development. To the extent that creativity is dependent on the presence of grammar, then creativity too is biologically based.

A second important claim made by Chomsky is for the modularity of the mind, in the sense that it consists of a set of autonomous processing units, each one with its own set of rules specialized for domain-specific information. Processing within a module is informationally encapsulated such that the workings of the module precede independent of activity in other units; the output of a module is usually available to general intellectual capacities. With respect to language, the argument would be that the processing of language occur independent of other intellectual activities and, within language itself, the processing of syntax proceeds independent of pragmatics or semantics. Employing the standard Chomskyan methodology of logically considering the intuitions about the acceptability of sentences, often those artificially constructed to make a point salient, one can easily show that a sentence such as “colorless green ideas sleep furiously” is intuitively understood as syntactically acceptable yet meaningless. Thus, one can conclude that the processes that lead to acceptable syntax is not based on whether or not the sentence is semantically interpretable, and, by extension, that syntax and semantics are separable.

An implication of a modular approach for our understanding of creativity is that, if correct, much of creative processing would reside in specialized processing units separable from language and, more generally, general intellect. Moreover, this position leads to the position that different facets of creative activity are each served by their own modules, with their own set of mental operations. Along these lines, the linguist Ray Jackendoff has suggested separate modules exist for music, language, and dance-related activities. There is some evidence as well that different domains of creative activity are mediated by different brain structures, a position consistent but not necessary for modularity. As extended further, a strict modular approach would be in line with multifactorial approaches to understanding creativity, such as the structure model proposed by Guilford, and inconsistent with theoretical models that posit a more general approach to creativity, such as through a generalized associate network (and implemented in such tests as the Remote Associate Test).