多元智能理论的科学性（加德纳答沃德豪斯女士的批评）

GARDNER AND MORANRESPONSE TO LYNN WATERHOUSE The Science of Multiple Intelligences Theory: A Response to Lynn Waterhouse Howard Gardner and Seana Moran Graduate School of Education Harvard University For a scholar, a fate worse than being criticized is being ignored. Waterhouse (2006) has done Howard Gardner the courtesy of reading much of the primary and secondary literature on multiple intelligences (MI) theory. Although the authors disagree with several of her in- terpretations and conclusions, we appreciate her efforts as well as the opportunity to respond. We have 2 main criticisms: (a) Waterhouse misunderstands and oversimplifies MI theory and (b) Waterhouse’s own line of argument undermines her claim that MI theory is not supported by the literature. This response reorients and clarifies for the reader the usefulness and impli- cations of MI theory with the goal of demonstrating why Waterhouse’s critique misses the mark in a number of respects. Gardner introduced multiple intelligences (MI) theory in the book, Frames of Mind, published in 1983. In arriving at his theory, Gardner combined the empirical findings of hundreds of studies from a variety of disciplines (see the extensive bibliography in the original book). Although he included psychometric and experimental psychology, he did not limit his base of support to just these disciplines. Rather, MI theory also encompasses cognitive and develop- mental psychology, differential psychology, neuroscience, anthropology, and cultural studies. Gardner’s major claim is that a description of individu- als in terms of a small number of relatively independent computational capacities is more useful to cognitive scien- tists, psychologists, and educators than a description in terms of an innumerable collection of sensory-perceptual modules, on the one hand, or a single, all-purpose intelli- gence, on the other. An intelligence is defined as a biopsychological potential to process information that can be activated in a cultural setting to solve problems or create products that are of value in a culture. This definition sets the stage for our first comment on Wa- terhouse’s (2006) account of MI theory. Her view fails to en- compass the several levels on which MI theory examines intelligences—as composites of fine-grained neurological subprocesses but not those subprocesses themselves, as biopsychological information processing capacities, and as the bases on which an individual can participate in meaningful activities in the broader cultural milieu. Indeed, Waterhouse misses this latter level altogether, except for her assertion that MI theory is a questionable basis for education, a key cultural institution. MI THEORY TAKES A MIDDLE ROAD Gardner has never claimed that MI theory represents “the” definitive description of human cognitive capaci- ties. Rather, he maintains that relatively independent yet interacting intelligences provide a better understanding of the variety and scope of human cognitive feats than do competing accounts. This debate of whether intelligence is a singular individ- ual quality or a plethora of components (Guilford’s 1967 structure of intellect model had 120!) has waxed and waned throughout the 20th century. Spearman (1904), Binet and Si- mon (1909/1976), L. M. Terman (1925) and L. W. Terman and Oden, 1947, and Wechsler (1958) paved the path for IQ, the “first factor” from paper-and-pencil IQ tests that has been correlated with other paper-and-pencil tests and linguistic and logical scholastic success—but far less impressively with real-world success (see Moran & Gardner, 2006). In- deed, the lack of extraordinary success among the children with high IQs in Terman’s 70+-year longitudinal study is re- markable in showing the limits of IQ as a conceptualization of intelligence. Yet, because IQ tests seem easy to administer EDUCATIONAL PSYCHOLOGIST, 41(4), 227–232 Copyright © 2006, Lawrence Erlbaum Associates, Inc. Correspondence should be addressed to Howard Gardner, Graduate School of Education, Harvard University, Larsen Hall 201, 14 Appian Way, Cambridge, MA 02138. E-mail: hgasst@pz.harvard.edu and score, and because they are well entrenched in the educa- tional psyche, they remain the standard in most schools, de- spite lack of agreement about what they actually measure and predict (see Neisser, 1998, for a critical discussion). At the other end of the spectrum are people who equate in- telligence with sensory-perceptual abilities, such as verbal, perceptual speed, number, word fluency, space/visualization, and mechanical acumen (Thorndike, 1921; Thurstone, 1938; Vernon, 1950). One advantage of this slant is that these abili- ties are easy to “see” experimentally: They can be manipu- lated by standard cognitive psychological and social psycho- logical research intervention paradigms. As such, this approach tends to be favored by psychologists. Both the IQ and the sensory-perceptual abilities foci continue today. Yet, they may be cases of the cliché of the man looking for his dropped car keys underneath the street lamp because that is where the light is. Intelligence as IQ or sensory-perceptual ability is easy to test, so many educators and psychologists understandably prefer these approaches. An MI approach demands a change of minds among re- searchers and educators: It requires an interdisciplinary per- spective, cultural sensitivity, and an interactionist-dynamic research methodology. The first two reasons support Gardner’s decision to incorporate anthropological studies and case materials from a variety of cultures in devising and revising his theory (see Gardner, 2006a). The third reason supports his decision to include developmental findings and to push for assessment criteria and environments that are in- telligence-fair, are individual-focused (rather than “average kid”-focused), and capture the often dynamic interactions among intelligences. The advantage of this approach is that it better explains the wide variety of “intelligent” performances among chil- dren and adults depending on level of training, context, cul- ture, and innate predisposition. An MI approach better ad- dresses the incongruities and imbalances of intelligent behavior not only between individuals but also within indi- viduals. Finally, an MI approach does not overprivilege the “average” person—rather, it makes room in the scholarly de- bate for experts whose intelligence profiles fit perfectly with a cultural domain; creators whose intelligence profiles are in- congruous with a cultural domain in a fruitful, surprising way; and savants and brain-damaged patients who exhibit a striking disparity among abilities. MI THEORY COMPRISES SEVERAL LEVELS OF ANALYSIS Where should “intelligence” be situated? At the modular level of specific neural processes, the middle level of coordi- nated intelligences, or the social level of how intelligences intersect with cultural domains? On the fine-grained end of the spectrum, each intelligence makes use of psychological subcomponents (e.g., face detection or object tracking). In his own work, for example, Gardner (1994; Gardner, Brownell, Wapner, & Michelow, 1983) dissected several as- pects of linguistic intelligence, including sensitivity to syn- tax, metaphor, and narrative. He also drew on others’ work to understand and formulate how these subcomponents com- prise an intelligence—in this case, linguistic intelligence. We are pleased by the extent to which this identification of subcomponents has been reinforced by findings in neurosci- ence (e.g., the discovery of many specific neural systems, mediating capacities like theory of mind, recognition of natu- ral kinds, understanding of self, understanding of others) and developmental studies (e.g., identification of core systems of numerical, linguistic, and causal reasoning). For the summa- ries of these findings, see Gardner 2006a and 2006b. At the broader end of the spectrum, intelligences interact with the opportunities and supports of social groups, such as profes- sions and vocations. Gardner’s work over the last 2 decades has focused on this interaction (Gardner, Csikszentmihalyi, & Damon,2001;Kornhaber,Krechevsky,&Gardner,1990).Social groups organize information into disciplines and domains (i.e., bodies of knowledge), toward which individuals can mobilize one or more intelligences to produce proficient and/or expert be- havior (Connell, Sheridan, & Gardner, 2004). As a result of this work—aswellas theworkofCsikszentmihalyi (1988),Hutchins (1980),Salomon(1993),andmanyothers—wehaverefashioned theconceptofintelligencesothatitcompriseswhattheindividual brings and what the cultural and social environments contribute to a particular cognitive performance. In the middle of this spectrum are the intelligences themselves. Based on the eight criteria outlined in chapter 4 of Frames of Mind, Gardner (1999) holds that there are eight intelligences, each oriented to a specific type of infor- mation: linguistic, logical-mathematical, musical, spatial, bodily kinesthetic, naturalistic, interpersonal, and intrapersonal. He has analyzed the potential of a ninth intel- ligence, existential, but is not yet convinced it fulfills all of the criteria. Note that information is not synonymous with sensory input. Rather, information comprises a collection of such inputs in any format that can be interpreted, under- stood, and made use of by the person (or, more precisely, by his or her computational capacities). MI theory also stresses that the interaction among these intelligences is important for understanding how people’s minds work. For example, the intelligences can be grouped together for various purposes (e.g., those that deal with ob- jects, with persons, with more abstract entities). One key modification of the theory entails two overarching intelligences profiles: searchlight and laserlike (Gardner, 2006a). Waterhouse (2006) regards these terms as additional intelligences, but they are not. Intelligence profiles describe the strength of intelligences relative to each other. Search- light profiles— especially characteristic of politicians and businessmen—involve a ready shifting among intelligences that are often of comparable strength. Laserlike profiles—es- pecially characteristic of artists, scientists, and schol- 228 GARDNER AND MORAN ars—demonstrate one or two powerful intelligences used in great depth that overshadow the other intelligences. Deployment of the Word Intelligence In addition to Waterhouse’s (2006) mischaracterization of the nature and scope of MI theory, her article includes several specific terminological and level-confusion errors. Consider, for example, the first paragraph of the article: 1. Gardner has never combined interpersonal and intrapersonal intelligence into a single intelligence. Interper- sonal intelligence processes information related to other peo- ple and interacting with them. Intrapersonal intelligence pro- cesses information about the self. They are clearly different. 2. Naturalist intelligence does not include empathy for natural things. Empathy is an emotional capacity not an in- formation processing capacity. Naturalist intelligence pro- cesses information related to distinguishing among natural and manmade objects, which is evolutionarily derived from the hominid capacity to recognize, group, and label distinc- tions among natural phenomena. 3. Gardner never described the searchlight and laserlike profiles as “additional intelligences.” As noted previously, these conceptualizations describe the ways in which the sev- eral intelligences interact and are deployed. In later passages, Waterhouse (2006) confuses an intelli- gence, which is an individual’s biopsychological informa- tion processing capacity, with a skill, which is a cognitive performance that includes the supports and constraints of the environment (see Fischer, 1980). In our terms, she col- lapses the middle intelligence level with the broad cultural domain level. Further on, she characterizes the object and location perceptual neural pathways as additional “place” and “object” intelligences—confusing the levels of sensory input and subcomponents with the middle level of intelligences. As an additional example, she characterizes Kahneman’s decision-making styles as still more intelligences—the “intuitive” and the “delibera- tive”—which confuses the individual intelligences with profiles or means of deploying those intelligences. These errors of detail—misusing MI theory terms and using em- pirical findings at the inappropriate level of analysis—cast doubt on Waterhouse’s (2006) analyses and conclusions. MI THEORY AS AN EDUCATIONAL “PROGRAM” Waterhouse (2006) castigates educators for building prac- tices on an “unproved theory.” We would turn this accusation on its head. It has taken a century and many millions of dol- lars to bring IQ testing to its current, not especially impres- sive (although highly reliable!) status. IQ-based education, which Waterhouse lauds, has yielded no genuine implica- tions for classroom instruction or the improvement of performances, only feelings of pride or shame on the part of test-takers or their families. It should be stressed that Gardner himself has never put forth an educational recipe growing out of MI theory. At most, he has indicated some general implications—individu- alizing education, approaching topics through multiple entry points—that are consistent with the theory. Waterhouse’s (2006) assertion that educational applications of MI theory are harmful is without merit. This claim is offered without evidence and by ignoring considerable counterevidence. With colleagues, Kornhaber (Kornhaber, Fierros & Veenema, 2004) studied 41 schools that had used MI-in- spired practices for several years. She documents numerous ways—quantitative as well as qualitative—in which these schools and their students have benefited. In our view, it is up to educators to decide whether ideas derived from, inferred from, or catalyzed from MI theory are useful to them. RIVAL VIEWS OF SCIENCE We are puzzled when Waterhouse (2006) asserts that MI theory is not empirical, is not based on empirical findings, or has no support in the empirical literature. The theory originated entirely from empirical findings. In our view, Waterhouse embraces a naïve view of science, which con- temporary philosophers or historians of science rarely hold. As put forth by Diamond (2005), science is misrepresented as “the body of knowledge acquired by performing repli- cated experiments in the laboratory” (p. 17). Rather, as Dia- mond goes on to point out, science is (and has been from its origins) a much broader enterprise: the acquisition of reli- able knowledge about the world. Many fields— population biology, astronomy, epidemiology, geology, and paleontol- ogy, to name a few—proceed by incorporating and contextualizing relevant empirical findings. That is, science progresses not only through experimentation but also by synthesizing the experimental, observational, and theoreti- cal work of others to build a foundation for future research. MI theory was put forth deliberately as a work of synthe- sis: a work that organizes and integrates large bodies of em- pirical work from a variety of disciplines. Rather than utiliz- ing only the experimental and psychometric psychological findings, which were the dominant approach to intelligence at the time, Gardner cast a wide net that included neurosci- ence, cognitive science, anthropology, and evolutionary sci- ences. This broader view allowed Gardner to reconceptualize intelligence(s)—that is, to understand the concept in a new light free from the constraints of a single disciplinary lens. Gardner continues to assess and integrate new empirical findings from these and other various domains to refine MI theory (e.g., Gardner, 2006a). This ongoing process of analy- sis and synthesis has resulted in the addition of intelligences, RESPONSE TO LYNN WATERHOUSE 229 the conceptualization of intelligence profiles, and the modest educational guidelines described previously. In each of these modifications, the changes have been guided by the usual considerations of logic, clarity, and usefulness expected in scientific theorizing and synthesizing. At the most fundamental level, we believe that Water- house (2006) does not acknowledge or understand the enter- prise in which Gardner has been engaged: synthesizing the empirical findings of others. Yet, there is a paradox: Synthe- sis is precisely the enterprise of her own article—assembling and making sense of a body of literature. Moreover, in her own synthesizing efforts, she makes use of categories that are precisely at the same level as Gardner’s intelligences—she readily invokes emotional intelligence, musical aptitude, spatial skill, and the like. Therefore, her own presentation un- dermines the heart of her critique. ON TESTING Waterhouse (2006) berates Gardner for not testing MI theory, and she quotes others who claim that MI theory has not been tested. As a work of synthesis, MI theory does not lend itself easily to testing through paper-and-pencil assessments or a one-shot experiment. Rather, it is repeatedly assessed and re- formulated as new empirical findings from a variety of disci- plines are analyzed and integrated. Theories such as evolu- tion or plate tectonics or MI develop through the continuing accumulation of evidence, which makes the theory more or less plausible, more or less relevant for further research, and more or less useful to practitioners. It is possible to develop a set of tests that purport to exam- ine the core components of each intelligence proposed in MI theory. Were such tests to be developed, one could then ex- amine the correlations among the subcomponents of each in- telligence and correlations (or lack thereof) across the tests that purport to examine each separate intelligence. However, in her adherence to an IQ view of intelligence, Waterhouse (2006) claims that intercorrelations among subcomponents refute rather than support an MI perspective. She does not recognize that such correlations also could imply the con- cepts of searchlight and laserlike intelligence profiles—in other words, how the intelligences interact. Infact, someuseful testshavebeendeveloped. In the1980s, Feldman, Krechevsky, Chen, Gardner, and other colleagues created“intelligence-fair”assessmentsof the intelligences for preschool children. An intelligence-fair assessment means it assesses a particular intelligence in its most natural mi- lieu—for example, the bodilykinesthetic intelligence through movement activities or the interpersonal intelligence through social interaction—rather than through ersatz paper-and-pen- cil measures. These researchers piloted the Project Spectrum measures on select groups of young children in the Boston area,whichyieldeddistinctive intelligenceprofiles inchildren asyoungas3or4years.These findingsaredescribed inanum- ber of articles and in three books (Gardner, Feldman, & Krechevsky, 1998a, 1998b, 1998c). Educators in the United States and around the world—from Scandinavia to East Asia—continue to use these materials. This project taught Gardner another key lesson: De- veloping tests is an expensive and time-consuming process whose final results may well be misused. Even as the IQ test yields premature, all-purpose descriptions of “smart” and “dumb,” a Spectrum assessment unthinkingly might lead to labeling a child as “music smart” or “interpersonally stupid.” Gardner would prefer to spend more resources helping learner