Defining and measuring human intelligence are probably more broadly researched and written about than any other topics in either psychology or education (Roberts & Lipnevich, 2012). The literature on intelligence consists of many thousands of studies, as well as countless books, chapters, and popular articles. This broad base of information reflects the reality that the topic of intelligence is extraordinarily complex and often interwoven with strongly held social, political, and educational ideologies. According to most traditional theories going back more than 100 years, intelligence includes a single unifying mental ability that underlies all specific types of abilities. In contrast, according to some contemporary and popularized views, intelligence is made up of several essentially separate mental abilities with little or no role played by a central underlying component. The debate has numerous implications for educational practice and policy. Many scholars and educators favor the perspective that there are different varieties of intelligence because the theory seems more egalitarian and optimistic than a focus on core mental ability (Hunt, 2011; Roberts & Lipnevich, 2012). However, reconciling theories positing multiple intelligences with the enormous body of existing intelligence research has proven to be difficult.
Interest in formally measuring human intelligence began with the work of Francis Galton (1869) whose early work on the potential heritability of mental ability laid the groundwork for a great deal of controversy that lingers to this day. In the more than 100 years since Galton’s work, researchers and scholars have faced chronic difficulty in achieving consensus about what intelligence actually is. The complexity of human thought makes it challenging to develop a definition that encompasses the core aspects of mental ability while satisfying the perspectives of a myriad of stakeholders from a variety of disciplines. Despite the varying definitions that have been proposed, a few common threads run though most conceptualizations of intelligence. Nearly all scholars who study intelligence agree that intelligence involves “the ability to reason, solve problems, think abstractly, and acquire knowledge” (Gottfredson, 1997: 93).
For the past 100 years, the prevailing models of human intelligence have emphasized a single central characteristic, referred to as general intelligence, that links and perhaps powers all other cognitive abilities. Galton (1869) was the first to propose that intelligence consists of a broad general component, and general intelligence was first identified statistically by Charles Spearman (1904; 1927). Spearman observed that people’s performance on cognitive test items tended to correlate with their performance on similar items. Perhaps few observers were surprised that test-takers who did well on one type of math item tended to do well on other math items. However, Spearman also found that performance on one type of task tended to predict performance on very different types of tasks. For example, people who did well on math items also tended to do well on items measuring language skills.
Spearman found that scores on a wide variety of cognitive ability tests were positively correlated: people who did well on one type of test tended to do well on others. In his first major publication on the topic, Spearman (1904) analyzed data from children completing a variety of sensory and cognitive tests. He found positive correlations between the scores on the various tests – even in cases where the tests were very different in nature. For example, he reported that scores on a test of simple auditory discrimination – the ability to identify two sets of tones as the same or different – correlated with tests of academic ability and common sense. Spearman hypothesized that correlations between measures of different sensory and cognitive abilities are due to a “common intellective function” (p. 272). He initially proposed this idea somewhat tentatively, noting that the idea of a unifying cognitive component was so radical that a great deal more corroborating evidence needed to accumulate before more definitive conclusions could be drawn.
Spearman invented a now well-known statistical technique called factor analysis that enabled psychometricians – those who measure psychological characteristics – to reduce a large number of measured variables to a smaller and more useful number of underlying elements. For example, one could begin with the responses from a large number of people to a large number of mental ability test items and determine whether there were broader common abilities underlying people’s performance on different types of items. This technique allowed Spearman and subsequent researchers to conduct a much more sophisticated evaluation of the cognitive components underlying performance on ability tests. Spearman observed that there was nearly always one main factor that emerged tying together all the diverse cognitive tasks – explaining the correlations between very different types of skills. He referred to this factor as general mental ability, or g. On intelligence tests, this general factor is represented by IQ scores. Spearman also noted that differences between people on the general ability factor did not account for all the differences in test performance. He therefore concluded that performance on any particular type of test is a factor of both general mental ability and some more precise ability or talent specific to the particular test.
To explain how a core general ability could affect performance on a broad array of diverse mental tasks, Gottfredson (1997) asserted that overlapping higher-order thinking skills underlie success across cognitive tasks regardless of the specific test content. She further states that the more complex the task, the more the ability to engage in higher-order thinking will affect performance. In a metaphorical illustration, Kaplan and Saccuzzo (2013) use the analogy of a power station providing electricity for a city. They explain that although some lights in some places are brighter than others, a change in the amount of power from the main station would affect the brightness of all the lights in the city. Similarly, greater or lesser g would affect the functioning of all specific mental abilities. Spearman himself (1927) likened g to a kind of mental energy that powers a variety of more specific abilities.
Over the 100 years since Spearman claimed the existence of a broad general factor of human intelligence and psychologists began formally measuring intelligence, researchers have conducted thousands of studies supporting the existence of the general ability factor. Two recent examples from this long history serve to illustrate the durability of the finding that some common factor underlies performance on different types of tests. In the first of these studies (Johnson, Bouchard, Krueger, McGue, & Gottesman, 2004), researchers had 436 American adults complete a total of 42 ability subtests from three test batteries. The subtests varied in terms of the skills emphasized (verbal knowledge, nonverbal reasoning, inductive reasoning, pattern recognition, etc.) and also in format (multiple-choice versus free response). Johnson and colleagues separately factor analyzed the subtests associated with each of the three test batteries. Consistent with hundreds of previous studies, they found a common factor underlying performance on each battery. Moreover, they found that the general factors underlying the three batteries correlated virtually perfectly with each other despite divergent test content and response formats. The researchers concluded that this pattern constituted the “most substantive evidence of which we are aware that psychological assessments of mental ability are consistently identifying a common underlying component of general intelligence” (p. 104). In a replication of their study, Johnson and colleagues (Johnson, te Nijenhuis, & Bouchard, 2007) analyzed data from 500 adults who had completed 46 separate ability tests from five batteries. All the tests were different from those used in their 2004 study. The researchers again found that the g components from the five batteries correlated highly with each other – again suggesting the existence of a single factor underlying performance both within and across diverse cognitive test batteries.
The statistical evidence for the existence of general mental ability might not be particularly compelling if it was limited to correlations between different types of tests. Just as there is a long history of research suggesting the presence of a general factor underlying human intelligence, there is likewise a great deal of research indicating the importance of this general factor in predicting positive outcomes in a host of domains. To name just a few, g is positively associated with income, health behaviors, longevity, and job performance; g is negatively correlated with criminal behavior (Lubinski, 2004). General mental ability is also more highly correlated with occupational level than any other variable (Jensen, 1986). With respect to job performance in particular, g is correlated with performance on jobs at all levels of complexity (Gottfredson, 2002; Schmidt & Hunter, 2004). Schmidt and Hunter also reported that g is associated with performance in job training programs because people with greater general mental ability learn more and learn faster. They also noted longitudinal data indicating that g is associated with later income even after researchers control for other variables such as family socioeconomic status, neighborhood characteristics, and school quality. In fact, general mental ability measured in childhood is positively correlated with income more than 30 years later.
Most researchers agree that specific abilities exist along with g, but research suggests that none approach g in terms of their role in a variety of life circumstances (Jensen, 1986; Gottfredson, 1997). For example, tests of specific abilities designed to predict performance on specific jobs tend to provide no useful predictive information beyond what is indicated by general mental ability (Schmidt & Hunter, 2004). That is, specific ability tests may measure both specific abilities and g, but it is g that is predicting job performance. Jensen cited further evidence that if it were possible to develop an ability test that did not measure general mental ability, the test would not predict performance in any domain and would therefore be useless. More conceptually, Traub (Gardner & Traub, 2010) asserts that to conclude that there is no general intelligence would mean claiming that the brain has “little or no executive capacity to direct and integrate the mind’s activity” (p. 46).
The preceding evidence on general mental ability provides the backdrop with which alternative models of human intelligence are compared. Gottfredson (2002) referred to g as “probably the best measured and most studied human trait in all of psychology” (p. 25). Reviewing many decades of research, Jensen (1986) reasserted that all varieties of mental tests – no matter how diverse the actual tasks – are positively correlated, which indicates that they are all – at least in part – measures of some common intellectual component. He referred to this pattern of correlations among tests as “about as inexorable as gravitation” (p. 305). Jensen further pointed out that the central g factor accounts for more variation in test performance than any other factor, and often for more variation than all other factors together.
Throughout the history of intelligence research there have been critics of the idea that human cognitive ability is characterized by a central common element. Thanks in part to the work of Galton (1869) and subsequent researchers studying the heritability of intelligence, the concept of g long ago acquired connotations of biologically determined and environmentally immutable notions of intelligence. As time went on, concerns about fairness and social justice grew – particularly regarding racial disparities in test performance – causing people to question the reality of g. Critics correctly claimed that g cannot account for all aspects of human cognitive ability and performance. This claim in itself does not contradict the evidence for the existence of g; Spearman himself (1927) emphasized that performance on any particular task is affected by both general and specific abilities. However, there is quite a distinction to be made between this defensible viewpoint and the claim that there is simply no such thing as general mental ability, or that g is not important except with respect to a narrow range of academic tasks – claims that often coincide with models positing independent types of intelligence (e.g., Gardner & Moran, 2006).
Among the many models portraying cognitive ability as consisting of multiple independent varieties of intelligence while downplaying the idea of general mental ability, a few in particular stand out. One of the earliest models was developed by Thurstone (1938), who proposed that intelligence was made up of thirteen separate abilities – although in subsequent models he reduced the number to seven. Later Guilford (1967) proposed a model of intelligence consisting of 120 abilities that operate independently. However, the multiple intelligence model that has been most influential was proposed by Howard Gardner (1983). Perhaps the most noteworthy difference between Gardner’s model and those of Thurstone and Guilford is that Gardner’s model includes several abilities that most researchers – and perhaps most laypeople – do not generally consider to be part of intelligence. This conceptual expansion is the source of much of the difficulty in reconciling multiple intelligence models with the concept of general mental ability.
Gardner defined intelligence 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” (1999: 33–34). He proposed seven distinct types of intelligence with the labels linguistic, logical-mathematical, spatial, musical, bodily-kinesthetic, interpersonal, and intrapersonal (Gardner, 1983). The first three abilities in this list are consistent with skills assessed on traditional intelligence tests, as is Gardner’s eighth intelligence – naturalistic – proposed in a later edition of the model (Gardner, 1999). The remaining intelligences consist of abilities that are not usually included in definitions of intelligence (see Table 1 for Gardner’s description of each intelligence).
Table 1 Gardner’s eight intelligences
| Linguistic | The “sensitivity to spoken and written language, the ability to learn languages, and the capacity to use language to accomplish certain goals,” demonstrated in the skills of those such as “lawyers, speakers, writers, and poets.” |
| Logical-mathematical | The “capacity to analyze problems logically, carry out mathematical operations, and investigate issues scientifically,” demonstrated in the skills of those such as “mathematicians, logicians, and scientists.” |
| Spatial | The “potential to recognize and manipulate the patterns of wide space as well as more confined areas,” demonstrated in the skills of those such as “navigators and pilots,” as well as “sculptors, surgeons, chess players, graphic artists, or architects.” |
| Musical | “Skill in the performance, composition, and appreciation of musical patterns,” demonstrated in the skills of those such as “composers, conductors, and musical performers.” |
| Bodily-kinesthetic | The “potential of using one’s whole body or parts of the body to solve problems or fashion products,” demonstrated in the skills of those such as “dancers, actors, and athletes,” as well as craftspersons, surgeons, bench-top scientists, mechanics and many other technically oriented professionals.” |
| Interpersonal | The “capacity to understand the intentions, motivations, and desires of other people and, consequently, to work effectively with others,” demonstrated in the skills of those such as “salespeople, teachers, clinicians, religious leaders, political leaders, and actors.” |
| Intrapersonal | The “capacity to understand oneself, to have an effective working model of oneself – including one’s own desires, fears, and capacities – and to use such information effectively in regulating one’s own life,” as demonstrated in “those who excel in introspection.” |
| Naturalistic | “Expertise in the recognition and classification of the numerous species of his or her environment,” demonstrated in the skills of those such as “hunters, farmers, and those who study the natural world.” |
Note: Content quoted from Gardner (1999: 41–43, 48; 2011: 126).
In developing his theory, Gardner (1983) drew on literature from a variety of disciplines, including psychology, biology, sociology, anthropology, and even the humanities. By his own acknowledgment (Gardner, 2011), he was not trained in psychometric principles when he developed his model of intelligence. This fact might be seen as a strength or a liability depending on one’s point of view. He also acknowledges that his choice to use the term “intelligences” was “primarily strategic” in order to garner attention for the model (Gardner, 2011: 128). Other researchers have suggested that Gardner’s work was in part motivated by his moral objection to what he perceives to be Western society’s emphasis on a narrow definition of what it means to be intelligent (Barnett, Ceci, & Williams, 2006).
Gardner (1983) identified eight criteria to determine whether a particular characteristic constituted a separate form of intelligence. These criteria include localization in a particular area of the brain as indicated by the potential for the ability to be destroyed in isolation by brain damage, and the existence of “prodigies” or “idiot savants” (p. 63) who display widely varying abilities – often to the extent that a single ability is exceptionally high while the rest are exceptionally low. Gardner has at times implied that all eight of the criteria must be met to demonstrate that an ability reaches the status of a separate intelligence (1983; Gardner & Moran, 2006), but has sometimes asserted that an intelligence must meet “all or a healthy majority” of the criteria (2006a: 7). Roberts and Lipnevich (2012) argue that such inconsistency serves to “strip the process of selection of its scientific rigor” (p. 44), and assert that most of the criteria are easy to meet so there is potentially no limit to the number of characteristics that could be identified as intelligences. They suggest that such a slippery slope of inclusion might mean that intelligences such as bodily-kinesthetic would need to be further divided into even more specific intelligences such as football, golf, and dance intelligence. To do otherwise, they suggest, would be to assume that someone who is good at football could have been just as good at dancing.
Gardner’s theory resonated with educators – many of whom were disenchanted with the concept and presumed implications of general mental ability (Waterhouse, 2006; Roberts & Lipnevich, 2012). Kincheloe (1999) advocated rethinking the concept of intelligence based on a more inclusive view of education that validates alternative forms of intelligence to those currently recognized by psychology. He proposed that doing so would increase inclusiveness by “admitting new members to the exclusive community of the talented” (p. 1). Hunt (2011) asserts that multiple intelligence theory was “an easy sell to educators” (p. 117) because it was so optimistic. According to Hunt, the theory is consistent with the way many people think about intelligence: that there are many separate types of ability and that everyone is good at something. The multiple intelligence model therefore seems more egalitarian than the general intelligence model. Roberts and Lipnevich note the appeal of being able to refer to a child who does poorly in math but is a good musician as having musical intelligence. The model became popular with both teachers and parents because it supported the assumption that “all children are special” (Lohman, 2001: 221). Lohman also points out the potential risk of equating different abilities. Having low musical or bodily-kinesthetic intelligence would be very different in terms of likely life circumstances than having low general intelligence.
Despite its popularity among educators, multiple intelligence theory has been the subject of a great deal of debate and criticism from intelligence researchers. Indeed, the idea that there are many distinct and independent forms of intelligence is difficult to reconcile with the large body of empirical research indicating the existence of general mental ability. Gardner does not deny that statistical analyses of cognitive tests reveal a central factor emerging from correlations between test scores, but he claims that this g factor emerges primarily because the tests share similar formats, and are all affected by a very limited array of abilities – specifically verbal and logical-mathematical skills (Gardner, 2006a). He also claims that g predicts little outside traditional school performance (Gardner & Moran, 2006). In response to such propositions, Lohman (2001) bluntly states that “even the most cursory examination of human abilities literature shows that every one of these claims at best overstates and at worst is simply false” (p. 221). Moreover, there has been little research directly investigating the validity of multiple intelligence theory (Gregory, 2011).
In perhaps the only attempt to directly test the validity of the multiple intelligence framework relative to the general intelligence model, Visser, Ashton, and Vernon (2006a) administered a variety of tests to 200 undergraduate and graduate students, university employees, and friends and relatives of the undergraduates. The participants ranged in age 17–66. The researchers used a variety of established tests to measure abilities corresponding to each of Gardner’s eight intelligences – two tests for each intelligence. To ensure that shared verbal demands would not inflate the correlations between tests, Visser and colleagues included nonverbal measures in their study. Some tests were completely nonverbal, and for others the verbal demands were so minimal as to preclude the possibility that they could lead to inflated correlations – particularly among such an educated sample. Many of the tests involved tasks beyond those required on traditional paper-and-pencil tests such as identifying routes on a map, folding paper for a spatial ability test, and performing physical dexterity tasks. The researchers also administered a well-established measure of general intelligence.
Visser and colleagues (2006a) found that most of the correlations between tests assessing traditional cognitive abilities akin to Gardner’s linguistic, spatial, logical-mathematical, naturalistic, and even interpersonal intelligences were positive and significant, and all five of these abilities were substantially correlated with a separate measure of general mental ability. Tests assessing abilities not typically considered part of intelligence because they are heavily affected by noncognitive factors – specifically musical, intrapersonal, and bodily-kinesthetic skills – were not significantly correlated with the test of general mental ability. However, a factor analysis of all 16 tests revealed a first ability component that accounted for far more variation between test takers than any other factor.
Visser and colleagues (2006a) concluded that the correlations between diverse tests, the emergence of the general factor, and the correlations of specific tests with a separate measure of general intelligence are inconsistent with multiple intelligence theory. They also emphasized that common verbal influences across tests could not account for the results because the tests minimized verbal influences; even the spatial ability tests – which were almost entirely nonverbal – were positively associated with g. This pattern is consistent with the established evidence going back to the work of Spearman a century ago that “A test’s relative standing on g could not be inferred from its superficial characteristics, such as the sensory or response modality involved, whether verbal or nonverbal, numerical or figural, paper-and-pencil test or performance test, or other formal features” (Jensen, 1986: 310). Jensen asserts that tests measure g to the extent that they require complex cognitive processing or mental manipulation – regardless of the specific nature of the test itself.
Not surprisingly, these conclusions were met with skepticism from Gardner (2006b), who criticized Visser and colleagues for using tests emphasizing skills traditionally thought of as cognitive. He stated that the spirit of multiple intelligence theory is to expand the definitions of cognition and intelligence. He further argued that the common factor linking various abilities might arise because the tests all in some way reflect a narrow set of cognitive skills emphasized in traditional schools. In a response to Gardner’s critique, Visser, Ashton, and Vernon (2006b) cited evidence that many tasks containing no academic content – such as putting blocks together to copy particular shapes – are highly associated with general mental ability. They went on to cite evidence that g is correlated with biological processes such as cerebral glucose metabolism. Researchers have also found that g is correlated with speed of neural transmission (McRorie & Cooper, 2004), clearly a factor not taught in schools. Perhaps most compelling, Jensen (1986) summarized research linking measures of average brain wave activity with general mental ability. He cited research using a paradigm where participants’ brain wave activity is measured via electrodes on the scalp while the participant sits in a reclined chair listening to random auditory clicks, making no voluntary responses whatsoever. Average brain wave measurements taken during this task are strongly correlated with general mental ability – despite the fact that the task requires no conscious problem-solving or cognitive response.
Researchers have also taken issue with Gardner’s (1983, 1999) criterion that separate intelligences reflect localized neural processing in specific brain regions, and his claim (Gardner, 2006a) that evidence for multiple intelligences is provided by the fact that brain damage can cause the loss of some skills and not others. Roberts and Lipnevich (2012) point out that most cognitive abilities cannot be localized to specific parts of the brain, referring to this fact as a “major problem with Gardner’s view” (p. 45). Waterhouse (2006) cites evidence that not only cognitive abilities, but even motor abilities such as walking and gesturing are not localized to specific regions of the brain but rather involve multiple brain regions. Barnett and colleagues (2006) further assert that the loss of certain abilities due to brain damage and the localization of certain abilities in the brain do not suggest that those abilities are different intelligences, nor does it disprove the existence of some central ability that drives performance on all cognitive tasks. White (2004) agrees, noting that the loss of specific functions due to brain damage demonstrates only that some physiological condition necessary to perform the function is absent – not that there are distinct types of intelligence.
It often appears that the contrasting conclusions reached by scholars agreeing that intelligence consists of a general ability and those favoring a model of intelligence consisting of multiple independent components are due primarily to differing perspectives on the nature of intelligence and the nature of scientific evidence. Hunt (2011) argued that Gardner’s approach to identifying support for his theory is heavily weighted by subjective reasoning and reflection, which differs from the more purely data-driven approach favored by most intelligence researchers. Indeed, Gardner (2006a) provides interesting anecdotal examples for each intelligence that would be compelling to some students of the theory but not to others. His opinion that “it is up to educators to decide whether ideas derived from, inferred from, or catalyzed from MI theory are useful to them” (Gardner & Moran, 2006: 229) appears to reveal a willingness to accept subjective impressions as evidence. Other researchers have argued that Gardner’s model is not based on statistical evaluation, but rather on Gardner’s subjective view of how human abilities are organized (Roberts & Lipnevich, 2012). Even Gardner’s definition of intelligence differs from traditional definitions centered on cognitive abilities in that it includes several largely noncognitive capacities. In some ways, those favoring g and those favoring multiple intelligences are not even talking about the same thing.
Intelligence scholars have often been harsh in their criticism of the multiple intelligence model, with many concluding that there is no evidence for the theory (Waterhouse, 2006; Hunt, 2011; Roberts & Lipnevich, 2012). One particular difficulty is that Gardner has not articulated specific ways that each of the intelligences could be measured. He is unapologetic about this fact, stating that multiple intelligence theory is a work of “scientific synthesis” (Gardner, 2006b: 505) that does not lend itself to traditional testing, but is revised as new findings from various disciplines emerge (Gardner & Moran, 2006). This approach renders the theory somewhat immune to direct empirical evaluation. Gardner and Moran go on to refer to paper-and-pencil measurement as an “intrusion,” advocating instead for “intelligence-fair” assessments where abilities are assessed as they are naturally expressed (p. 230). Lohman (2001) questions this assertion, pointing out that the types of assessment advocated by Gardner can assess knowledge and skills beyond those on standardized tests, but they overlap greatly with standardized tests in terms of what they measure, they are more time-consuming, and they are more expensive. Barnett and colleagues (2006) likewise question the wisdom of using broader, long-term ability assessments not only because they demand far greater resources, but because such unstandardized measures preclude appropriate comparisons of students or programs because they are vulnerable to so many subjective judgments and personal biases. Moreover, Gardner’s (1999) insistence on real-life assessment, coupled with his assertion that real-life tasks often require a combination of intelligences, would seem to make it difficult or impossible to satisfactorily assess the separate intelligences (Visser et al., 2006a).
Although Gardner has not involved himself with assessing multiple intelligences, he is not opposed to others attempting to do so (Gardner & Moran, 2006). However, most tests that have been developed to assess the intelligences are self-report measures rather than true ability tests. In other words, researchers collect data by having participants rate themselves on the various intelligences (e.g., Furnham, 2009). This method is problematic because self-report measures of ability tend to correlate only modestly with actual ability measures (Visser et al., 2006a). One study showed that participants’ self-ratings of their ability on each of Gardner’s eight intelligences correlated weakly with ability tests selected to assess the intelligences – with correlations ranging from 0 to .38 (Visser, Ashton, & Vernon, 2008). Such findings suggest that self-report measures are exceptionally imprecise indicators of actual ability.
Through his work on his multiple intelligence theory, Gardner has certainly made many important points about the nature of intelligence. Barnett and colleagues (2006) note that although Gardner has not provided sufficient evidence for his theory’s validity, his work has had the valuable effect of drawing attention to the topic of intelligence and the limitations of traditional models. He has emphasized the diverse nature of mental abilities, pointed out that no two people exhibit the exact same pattern of cognitive abilities, advocated that teachers should nurture a variety of student talents, and questioned the idea of selecting people for opportunities based only on a measure of general mental ability (Gardner, 1999; 2006a). Importantly, there is nothing inherent in the theory of general intelligence that conflicts with any of these concerns. Virtually all scholars of general intelligence from Spearman onward agree that performance on any mental task is affected both by general intelligence and specific abilities. They simply deny that the different abilities operate in isolation from a common factor of ability. Hogan (2007) notes that the emphasis on maximizing all students’ potential may be appropriate in education, but is not a sufficient basis for intelligence models.
There is no question that Gardner’s model of intelligence has been quite influential in educational settings. There are many programs in schools across the world – in both Western and non-Western countries – whose developers based their work on the multiple intelligence framework (see Visser et al., 2006a; Waterhouse, 2006; Chen, Moran, & Gardner, 2009). Despite the influence of the theory, however, evidence for the existence of intelligences that operate independently without the influence of a general intelligence factor is sorely lacking. Gottfredson (2002) asserts that g involves the ability to “reason, learn, and solve problems” (p. 27), which helps to explain why it is affects performance across domains and across the lifespan. She also notes that no one has been able to develop a meaningful ability test that does not measure general intelligence. Without evidence for the existence of intelligences that are independent of each other and independent of a unifying ability, developing new teaching strategies emphasizing alterative educational values could do students a disservice if real-life opportunities continue to demand traditional intellectual abilities such as language, math, and reasoning skills (Barnett et al., 2006). Unfortunately, claims that multiple intelligence theory is “a proven approach to education for the twenty-first century” (Hoerr, 2003: 94) are generally made with little or no systematic evidence to support them.
Gardner (2011) criticized “the psychometric establishment …who believed (and continue to believe) that they have the right to define intelligence, to determine how it is measured, and to resist efforts to pluralize the concept” (pp. 127–128). In stark contrast to Gardner’s perspective, Waterhouse (2006) suggested that the multiple intelligence model is easy to understand because it simply divides intelligence into separate components based on specific content, and therefore allows people to believe that they understand how human cognitive processes work even if the evidence is at odds with the theory. To date, there are many anecdotes and opinions, but little empirical evidence to conclude that human cognitive ability consists of many independent intelligences.