Return to the Table of Contents.
F E B R U A R Y 1 9 9 9
RITING as a scholar rather than as a layperson, I see two problems with the notion of emotional intelligence. First, unlike language or space, the emotions are not contents to be processed; rather, cognition has evolved so that we can make sense of human beings (self and others) that possess and experience emotions. Emotions are part and parcel of all cognition, though they may well prove more salient at certain times or under certain circumstances: they accompany our interactions with others, our listening to great music, our feelings when we solve -- or fail to solve -- a difficult mathematical problem. If one calls some intelligences emotional, one suggests that other intelligences are not -- and that implication flies in the face of experience and empirical data.
The second problem is the conflation of emotional intelligence and a certain preferred pattern of behavior. This is the trap that Daniel Goleman sometimes falls into in his otherwise admirable Emotional Intelligence. Goleman singles out as emotionally intelligent those people who use their understanding of emotions to make others feel better, to solve conflicts, or to cooperate in home or work situations. No one would dispute that such people are wanted. However, people who understand emotion may not necessarily use their skills for the benefit of society.
For this reason I prefer the term "emotional sensitivity" -- a term (encompassing my interpersonal and intrapersonal intelligences) that could apply to people who are sensitive to emotions in themselves and in others. Presumably, clinicians and salespeople excel in sensitivity to others, poets and mystics in sensitivity to themselves. And some autistic or psychopathological people seem completely insensitive to the emotional realm. I would insist, however, on a strict distinction between emotional sensitivity and being a "good" or "moral" person. A person may be sensitive to the emotions of others but use that sensitivity to manipulate or to deceive them, or to create hatred.
I call, then, for a delineation of intelligence that includes the full range of contents to which human beings are sensitive, but at the same time designates as off limits such valued but separate human traits as creativity, morality, and emotional appropriateness. I believe that such a delineation makes scientific and epistemological sense. It reinvigorates the elastic band without stretching it to the breaking point. It helps to resolve the two remaining struggles: how to assess, and what kinds of human beings to admire.
Once we decide to restrict intelligence to human information-processing and product-making capacities, we can make use of the established technology of assessment. That is, we can continue to use paper-and-pencil or computer-adapted testing techniques while looking at a broader range of capacities, such as musical sensitivity and empathy with others. And we can avoid ticklish and possibly unresolvable questions about the assessment of values and morality that may well be restricted to a particular culture and that may well change over time.
Still, even with a limited perspective on intelligence, important questions remain about which assessment path to follow -- that of the purist, the simulator, or the skeptic. Here I have strong views. I question the wisdom of searching for a "pure" intelligence -- be it general intelligence, musical intelligence, or interpersonal intelligence. I do not believe that such alchemical intellectual essences actually exist; they are a product of our penchant for creating terminology rather than determinable and measurable entities. Moreover, the correlations that have thus far been found between supposedly pure measures and the skills that we actually value in the world are too modest to be useful.
What does exist is the use of intelligences, individually and in concert, to carry out tasks that are valued by a society. Accordingly, we should be assessing the extent to which human beings succeed in carrying out tasks of consequence that presumably involve certain intelligences. To be concrete, we should not test musical intelligence by looking at the ability to discriminate between two tones or timbres; rather, we should be teaching people to sing songs or play instruments or transform melodies and seeing how readily they master such feats. At the same time, we should abjure a search for pure emotional sensitivity -- for example, a test that matches facial expressions to galvanic skin response. Rather, we should place (or observe) people in situations that call for them to be sensitive to the aspirations and motives of others. For example, we could see how they handle a situation in which they and colleagues have to break up a fight between two teenagers, or persuade a boss to change a policy of which they do not approve.
Here powerful new simulations can be invoked. We are now in a position to draw on technologies that can deliver realistic situations or problems and also record the success of subjects in dealing with them. A student can be presented with an unfamiliar tune on a computer and asked to learn that tune, transpose it, orchestrate it, and the like. Such exercises would reveal much about the student's intelligence in musical matters.
Turning to the social (or human, if you prefer) realm, subjects can be presented with simulated interactions and asked to judge the shifting motivations of each actor. Or they can be asked to work in an interactive hypermedia production with unfamiliar people who are trying to accomplish some sort of goal, and to respond to their various moves and countermoves. The program can alter responses in light of the moves of the subject. Like a high-stakes poker game, such a measure should reveal much about the interpersonal or emotional sensitivity of a subject.
A significant increase in the breadth -- the elasticity -- of our concept of intelligence, then, should open the possibility for innovative forms of assessment far more realistic than the classic short-answer examinations. Why settle for an IQ or an SAT test, in which the items are at best remote proxies for the ability to design experiments, write essays, critique musical performances, and so forth? Why not instead ask people actually (or virtually) to carry out such tasks? And yet by not opening up the Pandora's box of values and subjectivity, one can continue to make judicious use of the insights and technologies achieved by those who have devoted decades to perfecting mental measurement.
To be sure, one can create a psychometric instrument for any conceivable human virtue, including morality, creativity, and emotional intelligence in its several senses. Indeed, since the publication of Daniel Goleman's book dozens of efforts have been made to create tests for emotional intelligence. The resulting instruments are not, however, necessarily useful. Such instruments are far more likely to satisfy the test maker's desire for reliability (a subject gets roughly the same score on two separate administrations of the test) than the need for validity (the test measures the trait that it purports to measure).
Such instruments-on-demand prove dubious for two reasons. First, beyond some platitudes, few can agree on what it means to be moral, ethical, a good person: consider the differing values of Jesse Helms and Jesse Jackson, Margaret Thatcher and Margaret Mead. Second, scores on such tests are much more likely to reveal test-taking savvy (skills in language and logic) than fundamental character.
In speaking about character, I turn to a final concern: the relationship between intelligence and what I will call virtue -- those qualities that we admire and wish to hold up as examples for our children. No doubt the desire to expand intelligence to encompass ethics and character represents a direct response to the general feeling that our society is lacking in these dimensions; the expansionist view of intelligence reflects the hope that if we transmit the technology of intelligence to these virtues, we might in the end secure a more virtuous population.
I have already indicated my strong reservations about trying to make the word "intelligence" all things to all people -- the psychometric equivalent of the true, the beautiful, and the good. Yet the problem remains: how, in a post-Aristotelian, post-Confucian era in which psychometrics looms large, do we think about the virtuous human being?
My analysis suggests one promising approach. We should recognize that intelligences, creativity, and morality -- to mention just three desiderata -- are separate. Each may require its own form of measurement or assessment, and some will prove far easier to assess objectively than others. Indeed, with respect to creativity and morality, we are more likely to rely on overall judgments by experts than on any putative test battery. At the same time, nothing prevents us from looking for people who combine several of these attributes -- who have musical and interpersonal intelligence, who are psychometrically intelligent and creative in the arts, who combine emotional sensitivity and a high standard of moral conduct.
Let me introduce another analogy at this point. In college admissions much attention is paid to scholastic performance, as measured by College Board examinations and grades. However, other features are also weighed, and sometimes a person with lower test scores is admitted if he or she proves exemplary in terms of citizenship or athletics or motivation. Admissions officers do not confound these virtues (indeed, they may use different scales and issue different grades), but they recognize the attractiveness of candidates who exemplify two or more desirable traits.
We have left the Eden of classical times, in which various intellectual and ethical values necessarily commingled, and we are unlikely ever to re-create it. We should recognize that these virtues can be separate and will often prove to be remote from one another. When we attempt to aggregate them, through phrases like "emotional intelligence," "creative intelligence," and "moral intelligence," we should realize that we are expressing a wish rather than denoting a necessary or even a likely coupling.
We have an aid in converting this wish to reality: the existence of powerful examples -- people who succeed in exemplifying two or more cardinal human virtues. To name names is risky -- particularly when one generation's heroes can become the subject of the next generation's pathographies. Even so, I can without apology mention Niels Bohr, George C. Marshall, Rachel Carson, Arthur Ashe, Louis Armstrong, Pablo Casals, Ella Fitzgerald.
In studying the lives of such people, we discover human possibilities. Young human beings learn primarily from the examples of powerful adults around them -- those who are admirable and also those who are simply glamorous. Sustained attention to admirable examples may well increase the future incidence of people who actually do yoke capacities that are scientifically and epistemologically separate.
In one of the most evocative phrases of the century the British novelist E. M. Forster counseled us, "Only connect." I believe that some expansionists in the territory of intelligence, though well motivated, have prematurely asserted connections that do not exist. But I also believe that as human beings, we can help to forge connections that may be important for our physical and psychic survival.
Just how the precise borders of intelligence are drawn is a question we can leave to scholars. But the imperative to broaden our definition of intelligence in a responsible way goes well beyond the academy. Who "owns" intelligence promises to be an issue even more critical in the next century than it has been in this era of the IQ test.
Howard Gardner teaches human development at the Harvard Graduate School of Education. Among his books are Multiple Intelligences (1993), Extraordinary Minds (1997), and The Disciplined Mind, to be published this spring.
Illustrations by Malcolm Tarlofsky
Copyright © 1999 by The Atlantic Monthly Company. All rights reserved.
The Atlantic Monthly; February 1999; Who Owns Intelligence?; Volume 283, No. 2; pages 67-76.