"Intelligences" vs. "Languages"

I was pleased to see The New York Times article by Siobhan Roberts about Stanislas Dehaene's claim that geometry is a unique human capacity just as is language (link here). He proposes, therefore, that humans have "multiple languages." When I originally developed “multiple intelligences theory” forty years ago, I used the analogy of eight separate computers (one for each intelligence). I was critiquing the notion of a single all-purpose computer which is the construct underlying the standard IQ view of intelligence. (In the vernacular, according to the traditional view, people are either smart, average, or dumb).

At the time that I developed "MI theory," building on the work of Chomsky, Fodor, and others, I was using “natural language” as my prototype intelligence. And within language, my focus fell on syntax—semantics, much less so, pragmatics not at all. 

Reading about Dehaene's use of the term "multiple languages" led me to blog (link here) that he might just as well have used the term "multiple intelligences." Dehaene responded that his use of the term language refers to a formal "computer-like language whose expressions correspond to human mental representations and whose minimal description length captures psychological complexity." He went on to say, "As far as I know, multiple intelligences are not so well defined."

When I reflect on my original list of seven intelligences, I acknowledge that the personal intelligences are not well-described in traditional computational terms. Selma Mehyaoui seems to be alluding to personal intelligences when, responding to Dehaene, she says that language is not just a matter of representation but also involves pragmatics. I also don’t think of bodily-kinesthetic intelligence in traditional computational terms, but as robots become “smarter” they may also reveal the extent to which the uses of the body can be captured in computational language.

However, I submit that four of the original list of multiple intelligences do indeed fit the Chomskian and Dehaene model of language: not only linguistic intelligence and spatial intelligence (Dehaene's geometry language), but also mathematical and musical intelligence.

Is the Brain a Computer?

In June of this year, Gary Marcus, an NYU professor and contributor to The New York Times, published a piece entitled "Face It, Your Brain Is a Computer". What follows is Howard Gardner's response to this article.

---

Notes by Howard Gardner

When I am describing my view of intellect,  I often contrast it with the standard view of intelligence. And I invoke a computer metaphor. The old view posits a single all purpose computer: if it computes well, you are smart in everything; if it compute poorly, well, you are out of luck—all cognitive doors are closed.

My view, in contrast, posits the existence of several computers.  Each computes a certain kind of information in a way appropriate to that computer.  And so the musical computer deals with sounds, rhythms, timbres,  harmonics, while the spatial computer deals with the arrangement of objects or movements in local or global space.  A corollary is that the strength (or weakness) of one computer does not entail similar or different evaluations of the strength of another computer. Person A can be strong in spatial and weak in musical intelligence;  person B can display the opposite profile.

Clearly, the invoking of computers is a metaphor.  No one believes that an IBM computer (or several) or a microchip (or hundreds) is literally located in the skull.  Rather, the argument between Marcus’ view, on the one hand, and my view, on the other, is whether it is more helpful to think of one all purpose computer, or several more specific and more dedicated computers.

An analogy may be helpful . We all learn about the world through our sense organs. But there is a big difference between the claim that all sensory organs work in basically the same way, and the claim that each sensory organ has evolved so as optimally to handle certain kinds of inputs in certain ways.  I find the latter view much more useful.

Marcus raises a broader question (“Does it make sense to think of the brain as a computer?”) and has a simple answer (“Yes it does”).  But as he himself points out, we now recognize different kinds of computer with different kinds of computations. MI theory simply extends this form of conceptualization to the variety of cognitive processes of which human beings are capable.