Kagan's Articles - FREE Kagan Articles

Dr. Spencer Kagan

Trialogue: Brain Localization of Intelligences

Dr. Spencer Kagan, Dr. Howard Gardner, and Dr. Robert Sylwester

To cite this article: Kagan, S., H. Gardner & R. Sylwester. Trialogue: Brain Localization of Intelligences. San Clemente, CA: Kagan Publishing. Kagan Online Magazine, Fall 2002. www.KaganOnline.com

This quarter, rather than publishing my usual theory piece, you will find a "trialogue," a three-way correspondence between Howard Gardner, Robert Sylwester, and myself.

Some years ago, I wrote to Howard Gardner that I was troubled by his emphasis on brain localization as the most important criterion of an intelligence. My letter and Howard's response were published in a very early edition of this magazine (reprinted at the end of this trialogue). Since then, there has been an explosion of research utilizing active brain imaging, giving us an increasingly differentiated picture of what is happening in the brain in real time while people engage in various types of mental activity. Quite recently, a very carefully controlled study in Cerebral Cortex by Parsons and Osherson revealed amazingly little common brain activation when participants did deductive v. probabilistic logic problems. This, in my view, supported my doubts about declaring the logical/mathematical intelligence (or any other intelligence) based on brain localization. When I shared this information with Howard Gardner and Robert Sylwester, what emerged was a engaging and instructive Email correspondence, a trialogue, which is reproduced below. I hope you find it as thought-provoking as we did.


From: Spencer Kagan
Date: Friday, July 5, 2002
To: Howard Gardner
CC: Robert Sylwester
Subject: Brain Localization of the Intelligences

Dear Howard,

Some years ago we had an e-mail discussion regarding brain localization of the intelligences. If you remember, we posted my critique and your response on the Kagan Website.

At that time I argued that since different facets of the same intelligence were located in different parts of the brain, it is problematic to make brain localization the most important criterion of defining intelligence. My argument was this: If we mean by "brain localization" localization anywhere, even if different facets of an intelligence are located in different parts of the brain, then every skill passes the criterion of being an intelligence because every cognitive skill is associated with activity somewhere in the brain. If on the other hand we mean by "brain localization" that various facets of intelligence have to be located in one area or in one brain system, then some of the declared intelligences do not stand the test. I pointed out that different facets of the some of your favored eight intelligences are located in different, relatively unrelated areas of the brain. Further, this independence is supported by psychometric data showing very low or even negative correlation among facets of the presumed same intelligence, something you found in your own research.

This was not an argument against the usefulness of defining the eight intelligences (which, as you know, I find very useful); it was an argument against the idea that eight relatively distinct intelligences are supported by brain localization studies.

Last October, a brain localization study was published which I felt very elegantly supported my point, and for some time have wanted to make you aware of it, if you are not already, but only now have found the time. I would very much appreciate any comments you might have about the study or about my interpretation.

The article:
Parsons, Lawrence M. & Osherson, D. New evidence for distinct right and left brain systems for deductive versus probabilistic reasoning. Cerebral Cortex, Oct 2001, 11, 954-965.

In what I think is a brilliant study, Parsons & Osherson used active brain imaging while subjects were engaged in logic problems. Using the same exact stimuli (a wonderful control) different parts of the brain were active if the set was to engage in deductive reasoning than if the set was to engage in probabilistic reasoning. Based on this very tightly-controlled study, deductive reasoning could be located primarily in a very specific area of the right hemisphere—a finding contrary to the oft promoted notion that the left hemisphere is dominant for logic/math tasks. Further, more central to the point here, the researchers found little overlap between those parts of the brain active during deductive and probability reasoning. Probability reasoning was located primarily in areas of the left hemisphere. The findings are consistent with the now popular view that the right hemisphere is for more general or global discriminations and the left for more fine-grained processing. I have summarized some of the key findings below.

Most importantly for MI theory, there is little overlap between these two facets of the logical/mathematical intelligence. The findings seem to me a very strong challenge to the notion that brain localization studies support the existence of one intelligence called "logical/mathematical." Two activities, which are both clearly logical, are located primarily in different parts of the brain (opposite hemispheres!) with only 8% common activation—extraordinarily low common activation for tasks so similar.

Similar arguments can be made on empirical basis for the other intelligences. For example, London Taxi cab drivers asked to imagine getting from one place to another (presumably a facet of the visual/spatial intelligence) show high activation in the right but not left lobe of the hippocampus. When they are asked to imagine places, to call up a strong visual image of monuments in London, that same right part of the hippocampus is relatively quiescent and other parts of the brain show enhanced activity. If visual and spatial can be located in different parts of the brain, how does brain localization support the existence of one visual/spatial intelligence? Isn't it problematic then to make brain localization the strongest criterion for declaring a set of skills an intelligence?

I have no problem with talking about the "visual/spatial intelligence" or the "logical/mathematical intelligence" for pragmatic reasons if it is understood that these are convenient ways to group and think about sets of skills: Eight intelligences are hard enough for teachers to remember and attempt to develop. But I do have a problem with cloaking those intelligences with the mystique of being based on brain science if brain science with increasing clarity is revealing that the skills do not group that way in the brain. Based on the brain localization studies, we would have to conclude there is nothing more in common between deduction and probabilistic reasoning than there is between probabilistic reasoning and understanding one's own dreams.

Howard, I have several intents in writing:
1) I will be giving a workshop this summer on brain science and instruction and want to represent your thinking accurately. I would very much appreciate your thoughts on this issue;
2) I love MI theory, and do not want to see its rationale based on brain localization if in the long run that will be overthrown and out with the bath water will go the beautiful baby;
3) I am curious if you have given this issue more thought and have contemplated revising MI theory based on the outpouring of new brain localization evidence supporting a mini-modular view of the intelligences.

I would very much appreciate your thinking on this,



Brief summary of Parsons, Lawrence M. & Osherson, D. New evidence for distinct right and left brain systems for deductive versus probabilistic reasoning. Cerebral Cortex, Oct 2001, 11, 954-965.


The following sentences bear on a man (he) drawn randomly from a small Texas town.

Premise: If he owns a computer and plays chess for a hobby then he is a science teacher.

Premise: He owns a computer.

Conclusion: He is a science teacher.

Deduction task: Is the conclusion logically valid? That is, given the two premises, does the conclusion necessarily follow?

Probability task: Given the premises, is the conclusion more likely to be true than false?

Most important results:

1) Only 8% of brain areas were active in both the probability and deductive reasoning tasks, even though the two tasks involved identical stimuli.

2) For both the probability and deduction tasks, there is no stimulation above that explained by comprehension in Broca's area, sub-Broca's area, or Wernicke's area (the left hemisphere language sites).

3) For the deduction task, 65% of the significantly activated locations occur in the right hemisphere. Largest activation by far:

a) right middle temporal cortex, just below a region homologous to Wernicke's area and area homologous to Broca's area. (consistent with prior work showing damage to right frontal areas associated with difficulty in making inferences and in interpreting propositions joined or modified by logical connectives; authors)
b) right amygdala (A ha! 70% report sudden insight associated with pleasurable release of tension in deduction but not probability problem solving)
c) right basal gaglia (in prior work shown to support rule-based deduction)
d) medial and dorsolateral prefrontal cortex (executive attention, strategy functions, controlling and monitoring working memory)
e) temporoparietal and anterior cingulate areas (selective and sustained attention; response selection)

4) Deduction produced 1/3 more activation than probability Authors conclude "right hemisphere "logic-specific network."

5) For the probability task, 59% occur in the left hemisphere; those occurring in the right hemisphere explainable for the most part in terms of attention and some subcortical structures. During deduction but not during probability, large intense activations occur in

a) Left hemisphere area of inferior frontal area (which has been implicated in retrieval of semantic information and use of working memory)
b) Posterior cingulate (attention and or long-term episodic memory)
c) Parahippocampal and medial temporal (declarative, semantic memory.

6) Both males and females show same pattern of right hemispheric prevalence for deduction and left hemisphere for probabilistic reasoning.

7) Results can be interpreted as right hemisphere for course and left for fine discriminations: probabilistic reasoning demands fine differentiation—small differentiation of plausibility whereas deduction is all or none, necessary or not.

From: Howard Gardner
Date: Saturday, July 6, 2002 10:41
To: Spencer Kagan
CC: Robert Sylwester
Subject: Re: Brain Localization of the Intelligences

Dear Spencer,

Thanks for your thoughtful note. I don't remember what I wrote you a few years ago but here are my current thoughts.

There is no question that brain science has advanced enormously since 1980 (almost a quarter of a century ago) when I did my original research for the theory of multiple intelligences. Were I to rewrite FRAMES OF MIND today, it would have to be a very different publication. I see no likelihood of my doing that, though, as you know, there are updates in works like INTELLIGENCE REFRAMED.

That said, as you know, I have always acknowledged—indeed, stressed—that each intelligence has separable components (I often call them subintelligences). I would expect them to be represented differently in the human brain (e.g. perception of wide spaces as opposed to local spaces; or perception of melody as opposed to perception of rhythm). And I am increasingly persuaded that there are differences across individuals in how they represent the "same" intelligence; that is one discovery that is more patent in the last twenty years as more precise forms of imaging have been developed. Still, to the extent that the subintelligences come to work together, they will form more of a functional unit (as occurs, for example, in any working musician).

That said, what advantage still exists in talking about a "brain basis" for each intelligence? With the above noted caveats, I still find it to be important to insist on the brain basis. The interest in brain alerts to nonintuitive forms of classification —for example, that language works in similar ways whether it is mediated by ear, eye, or (in the case of blind people) touch or (in the case of deaf people) gesture. Also, the brain basis is important in determining whether a candidate intelligence—like the existential intelligence—is actually separate, or whether it is just another form of philosophical thought (in which case there is no need to tease it out).

Even if the intelligences represent a genuine scientific discovery, the ways in which they are developed, nurtured, and canalized are still an issue for culture, not biology. It is on this issue that I expect to work in the coming years. So while my analysis still begins with the brain and evolution, I don't expect it to end there.

With best wishes.

From: Robert Sylwester
Date: Wednesday, July 10, 2002
To: Spencer Kagan
CC: Howard Gardner

Dear Friends

Thanks for the exchange on the localization of intelligent behavior, Spencer and Howard.

Yesterday, my wife and I attended a performance of the St. Matthew Passion at the Oregon Bach Festival. This commissioned version of the Passion was composed and conducted by Tan Dun (perhaps best known for his score for the recent marvelous film, Crouching Tiger, Hidden Dragon).

Try to imagine a St. Matthew Passion composed by a Chinese Buddhist. The text and organization of such a choral/orchestral work follows a somewhat prescribed form, and both would come under the category of MUSIC (or musical intelligence?)—but Tan Dun's version was certainly musical light years away from what Bach did 250 years ago with his St. Matthew Passion.

Tan Dun's instrumentation consisted of a single violin and cello, lots of percussion, and the voices of the choir and two soloists (who played all the roles). The singers often used their voices as instruments. The percussion included 12 large bowls of water that musicians used to create a wide range of amplified water-related sounds, and also rocks that the choir members had collected from along a stream to pound together in strange rhythms. You get the idea. It was something else, mesmerizing.

Where Biblical text and rich harmonics were everything to Bach, Matthew's text was merely a point of departure for Tan Dun, who also drew from Buddhist texts to express a theme of universality within the universe (Bach communicating a perspective of contrasts—spirit/material, good guys/bad guys, etc). The harmonics similarly often seemed strange to those more familiar with European liturgical music.

Both compositions focused on the same concepts and used the same forms—and they are both simply marvelous. Both are MUSIC, but they're not identical in composition, performance, or hearing/interpreting.

I would imagine that if it were possible to compare the fMRI imaging results of a "musically intelligent" musician or chorister or audience member during the Bach and Tan Dun performances, scientists would discover cognitive system and processing activity similarities and differences within each brain.

Similarly, if you compare brain localization of cognitive functions in intelligence, as viewed by Gardner, Sternberg, Perkins, etc. differences in brain system activation would be apparent.

So there's a unity to a musical work such as the St. Matthew Passion (just as a human action is a unity), but localization creates the unity—whether it's various instruments, the soloists and choir, and the melodic/harmonic forms in the St. Matthew Passion or the various cognitive systems that create a human action.

So maybe I miss your point Spencer, but functional localization is pretty central to all biological systems—whether it's the components of a cell or a brain, or of all the plants, animals, and ground/water/light that make up an ecosystem.

That this consolidation of parts into wholes is dynamic creates the incredible natural diversity we enjoy, whether it's composing/conducting music or teaching a common curriculum to a classroom full of individual students.

Bob Sylwester
Emeritus Professor of Education
University of Oregon

From: Spencer Kagan
Date: Thursday, July 11, 2002
To: Robert Sylwester
CC: Howard Gardner
Re: Localization of the Intelligences

Bob and Howard,

Bob, Thank you for you thoughts on the brain localization of intelligences issue. And your intriguing interpretation of Tan Dun's Passion. And Howard, thank you very much for updating me on your thinking.

I still have a hard time reconciling two premises with a conclusion:

Premise 1: The most important criterion of an intelligence is brain localization (a cornerstone of MI theory);
Premise 2: Empirical work shows that when we do probability problems we engage mainly on one side of the brain and when we do deductive problems we engage mainly the opposite hemisphere and there is only 8% common activation.
Conclusion: Brain science supports the conclusion that deductive and probabilistic reasoning are part of the same intelligence.

In all our minds probability and deduction are both logical activities and historically we have conceptualized them as forms of logical reasoning. But our brain does not seem to know what is in our minds! At least it seems to have a different way of grouping and processing those stimuli. I am unaware of any brain localization studies that link the two type of reasoning and the recent tightly controlled study shows the two type of reasoning are almost completely independent, at least with regard to brain localization.

It still seems to me that logic demands that something has to give. 1) If we accept brain localization as the single most important criterion of an intelligence, we have to say probability reasoning and deductive reasoning are different intelligences because very strong empirical data which demonstrates deductive and probabilistic reasoning are supported by activation in very different parts of the brain. They could be considered part of one intelligence only unless or until we can find some common activation link that to my knowledge has not been found. Logic does not seem to be localizable, or at least not in one place. To say it is OK for an intelligence to be localized in different places would make the localization criterion meaningless because every cognitive activity can be localized somewhere. The localization criterion has teeth only if it means localized in one area or in one brain system. Or 2) If on the other hand we accept probability and deduction as part of one intelligence called logical, then we cannot accept brain localization as the most important criterion of an intelligence.

I would be the last to argue with functional localization. It seems to me though that the brain studies as well as new theorizing about hemisphere specialization are pointing to the conclusion that gross right-wrong deductive reasoning is functionally independent from fine-grained probabilistic reasoning. That in our minds we have conceptualized these two relatively independent functions as part of one thing called logic does not mean that they are in fact one type of function. The brain seems to think differently.

I did enough empirical research to be humbled. Often what was in my mind, my hypothesis, was not supported by the empirical data. I suspect that as more empirical data pours in, we will have to give up the notion that the various ways to be logical are manifestations of a single logical/mathematical intelligence. For years we thought of memory as a storehouse, a place, but brain science makes it clear than it is a process not a place. It is now clear that it is foolish to speak of having a good memory—there is not one thing called memory, but a number of very independent memory systems.

As I think you both know, I am quite comfortable grouping deduction and probability as facets of one logical intelligence for pragmatic purposes, but I am not comfortable with saying that that grouping is supported by the findings of brain science. Neither of your eloquent E-mails has pointed me in the direction of an empirical finding linking the two types of reasoning. And the recent remarkable research demonstrating their relative independence points strongly in the opposite conclusion.

I am almost certain audiences listening to Bach and listening to Tan Dun would show a tremendous amount of activation in common areas of the brain—far far more than 8%. So I do not have any problem thinking of a common musical intelligence from which their genius springs and to which it appeals.

But I must confess the two of you have me at a disadvantage. You are both very strong in the musical intelligence which is my weakest. So perhaps you are both able to hear a transcendent harmony in what to me sounds like findings discordant with a hypothesis. I cannot find harmony in the following three statements:
1) brain localization is the most critical criterion of an intelligence;
2) deductive and probabilistic reasoning are located in very different places in the brain; and
3) deductive and probabilistic reasoning are part of the same intelligence.

Neither of you seem too bothered by this. This poor fellow, however, trying to exercise his (perhaps disjointed) "logical intelligence," remains bothered.


From: Robert Sylwester
To: Spencer Kagan
Cc: Howard Gardner
Date: Saturday, July 13, 2002
Subject: Re: Localization of Intelligences


Some time ago during a different exchange, I suggested that you might read Elkhonon Goldberg's The Executive Brain: Frontal Lobes and the Civilized Mind (2001, Oxford). I don't know if you have, but he differentiates between our brain's almost insular, often innate, modular areas that process very specific tasks (such as heart beat and index finger movements), and less-dedicated cortical gradient areas in which smaller, widely distributed, and highly interconnected networks combine to process such more complex things as representations (a chair) and the elements of a problem.

Goldberg further proposes that the fundamental hemispheric organizing principle is that the right hemisphere processes novel challenges and the left hemisphere processes established routines that we develop for familiar challenges—the major issue when a brain confronts a problem being, Have I ever run into this before? So how does that fit into your deductive RH and probability LH processing?

My point is simply that the cognitive neurosciences are currently in such a volatile period that new discoveries are constantly questioning old interpretations, and they're leading scientists to propose new interpretations.

So one might think that if Goldberg is correct, the previous belief is wrong that the principal purpose of the human LH was to process language and the RH to process space. But not really, because language is an established routine, and most novel challenges we confront include spatial elements (and we can think of an infant creatively communicating via RH spatially-driven body language before discovering the greater efficiency of linear language).

So a lot of this comes down to how tight you are with established categories. Folks who think that the concept of a multiple intelligences brain dogmatically begins and ends with what Howard Gardner wrote 25 years ago may well be distressed by issues such as you're raising. I happen to believe that Howard developed a marvelous useful theory—but I'm also fascinated by the ways that Robert Sternberg, David Perkins, and others organize intelligence, and so categories don't have to be completely tidy and encompassing for me to be intrigued by them. And I'm not distressed when Howard periodically adds MI categories (Wow, how did we process that before we had a MI category for it!)

A SF newspaper music critic thought Tan Dun's St. Matthew Passion was terrible, principally because it wasn't like the one Bach wrote, as far as I could figure. He wasn't willing to get out of his European liturgical box and imagine another and very different way of doing it. Conversely, the Portland Oregonian's critic thought it was great because he searched for Tan Dun's perspective when he listened to it. So what form of music is true music for intelligent folks?

Or consider the current plight of all the Fundamentalist Christians who have been railing against Darwin for 150 years, but now find themselves taking medications that emerged out of Darwinian scientific principles. To be consistent, shouldn't they reject all medicines and genetic medical advances if they don't accept the fundamental principles of genetics?

So go with the categorical flow, Spencer. Almost everything we do is a collaboration of multiple brain processing systems—organized in ways wondrous to behold. That different scientists look at the research and organize the hundreds of systems differently isn't any stranger to me than the zillions of Christian denominations that have emerged out of the same relatively straightforward principally New Testament data base. But the Baptists think the Methodists have it all wrong and vice versa, and some music critics can't imagine that a Chinese Buddhist could compose a legitimate St Matthew's Passion (and in English to boot).

Some folks may feel betrayed to discover that Gardner's marvelously useful theory may now have warts. So what else is new in the theory business? Charles Darwin's theories also had a lot of problems initially. It took 100 years for DNA to be discovered and so solve many of them—and it's taken another 50 years to fine tune things so that genetic manipulations can now occur. 150 years from genetic theory to practical applications. MI is only 25 years old, and I don't know of any scientific theory that worked out all its bugs in 25 years.

Go back to the 1980s and its (now primitive) imaging technology. Howard made a huge creative leap on the basis of what was then known about brain modularity. Current brain research technology provides a much clearer view of brain activation patterns, but that doesn't necessarily denigrate what scientists thought earlier. Folks who tunnel-vision into one eternal perspective of anything have all kinds of problems trying to reconcile disparate things that emerge—whether its about the nature of intelligence or the form of liturgical music.

A major criterion for an intelligence in Gardner's theory is that it's something that all human do—talk, move about, contemplate themselves and others, quantify things, etc. Howard thought that if these are the things that folks do, each ought to be located somewhere in our brain. It's a pretty good idea. That later developments suggest that such systems are sometimes organized counter-intuitively (as Howard said in his response) is of little concern in science. Elkhonon Goldberg's novelty/familiarity take on the shift in our perspective of hemispheric specialization (that I referred to above) is a good example of an earlier idea that can be accommodated into a later idea.

I hope to live long enough to see the intelligence theories of Gardner, Sternberg, Perkins, and others work themselves out. A useful synthesis will emerge. Einstein's theories were validated about the time I was born in 1927—and I've lived to see Nintendo games (perhaps the apogee of the electronic revolution that emerged out of his theories).

Does this sufficiently further muddy the issue?

Bob Sylwester

From: Spencer Kagan
Date: Saturday, July 13, 2002
To: Robert Sylwester
CC: Howard Gardner
Subject: Re: Localization of the Intelligence


Thanks for your recent letter. I am back from doing eight workshops/keynotes in three cities this last week and frankly am quite near comatose, but very much enjoy getting your Email and the thinking it leads me to.

I have not yet read Goldberg but need to.

There is a level of analysis issue, I think, which is part of the problem in defining intelligences. At one level to talk/think about one intelligence is quite meaningful; at another level eight are fine; at another level eight is far too few.

At one level it makes sense to say some people are smart and others are less so. We all know a "George" who plays a great game of chess, can solve all kinds of problems, plans further down the road than the rest of us, creates novel and useful things, anticipates and avoids pitfalls of all sorts, takes advantages of opportunities others of us cannot see, has a great way of expressing himself, and shows himself to be a generally clever fellow in myriad other ways. We know other people who are, to use a suspect word, not as "smart."

Along comes Howard with MI theory. He says it is not as fruitful or accurate to think in terms of one intelligence but rather in terms of eight. George may be good at all the above mentioned things, but is a complete loss with anything to do with music, or a disaster in understanding others. We need to be more differentiated and think in terms of eight intelligences. Howard is taking a different, more fine-grained level of analysis, looking at different facets of what we have traditionally called smart. Why? It is more useful, allowing us to understand ourselves and others better and to be better teachers.

But then, if we apply that same kind of thinking, we can take any one of Howard's eight intelligences and go to an even finer level of analysis. We all know that being able to dress in color coordinated outfits does not predict ability to parallel park. Visual and spatial skills can be quite independent. The brain localization studies support that—distinct areas light up when people imagine themselves navigating than when they imagine themselves looking at a landmark. We could multiply the examples'. We all know the person who can recognize any song but cannot sing on key and the person who is a great leader but not all that empathetic, etc. etc.

How then are we to group all these myriad skills, talents, intelligences? That seems to me to be a very important question, especially as applied to education because it determines how we teach and which intelligences we work to develop. Howard offered the criterion of brain localization as the most important criterion, but one among many. It is an attractive criterion. In fact, it is very hard to argue against. The problem is that the brain localization studies do not support the existence of the eight relatively independent intelligences defined by MI theory.

If different areas of the brain are associated with visual v. spatial skills, as well as deductive v. probabilistic reasoning, then in the long run we will be better of thinking about these as distinct because different instructional strategies will be needed to engage and develop them.

I have no difficulty in using MI theory's eight intelligences in my thinking. MI theory has transformed the way I think about myself and others, and continues to be an extraordinarily useful tool.

I do have difficulty in accepting the claim that the eight intelligences of MI theory are a natural conclusion from brain localization studies. We have very strong hints that brain science is leading us to a different way of categorizing the intelligences, one which will be more aligned with the way the brain is organized. Time will tell.

For now I am happy to "go with the categorical flow" and give Howard tons of credit for having broken the set created by traditional IQ style thinking. But as I go with the flow, I do not go with the claim that MI's 8 is a category system which is strongly supported by brain localization studies. When clearly defined areas in opposite hemispheres light up with different types of logic problems it is not brain localization studies that are leading us to say ability to solve those problems are manifestations of the same intelligence.

Ultimately I will be a lot more satisfied when we get to a categorical system for the intelligences that aligns well with brain localization studies. I think that when we do, it will not be the eight of MI theory, but that we will see MI theory as a marvelous stepping stone that helped us get there.

Thanks again for your stimulating input.


From: Howard Gardner
Date: Mon, 15 Jul 2002
To: Spencer Kagan, Bob Sylwester
Subject: a few points

July 15 2002

Thanks for your memo of July 11. I want to start by emphasizing that your questions are entirely legitimate and I appreciate them. Indeed, one of the frustrations of my involvement with MI theory is that too infrequently have critics read carefully enough to raise good questions and think them through.

Let me make two points and offer a conclusion:

1. Perhaps a better term than 'localization' would be 'represented separately in the brain.' I see from your notes that 'localization' can be ambiguous. I have long been aware that not all individuals, not even all highly musical individuals, represent musical capacities in precisely the same zones of the brain. (Representation depends on how they themselves have encountered and 'made' music). However, it is highly likely that these individuals represent music in the brain differently than they represent logic, math, spatial information etc. If musical capacities were organized in exactly the same way and sites as linguistic capacities, then we would be dealing with one intelligence, and not two.

Indeed, this is one reason why I have been so interested in the "Mozart effect." If music and spatial abilities were indeed represented in the same area (same in the sense of one single region, but not necessarily the SAME region across individuals), then it would not make sense to consider them separate intelligences.

2. I am intrigued to learn that the brain treats deductive reasoning different from probablistic reasoning. Once one hears this, it is not surprising; but I would never have thought of it on my own. But I am not bothered by this distinction in the least —I think it is lovely. All this tells me is that the term 'logic' is very broad and covers more than one kind of thinking. If you look at my work on linguistic intelligence even 25 years ago, I was well aware that syntactic processing is handled entirely differently from semantic processing, let alone metaphor. But that did not lead me to conclude that there was no such thing as linguistic intelligence; only that it has various subfacets.

Conclusion: The profound question raised by your critique is what justifies us in talking about a set of 7-9 intelligences, rather than 50 or 100 separate capacities. I have been aware of this conundrum since the beginning. In a sense the 7 or more intelligence are a convenient way of grouping rather than an absolute categorical system. It is probably more accurate that we have 50-100 different cognitive systems, in the sense that each does something distinctive and each involves its own neural circuitry. I do continue to believe, however, that certain cultural roles require the various subcapacities to work together (e.g., the cultural role of musician requires individuals to yoke together rhythmic, timbre, harmonic sensitivities, etc.) and so in practice the intelligences make a convenient natural kind.

With best wishes to you both.

From: Spencer Kagan
Date: Monday, July 15, 2002
To: Howard Gardner, Bob Sylwester
Subject: Re: a few points

Dear Howard,

Thank you. Your latest clarifies for me your thinking and I think will be helpful for others as well. If you have the time, I would like you to respond to one additional question that is at the heart of my concern about this issue.

I fully accept that it would be foolish to conceptualize as different intelligences those that are associated with exactly the same or even mainly the same brain activity. This I have never questioned. Thus, I have not been concerned that MI theory splits things that really belong together. None of the eight intelligences involve exactly the same neural activation. My questioning is on the opposite side of the coin: Does MI lump things that really belong apart? On what objective basis do we group the abilities?

The years of empirical work I did in social development led me to be keenly aware that things I thought were manifestations of an underlying social orientation (empathy, cooperativeness, face recognition, leadership skills, need for affiliation, caring) were in fact relatively independent empirically. Time and again I was forced to become more differentiated in my thinking. If face recognition and leadership skills, for example, are represented by quite different processes in the brain and are not correlated empirically, is there anything in brain research that supports calling them both manifestations of a single thing called interpersonal intelligence? In our minds they are part of the same package, but in reality they operate quite independently.

It becomes critical then that we have a criterion for grouping things, for saying they are part of a single intelligence. Is there a rational or empirical reason to group deductive and probabilistic reasoning as part of a single intelligence other than we have always called them both logical thinking and it now feels right to see them as part of one intelligence? We all know that deductive and probabilistic reasoning are part of the logical intelligence. But how is that supported by brain science? Is it possible that what we know ain't so? Is it possible that the new brain imaging processes are telling us that how we think about the intelligences is not really how intelligence works?

The history of science is replete with examples of resistance to giving up ideas in the face of contrary evidence. It seems like the data on the two types of logical reasoning should have us question that they are two manifestations of one underlying intelligence. What empirical brain evidence is there that what we call deductive logic and probablistic reasoning are really related at all?

Or, more importantly, what criteria must any two abilities satisfy to be called part of a single intelligence?

A category system has two ways to go wrong: 1) Does it split that which belongs together? 2) Does it lump things which belong apart? I am far less worried that MI theory falls into the first trap than the second. I am confident that each of the eight MI intelligences is—to use your term (which I like)—represented separately in the brain. I am less confident there is an objective way, based on brain science, to say the various abilities subsumed under each of the intelligences are best thought of as manifestations of a single intelligence.

It seems to me, if we cannot find an objective way to group the various abilities, MI remains an extremely useful category system, but not one which is a conclusion based on evidence from brain science.

Or do I still miss something?

Again, I would greatly value your input.


From: Howard Gardner
To: Spencer Kagan
Date: Wednesday, July 17, 2002
CC: Robert Sylwester

I appreciate your note. The question that you raise is well put and fair enough. It is not easy to answer succinctly, however, because the issues are complex. On the one hand, we have evolution: why do we evolve logical capacities, musical capacities, linguistic capacities etc? Is there a certain "spandrel quality" such that actually independent information processing capacities are yoked together because they are synergistic? Perhaps that is what human language is—4 or 40 separate computations that got joined over the millennia.

On the other hand, we have entrenched cultural practices. Perhaps deductive and probabilistic capacities have quite independent origins and 'syntaxes' and yet, if they are required together to be a good scientist or accountant or lawyer, then these two capacities will pragmatically become linked. Similarly, musicians doubtless depend on several independent capacities and yet, you can't be a successful musician, unless you somehow manage to yoke them together; and then they become, as it were, a unit.

In other words, it is more difficult to apply the tests of "purity" and "dissociability" than any of us would like.

I wish that I could continue this trialogue, from which I have learned, but, as with MI theory, I need to give it some 'benign neglect' for awhile because other projects call.

With best wishes to you both.

From: Spencer Kagan
Date: Thursday, July 18, 2002
To: Howard Gardner, Robert Sylwester


Thank you very much for your comments. I too have learned in the process and very much appreciate your willingness to endure and respond to my questions.

As you have indicated, you are comfortable with my sharing your comments with others as long as they are shared in their entirety. I think it would be instructive to others wrestling with understanding MI theory for us to reprint the correspondence in the Kagan Online Magazine.

Again, thank you for your time and for sharing your insight. I am sure it will be of interest and help to others as well as myself.


Can Intelligences be Located?
by Dr. Spencer Kagan with Response by Dr. Howard Gardner

I have been writing a piece for our MI Trainer's institute on my responses to the most frequently asked questions in MI workshops. On many points I find myself in complete agreement with Dr. Howard Gardner. On one point, though, I found myself parting company with Dr. Gardner. The issue is brain localization as a criterion of intelligence. I sent my thoughts on the topic to Dr. Gardner who gave a very thoughtful and interesting response. What follows is the piece I wrote indicating problems with brain localization as a criterion for intelligence, and Dr. Gardner's response.

Brain Localization — An Unworkable Criterion

Dr. Spencer Kagan

Teachers at workshops often ask me where in the brain a particular intelligence is located. This question may be stimulated by Howard Gardner's statement in Frames of Mind that he took brain localization to be the most important criterion for an intelligence. Or perhaps the question is stimulated by older, now outdated views of the brain. There was a time when educators thought some people were left-brained and others right-brained and that some functions like language were located in the left brain and others like art were located in the right brain.

Modern brain science reveals a far more complex picture. For some time we have known that when we look at a picture, relations of objects are processed by certain parts of the brain, details by others. New brain imaging technology has recently revealed that "seeing" a picture (or processing any information) is a miracle of cooperation among millions of neuron groups in different parts of the brain. As we look at the world, some neurons respond to color, others to lines of certain angles, others to shapes, others to the direction of people's heads, others to the features on their face, and yet others to emotional expressions their features portray. The experience of reality as we know it is assembled. Only if certain different neuron groups are firing simultaneously (lets say red and triangle) do other groups fire, letting us experience an integrated reality (a red triangle). Every intelligent act we observe in a student is the result of an incredible number of different parts of the brain working in cooperation. Intelligences do not reside in any one place; they are the by-product of a fantastically large number of interactions among neurons in different parts of the brain.

What then can we make of the claim that the most important criterion for an intelligences is brain localization? We must conclude that that notion is outdated. There are two things which can be intended by the phrase "brain localization." We can mean either that the behavior is associated with activity in one spot in the brain (narrow localization), or we can mean that the behavior is associated with activity anywhere in the brain whether it is narrowly located or located in a number of places (broad localization). Either interpretation, however, does not support the notion that there are eight intelligences.

Every cognitive event is associated with neurons firing and neurotransmitters acting on receptor sites. That everything can be "located" in this broad sense makes broad localization an impossible criterion for an intelligence. A criterion which includes everything is no criterion.

If we turn to the narrow notion of localization, localization in one site, the criterion Gardner originally described in Frames of Mind, we find support for an incredible number of intelligences. We can locate sites responsible for receptive speech apart from sites for expressive speech, and within each we can locate many discrete "intelligences." The features of a person's face (leading to face recognition) are processed by different parts of the brain than the emotions portrayed by the face. Brain studies show we can lose the ability to process proper nouns without losing the ability to process common nouns, and within the domain of common nouns we can lose the ability to name tools while retaining the ability to name animals. There are literally millions of different "intelligences" which can be located in this way. The brain imaging technology is progressing at a rapid rate; there will come a time that we will find for humans what we have already established for other primates — that single neurons respond to stimuli as specific as the shape of a hand, or the position of a head.
With either notion of localization, broad or narrow, we do not end up with just eight intelligences. In the face of advancing brain science we find "brain localization" is not a workable criterion for a human intelligence. Intelligences cannot be localized because intelligence is the by-product of an incredible number of interactions within and between the hundred billion neurons in the brain.

Response from Dr. Howard Gardner:

Thanks for running your "usual answer" by me and for requesting my "take" on it. Here is my brief response, which you have my permission to share with interested parties, provided that you do so in toto. I ask that you not excerpt the material.

Any scientific position needs to be considered in the context of the circumstances under which it was developed and the uses to which it is being put. When I was doing my research for Frames of Mind, in the late 1970s and early 1980s, many scholars and laypersons still believed that the brain was an all-purpose devise, with zones and lobes virtually substitutable for one another. It was important to make the case, once and for all, for the highly localized nature of brain functioning and against the classical "equipotential" position. Nowadays, that battle has been fought and settled for most of the scientific community, though skirmishes occasionally arise on both the scientific and the lay front.
The evidence that has accumulated in the last two decades points to far greater localization of function, on the one hand, and to intriguing kinds of dynamic interactions among different columns, and zones, on the other. Any up-to-date account of brain functioning, particularly with reference to psychological functions, needs to convey these points.

Even at the time when I was preparing Frames of Mind, it was evident to me that some brain functions are far more localized than others, and some are far more uniform across individuals than others. Indeed, in Frames of Mind, I contrasted the highly differentiated and localized nature of linguistic intelligence with the more diffuse representation of personal (and even spatial) intelligences. And I also made the point that language is represented far more similarly among right handers than, say, musical intelligence which can be quite distinctively represented depending upon the mode in which music was first encountered and the kinds of music that exist within a culture. All of this was recognized even twenty years ago.

The question arises about whether the vast amount of neurological, genetic, and psychological information about human functioning profits from being organized in terms of a small number of relatively autonomous systems the phrase that I have always used. Not only do I believe that it is profitable to do so. I also believe that, in a broad sense, the discoveries over the past twenty years have been extremely hospitable to my recent division among kinds of intelligence; and that writings from evolutionary psychology (Pinker, Tooby and Cosmides), cognitive archeology (Mithen), and cognitive studies (Astington, Damasio, Rosnow) are consistent with the particular intelligences that I proposed.

There are, of course, contexts in which it is plausible to say that the human brain is a single information processing device. There are also contexts in which it is plausible to talk about dozens or even hundreds of different human faculties that may be individually described, studied, and perhaps even localized. Yet for both scientific and practical reasons, and particularly for educational purposes, there are still strong reasons for aggregating faculties in a small number of clusters, which I call the "intelligences." I believe that my list of intelligences remains a strong contender for the best such cartography. And I note with interest that Spencer Kagan continues to organize his own writings around the list that I've put forth.

Howard Gardner
Cambridge, MA.
July 1998