In response to:

Evolutionary Psychology: An Exchange from the October 9, 1997 issue

To the Editors:

The exchange between Stephen Jay Gould and Steven Pinker [NYR, October 9, 1997] regarding the nature of evolutionary psychology and more specifically the notion of cognitive spandrels was entertaining and informative. Which aspects, if any, of our mental functioning are spandrels, as opposed to adaptive mechanisms “designed” by natural selection to solve the kinds of problems faced by our ancestors in their struggle to survive and reproduce (quoting Pinker), is an extremely important question that requires serious discussion and debate. But it also involves a study of precisely what those “kinds of problems” may have been, and quite how our ancestors solved them.

In this regard a serious weakness in the current literature of evolutionary psychology is the almost blasé disregard for the archaeological, fossil, and palaeoenvironmental records that provide evidence to address these questions. Reading much of the evolutionary psychology literature as an archaeologist, I find it astonishing that gross generalizations are made about our Pleistocene hunter-gatherer past, with such little reference to the research of archaeologists who endeavor to reconstruct past lifestyles. Pinker did at least acknowledge the existence and value of the archaeological record, when commenting about whether reading may be a spandrel. But whatever its theoretical strengths or shortcomings as debated by Pinker and Gould, evolutionary psychology will not prosper until it takes the hunter-gatherer past as seriously as it claims to do so. The archaeological, fossil, and palaeoenvironmental records provide evidence about past behavior and cognition.

To suggest, as Howard Gardner does in his review of my book The Prehistory of the Mind (Thames and Hudson, 1996), coincidentally in the same issue as the Pinker-Gould exchange, that discussions of prehistoric behavior cannot go beyond speculation appears to be a further reflection of this academic arrogance that pervades the cognitive sciences. Archaeology can go just as far beyond speculation about past behavior as can, say, a cognitive-development-psychologist when speculating about what might be going on in a child’s mind. Perhaps a lot further. Archaeology is no longer a young or naive discipline. The last thirty years has seen a veritable revolution not simply in its use of scientific techniques to extract ever greater amounts of information from stone artifacts, broken bones, or ancient sediments, but in its adoption of a scientific methodology to evaluate claims about past behavior, and indeed past cognition. To dismiss the work of archaeologists as being unable to go beyond speculation reflects not just an intellectual arrogance, but an ignorance of the theories, methods, and techniques used in modern archaeology.

So while Stephen Jay Gould argues that evolutionary psychology could become a more fruitful science by respecting a plurality of evolutionary processes beyond natural selection, I believe it also needs to mature by engaging more directly with the evidence for past human behavior and cognition as inferred from the archaeological record. The discussion about spandrels is a case in point. Pinker, Gould, and others can bicker as long as they wish about whether, say, propensities for art, religion, or a fear of death are spandrels or whether they were selected to solve particular problems faced by ancestors. But it is possible to get to work on such issues by exploring the evolutionary history of these propensities by exploiting the fossil and archaeological records.

Obviously this is not an easy task; as Howard Gardner noted, the number of new discoveries in the last few years that may change one’s understanding has been daunting. So I am not arguing that there are any clear answers about cognitive evolution in these records of the past: cognitive abilities need to be inferred from reconstructions of past behavior, just as psychologists need to infer the cognitive abilities of children or apes from behavioral observations. But one can, I contend, reconstruct evolutionary histories for many of those aspects of human cognition that appear to be uniquely human (but not, the archaeological record suggests, necessarily unique to Homo sapiens sapiens). These evolutionary histories should inform us as to the likelihood that we are dealing with spandrels or “engineered mental solutions” to adaptive problems or at least lead to a more informed and interesting debate.

My own interpretation of the archaeological record as laid out in The Prehistory of the Mind concludes firmly in the favor of Gould that many, perhaps most, of those peculiarly human mental propensities, such as believing in supernatural beings and the pursuit of pure science, are indeed spandrels. These seem to appear suddenly and dramatically in our evolutionary past. Original adaptive values appear absent, but these propensities can be readily understood as byproducts of ways of thinking which did indeed help solve adaptive problems faced by our recent human ancestors.

Yet I am an archaeologist, not as adequately versed in evolutionary theory or human psychology as I would wish. Indeed, perhaps the greatest constraint on our understanding of how the human mind evolved is that none of us are sufficiently well versed in a sufficient number of disciplines to make sufficiently informed contributions about cognitive evolution. The structure of our academic institutions or at least those in the UK continue to inhibit multidisciplinary research, and more particularly teaching, on cognitive evolution. Some, such as my own of Reading University, are working to change this with new multidisciplinary graduate degree programs. At Reading a recent innovation has been an MA course taught jointly by archaeology, psychology, philosophy, and linguistics. Hopefully graduates from such courses may be able to take our understanding of cognitive evolution beyond that possible from those of us who have had our minds moulded by one discipline alone.


Steven Mithen
Department of Archaeology
University of Reading, UK

To the Editors:

Howard Gardner criticized Steven Mithen’s book The Prehistory of the Mind because the author proposed that ancient humans evolved a “general-purpose” form of intelligence. While I agree with Gardner that Mithen is guilty of a misconception in the way he formulated his proposal, I think that Gardner’s rather global criticism does not do complete justice to Mithen’s proposal.

Two quite different ideas are conflated in Mithen’s proposal, and they should be picked apart, because only one of them is wrong. The first is the notion that human intelligence could be the result of a general-purpose evolutionary adaptation. This idea is simplistic, and ignores most of what we know about evolution. Adaptations always occur under specific environmental pressures, and within the constraints of very specific brain designs. This results in specialized modifications, rather than general-purpose ones. This rule applies to apes and humans, just as it applies to other species. In fairness, Mithen may not have intended to convey what his terminology seems to suggest, but Gardner is right to criticize him on this point.

The second idea is domain-generality, the notion that some mental faculties have a much wider reach than others. This is an unassailable structural concept. The human mind has exquisite structure. One can imagine it as a set of pyramids, each with its own hierarchy of modules, each of which mediates some special skill. At the base of the pyramid stand many of the most basic reflexive functions. These are the specialists of the mind, narrow-band modules designed to carry out specific functions, such as focusing the eyes, or tasting food, with maximum efficiency. The mid-levels of the mind contain somewhat integrative functions, for instance, spatial maps of the environment, or images of one’s own body, both of which require synthesis across a few modules. And at the top levels, there are the most powerful integrative systems, some of which approach true domain-generality in their reach. Language is one of these. The fact that we can talk about what we hear, smell, touch, or feel, and also describe our cognitive maps of the environment, testifies that our language brain has broad general access to the knowledge gleaned by the specialized modules dedicated to hearing, smelling, touching, and feeling, as well as those concerned with spatial mapping. In that sense, it is a domain- general capacity.

Thus Mithen was right in arguing that humans have evolved some remarkably powerful domain-general capacities. The catch is that domain-generality in cognition should never be confused with the notion of general-purpose adaptation. The two are not the same. Domain-generality is a purely architectural conception, and has nothing whatsoever to say about the specific adaptations by which it is achieved. Language is not the only domain-general system in the human mind. As both Gardner and I have argued from quite different vantage points, we have other capacities that are domain-general in an architectural sense. Because of this, we have many ways of representing the same reality. An artist might draw his conception of an idea, while a speaker, a mathematician, a dancer, and an actor might each capture the same idea in a very different way. There is a good case to be made that each of these is a domain-general construction. Yet none of them can be called a product of “general intelligence.”

As he is undoubtedly discovering, Mithen wandered into treacherous waters when he formulated this part of his proposal. On the whole, however, his project is a good one. He is forging a distinctively archaeological “take” on the prehistory of the mind. In doing this, he might serve as a corrective influence on psychologists such as myself who might lean too heavily on cognitive evidence, and not heavily enough on the corresponding cultural record.

Merlin Donald
Department of Psychology
Queen’s University
Kingston, Ontario

Howard Gardner replies:

I am surprised by Steven Mithen’s response to my review; it seems unnecessarily defensive and inappropriately offensive. I wrote at length and respectfully of Mithen’s undertaking and commented substantively on his various claims. I cannot subscribe at all to his boast that archaeology can proceed as confidently about past behavior—and “perhaps a lot further”—than the field of developmental psychology. Like geology, cosmology, and astronomy and other historically grounded sciences, archaeology must deal with a necessarily scanty record of past events and is ever subject to radical recalibration on the basis of new findings (as both Mithen and I note). In contrast, psychologists participate in experimentally grounded science. Researchers are in a position to develop theories, test hypotheses, run control groups, and, most crucially, replicate results at will, varying factors in a systematic way. It is possible to establish the role assumed by language in the cognitive development of children all over the world; it will never be possible to establish, with a comparable degree of authority, the roles played by languages (and protolanguages) in the evolution of human cognition.
I find the distinction between the two kinds of science useful in evaluating the recent exchanges in these pages about evolutionary psychology. Stephen Jay Gould and Steve Jones acknowledge the fragmentary nature of the data on which evolutionary psychologists must base their (pre)historical reconstructions. As Jones notes, “The mind may have evolved through something quite invisible to experimental science.” In contrast, Steven Mithen and Steven Pinker hope that the interplay among the archaeological record, cross-species comparisons, findings about human development, and the principles of natural selection will yield a cumulative and increasingly reliable science.


Merlin Donald makes a judicious point in his letter. It is true that language has the power to refer to multiple domains. Note, however, that language can never substitute for music, dance, architecture, or other systems of meaning. As the dancer Isadora Duncan once phrased it, “If I could tell it to you, I would not have to dance it.”

This Issue

May 28, 1998