The Consideration of Evolutionary Conservatism Toward a Theory of the Human Brain

THE CONSIDERATION OF EVOLUTIONARY CONSERVATISM TOWARD A THEORY OF THE HUMAN BRAIN ALAN NATAPOFF* Introduction Our present understanding ofthe human brain springs primarily from a single great Darwinian principle: man is descended from infra-human ancestors and has living non-human relatives. Without some such chartering principle and the similarity it implies, there would be no rationale for the use oflower animals as experimental models. Given that motive, experimentation on lower animals has provided access to the understanding ofthose qualities of brain that are evolutionarily old. The uniquely and specifically human qualities ofthe human brain—the new qualities—have, ofcourse, no models in lower brains and present, therefore, special difficulties [i, pp. 189-93; 2, p. 284]. Lacking the rich possibilities ofnormal investigation , the experimental results available are too sparse to motivate a correct theory embodying the new qualities and to distinguish it from alternatives equally consistent with experiment. In sum, the problem is greatly underconstrained, and the normal experimental approaches, based on evolutionary similarity, do not, taken by themselves, form promising strategies for its resolution [3, p. 637]. The difficulties suggest their own remedy: seek new constraints and use the experimental results more efficiently. The choice of additional constraint , however, is not dictated by any pressing consideration. While any * Department of Biology and Research Laboratory of Electronics, Massachusetts Institute of Technology 26-007A, Cambridge, Massachusetts 02139. 1 gratefully acknowledge the useful suggestions , discussions, and criticisms ofDrs. Patrick D. Wall, Walle Nauta, Karl Kornacker, and Arthur Taub. The errors remaining are entirely mine. This investigation was supported, in part, by a Public Health Service Fellowship from the National Institute ofNeurological Diseases and Blindness; additional support was received from the National Institutes of Health (grant 5 ROi NB-04987-04); the U.S. Air Force (Aerospace Medical Division) under contract AF33 (6is)-3SSy, The Teagle Foundation , Inc.; and the Bell Telephone Laboratories, Inc. 445 such exploratory procedure would, therefore, be speculative, one hopes that it might lead to the accumulation and focusing ofexperimental results in the neighborhood ofa promising and suggestive theoretical skeleton. In these ways, the efficient husbandry ofscarce results may support a verification , where it could not have raised a theory. The major difficulty in fabricating our phenomenological model is the multiplicity ofchoices facing us at each critical stage. To choose among them we adopt an analogue ofthe statistical method ofmaximum likelihood [4; 5, p. 113]. From among the possible sets of choices, we try to find the set that would deliver the situation we actually observe with the greatest likelihood. Ordinarily, likelihood is not an unexceptionable basis for a choice among alternative formulations. Correct solutions may easily require, in general, the selection, at some stage, ofan alternative that is, taken by itself, less attractive than some other. In this case, however , where there are good reasons for wanting a solution (even a wrong one), the criterion of maximum likelihood gives one and can be defended . It dictates that we choose at each stage the alternative that is simplest, requires the fewest additional assumptions, and has the greatest underlying probability of occurrence—provided that choice is consistent with a similar choice at other points ofuncertainty. We do not calculate any numerical values of likelihood, and we do not hope, with such a drastic procedure, to get anything more than a suggestive skeleton of a theoretical structure. This choice ofadditional constraint is suitable because the human brain is evolutionarily recent, in a precise sense we will discuss in a later section. Very slow mechanisms could not have delivered the observed changes in the time available; moderately slow mechanisms would require the intervention ofspecial and rare processes. Without the special knowledge that research will one day provide, the addition of such special assumptions would render the resulting picture successively more arbitrary and unattractive . We have taken the extreme position of assuming that only the simplest and most likely chain ofalternative mechanisms could have produced our brain in the short time available. It is an overstatement of the facts, but it forms a viable strategy for generating a tentative solution. It provides a motivating constraint that leads to a verifiabletheoretical skeleton and is directly related to a central physical fact about the...