Theory and Method in the Neurosciences.

Theory and Method in the Neurosciences. Machamer, Peter K. , Grush, Rick, and McLaughlin, Peter, Eds. Pittsburgh: University of Pittsburgh Press, 2001. vi + 300 pp. , 65. 00 “At present we are combating a trend. But this trend will die out, superseded by others, and then the way we are arguing against it will no longer be understood; people will not see why all this needed saying. ” —Wittgenstein Is there an overarching theory that drives research in neuroscience and, if so, what is its relation to the methodology that is the engine of scientific discovery? This collection of 12 essays mostly by historians and philosophers of science attempts an answer to this question, and has the merit of a diverse set of opinions on a topic where all too often a single paradigm has dominated theory and research. The volume provides an excellent overview of the current state of the field with arguments for and against the guiding concepts. The book is worth buying solely for the paper by Antti Revonsuo, from the Department of Philosophy in Turku. Revonsuo points out the problems with the reductive shift to a lower level of explanation, which depends on the idea of proximate causality, the neglect of the “ultimate causality” of evolutionary explanations, the decomposition of elements, and the proliferation of local models independent of their coherence. He wonders if there may be “a fatal flaw in the assumptions of the mechanistic research strategy employed (and if so, it) …will sooner or later become permanently unproductive when applied to the study of the brain. ” The chapter should be attached to papers in cognitive neuroscience like a read-at-your-own-risk warning! The papers are of excellent quality but a few stand out as especially thoughtful. Holk Cruse describes work on models of walking in insects in relation to computer simulations. Rick Grush has a stimulating, though difficult, paper on the limits of computational neuroscience. James Bogen gives a thorough critique of functional brain imagery. Kenneth Schaffner concludes a fine paper on the dangers of extrapolating from animal models to humans by wishing “good luck to contemporary biology, which has its fingers crossed that simple mechanisms are the ones that are broadly conserved. ” William Bechtel takes a more conservative approach in a short history of neuroscience, selecting those studies that appear to him to have been seminal from a cognitivist perspective. This seems to me a wrong way to think about history, too much like propaganda for a personal agenda. To interpret the past in light of what the present holds dear is to misconceive the present as an arbiter of historical worth. The value of history, one could claim, is not that it is the royal road to the present, but rather, what lies in the wealth of possibility fallow in byways that scientific progress has left unexplored. Recall C. D. Broad’s remark that: “all good fallacies go to America when they die, and rise again as the latest discoveries of the local professors. ” Underlying these essays is a tension about what concept of the brain, indeed of the physical world (of which the brain is part), should guide our thinking about functional organization. We live in—and the brain is adapted for—the middle-sized world of ordinary objects. Strategies have evolved to resolve the causal determinacy of the very large with the indeterminacy of the very small. Agent causation and uncertainty are everyday examples of this balance. How we think about nature determines a theory of the brain, though a deep theory of the brain would surely influence concepts in physics. Werner Heisenberg has written, “we ask, ‘what does a proton consist of? Can an electron be divided or (is it) indivisible? Is a photon simple or compound? ’ But all these questions are wrongly put, because words such as ‘divide’ or ‘consist of’ have to a large extent lost their meaning. . . (and this leads) to developments that do not fit the real situation in nature. ” Is nature purely relational, is it composed of solids that interact, is it a bottomless nesting of categories? For my money, the E = mc2 of neuroscience will turn out to be a theory of the whole-part relation, not as sums of aggregates, or parts as components, but a qualitative shift from whole to part in a process of becoming or actualization. But the demonstration of that theory entails a radically different history of neuroscience and cognition, and a methodology to access subjective states. For now, it is reassuring that the dominant paradigm is under scrutiny, and that we see the beginnings of a search for new ways of thinking about these matters.