To appear in S. Gozzano and C. Hill (Eds.), The mental, The physical: New perspectives on type identity. Cambridge, Cambridge University Press Identity, Reduction, and Conserved Mechanisms: Perspectives from Circadian Rhythm Research William Bechtel Department of Philosophy, Center for Chronobiology, and Interdisciplinary Programs in Cognitive Science and Science Studies University of California, San Diego After briefly flourishing as a characterization of the relation between mind and brain in the 1950s (Place, 1956; Smart, 1959), the type identity theory was eclipsed for the rest of the century, supplanted by functionalism. Perhaps the most influential argument against identity theory and for functionalism was the claim that mental processes are multiply realized: the same mental phenomenon, for example hunger or pain, is realized in radically different ways in different brains, such as those of octopi and humans (Putnam, 1967). Although philosophers sometimes alluded to biological examples of multiple realizations, it was for the most part taken to be an obvious truth not requiring empirical support. As a clincher, it was sometimes noted that human brains differ (for example, in their number of neurons or the details of the wiring diagrams between neurons) and yet people share many beliefs (e.g., belief in the truth of the multiple-realization claim). Since one thing cannot be identical to two or more realizations, the alleged identity between psychological phenomena and brain processes seemed clearly refuted. An even broader conclusion was reached based on this rejection of the identity theory: psychology is autonomous from and should be pursued independently of neuroscience. This prescription fit well with the Zeitgeist in cognitive psychology and the emerging interdisciplinary field of cognitive science in the 1970s, when there were few tools available to relate findings in neuroscience to cognitive phenomena (Bechtel, Abrahamsen, & Graham, 1998). Cognitive research typically employed psychological evidence (e.g., reaction times) and computational modeling (often involving manipulation of symbol structures), but not evidence directly about the brain processes involved. Starting in the 1980s, however, psychologists and neuroscientists began to build bridges and increasingly linked their inquiries together (LeDoux & Hirst, 1986). The advent of techniques such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) played important roles in developing the field of cognitive neuroscience. Something very much like identity claims began to appear in cognitive neuroscience research as types of psychological processes were linked to specific types of brain processes. More recently something even more disturbing for those convinced by multiple realization arguments has begun to develop. Whereas PET and fMRI were the focus of much attention because they could be employed in the study of human brains, many neuroscience techniques, such as recording from implanted electrodes and inducing lesions, are far more invasive and not applied to humans for ethical reasons. But neuroscientists are increasingly relating findings on other species using more invasive techniques to those procured with PET and fMRI on humans, generating the very sort of type identity claims the advocates of multiple realization claimed were impossible.1 1 In earlier work, McCauley and I (Bechtel & McCauley, 1999; McCauley & Bechtel, 2001) characterized such identity claims as heuristic so as to emphasize that they are typically advanced as assumptions that then guideIdentity, Reduction, and Conserved Mechanisms p. 2 How do cognitive neuroscientists cope with the claimed multiple realization of mental processes? For the most part, they simply ignore the philosophers’ objections and pursue their science. So the more relevant question is how should philosophers, especially those who might feel the tug from the arguments for multiple realization, respond to these developments in cognitive neuroscience? To address this issue, I will first place it in a broader framework. The issue is not unique to psychology and neuroscience, but arises in the relation of physiology to chemistry, or indeed wherever two disciplines, one of which is viewed as more fundamental, offer different perspectives on the same phenomenon. Both philosophers and scientists employ the term reduction in characterizing relations between the results of higher-level and more basic-level inquiries that are supposedly jeopardized by multiple realization, but they typically understand reduction quite differently. In the first section, I will describe an understanding of reduction provided by the framework of mechanistic explanation that fits with the pursuits scientists label reductionistic. While this will help resolve some confusions that often intrude into the philosophical literature (e.g., that identity claims are themselves reductive), it will not resolve the problem of multiple realization. Rather it allows for its reformulation in terms of multiple types of mechanism generating the same type of phenomenon. It is indeed the case that there are differences between the mechanisms in different species that result in what are treated as the same phenomena. In the second section I will take up this issue directly, first arguing that when the same standards of typing are applied to phenomena as to realizations, in most instances what were viewed as the same phenomenon are in reality very similar but nonetheless different phenomena. Second, I will consider what happens when one uses a coarser grain to type neural phenomena. Then types might range across species and enable scientists to claim that the same type of mechanism in different species produces the same type of phenomenon. As has long been noted in philosophy, the notion of similarity must be appropriately constrained before it can be useful. I will introduce the notion of conserved mechanism that is widely employed in biological research to constrain appeals to similarity between mechanisms, and in the final section, discuss why this is such a powerful concept in contemporary biology. Throughout this discussion I will appeal to research on circadian rhythms as an exemplar as this is a field in which the issues concerning multiple realization, conservation of mechanism, and identity, can be clearly illuminated. A further virtue of this exemplar is