Louise Antony is a philosopher of mind who defends a form of non-reductive physicalism. She argues for a middle ground between eliminative materialists who deny the existence of mind (like Daniel Dennett and the Churchlands) and the dualists, panpsychists, or neutral monists who look for something other than a purely physical explanation of the mind. Antony divides her fellow defenders of a non-reductive physicalism into 1) those she calls "Neumanians," who are content to consider mind ineliminable by taking the descriptions, predictions, and explanations of "folk psychology" at face value, and 2) those, like herself, who want more robust reductive explanations of psychological phenomena. She wants to know how psychology can be embodied in material beings. She says:
The debate about non-reductive materialism is, I acknowledged, esoteric - it is an in-house dispute among committed materialists. But I must warn the reader that there is an even more outré dispute on the horizon. Non-reductive materialists do not all agree with each other about exactly what it means to call the mind material. Some (the philosophers I think of as "Neumanians") are willing to stop arguing once it has been shown that psychology is ineliminable, that the descriptions, predictions, and explanations of folk psychology must be taken at face value (Davidson. 1970; Baker, 1995; Burge, 1993). But others of us (and I am in this camp) think that a full defense of psychology requires more - an account of how psychology, with all its distinctive features, could be embodied in material beings. Such an account we contend, requires providing a reductive explanation of psychological phenomena. Thus, I intend to defend a version of non-reductive materialism that insists on ontological autonomy for the entities and properties of psychology, while demanding at the same time an account of psychological phenomena in terms of non-psychological phenomena.Antony's account is to embrace the functionalist and computationalist views of Jerry Fodor and Hilary Putnam. Minds are advanced Turing machines which process "representations." Creatures that reason and deliberate, exhibit intentionality that conceives things and actions which do not yet exist, creatures that have a view of the world as they want it, as opposed to the world that exists, and that have powerful predictive powers, creatures with these characteristics need only possess a capacity to generate, store, and manipulate "representations," she says.
A good naturalist would make this picture the starting point of scientific investigation - why not? The data are manifest; the picture offers an explanation. The first question to ask would be how to understand the notion of "representation" - what kinds of physical states and mechanisms could implement the information processing posited in the naive picture? Turing, of course, provided an answer, by demonstrating how, in principle, a completely physical and fully automatic representation-processing machine could be built. This would be a machine with structured internal elements that could be construed as symbols and internal states defined partly in relation to those symbols, built in such a way that the principles governing the causal interactions among the states (in conjunction with "inputs" and "outputs") mirror rational relations among the representational contents encoded in the symbols. It is important to the adequacy of Turing's model as a model of mind that the "mirroring" be quite strong, and it is - the physical features of the representational elements to which the machine's causal laws are sensitive are precisely the features that serve to encode the elements of the representational contents that are semantically relevant. The generality of the mirroring - the ability of the mechanism to track all the semantic relations that exist among the contents of the symbols - is due to the compositionality of the symbol system as a whole. The application of Turing's theory of automatic computation to psychology yields a satisfying precisification of the naive conception of mind: Thinking is fundamentally a matter of the manipulation of symbols - physical items with representational properties. The logically relevant aspects of the representational properties of the symbols are encoded in their syntactic forms, and the compositional structure of the symbol system mirrors the semantic and logical relations in which the representational contents of the symbols participate. Mental states are functional relations to mental symbols, and mental processes are computational processes defined over the mental symbols. The hypothesis that minds are like this is the hypothesis that minds have a "classical" architecture. In the 1970s this hypothesis was first articulated and defended, as the "language of thought" theory, by Jerry Fodor, perhaps the world's foremost champion of intentional realism,8 but it has received substantial development since then, notably by cognitive scientist Zenon Pylyshyn. The LOTT explains the central phenomena. The hypothesis that mental representations are syntactically structured explains how psychological processes can respect rational relations during deliberation. The hypothesis that agents' behavior is mediated by representations explains both intentional inexistence and opacity phenomena. And the hypothesis that representations are realized in physical structures whose forms strongly mirror syntactic structure explains how representations can have causal powers that track rational relations. Finally, the entire picture explains the projectibility of mentalistic discourse: it explains how beliefs, desires, and other mental states implicated in perception and action can constitute natural kinds, capable of grounding prediction and explanation. Not only does computationalism provide a satisfying account of folk psychological data, it has proved immensely fertile when extended beyond the realm of conscious and deliberate thought. Beginning with Chomsky's pioneering approach to language acquisition, and continuing with David Marr's theory of visual processing (Marr, 1982), the computationalist model has offered promising explanations of largely unconscious cognitive feats performed by human beings on a daily basis, such as face recognition. The idea that an innate "theory of mind" underlies our ability to quickly interpret the facial expressions of our con-specifics, and to give intentionalistic construals to characteristically human patterns of behavior, has gained wide acceptanor among psychologists: there is serious evidence that absence of such a "psychology module" might be the central deficit in autism. Computationalism has also been extended to the cognitive achievements of infrahuman animals, such as birds' acquisition of their species' songs and insects' spatial navigation, to account for animal cognition. So here is the situation: we have available to us an intuitively appealing model of mind, one that explains the central phenomena of mentality and that has generates! new and fruitful programs of research within the fields of human psychology ami ethology. It is striking, then, that there is so much resistance to this model within philosophy. But what is more striking than the resistance itself is the fact that critics of this picture have no alternative to offer.Antony discusses Jaegwon Kim's objections to a non-reductive physicalism.
In general, Kim and other reductionists need to show that there is a compelling difference between biology and psychology, such that we can rest content with a biologv that is autonomous from chemistry, but not a psychology that is autonomous from biology. I submit that no such difference will be - or can be - found. Biological theories earn their keep by providing fertile and explanatorily satisfying accounts of the phenomena we pick out under biological description. No one frets about how such theories will be "integrated" into the non-biological realm (although I understand that there have been such worries in the past), for it is presumed that the truth cannot be an enemy to the truth; that if biological phenomena are, as they certainly appear to be, part of the natural material world, that their existence is compatible with their being composed of chemical and ultimately physical stuff. Why cannot the same attitude be taken toward psychological phenomena? It is only if one assumes going into the game that "the mental" is somehow defined in contradistinction to the physical that there can even appear to be a problem about "locating" the mind in a physical world. Nonreductive materialists are thoroughgoing naturalists: we want only the same consideration for the psychological data as are according the data in any other domain. Think about it.