INTRODUCTION
The articles in this section introduce four influential but very different ways of thinking about the mind/brain. Each picture of the mind offers a conceptual framework both for thinking about some of the key problems in the philosophy of psychology and for thinking about how to integrate the many different disciplines studying the mind/brain. We can view these four ways of thinking about the mind/brain as offering different responses to what I have elsewhere termed the interface problem (BermĂșdez 2005).
The interface problem arises because we can study the mind/brain at many different levels. Philosophers typically think about the mind in very high-level ways. They are concerned with mental states such as beliefs and desires, hopes and fears, and their aim is to understand what these mental states are; how they are related to each other; and how they are realized by, or instantiated, in the brain. Psychologists, in contrast, often work at a somewhat lower level of analysis. Their interest is in the different capacities that we have for finding out about the world (both physical and social), for thinking about it, and for acting upon it. Psychologists want to understand both what these capacities are and what mechanisms underlie them. Moving still further down the hierarchy of explanation, neuroscientists study individual neurons and populations of neurons, trying to make sense of the structure and organization of the brain. These are all complementary investigations, and one of the most fundamental problems in the philosophy of psychology is working out how, if at all, they can be fitted together to form a unified theoretical perspective on the mind/brain.
Traditional philosophy of mind concerns itself with various versions of the so-called mindâbody problem. The mindâbody problem is standardly formulated at the level of individual mental states. How are individual mental states related to physical states? Philosophers of mind have answered this question in many different ways. According to behaviorists such as Ryle, mental states are complicated behavioral dispositions, so that a personâs mental life is ultimately determined by how they are behaving and how they are disposed to behave. Identity theorists, on the other hand, take mental states to be identical to brain states, while functionalists argue that mental states stand to neural states in the same relation that software programs stand to the computer hardware that implements them.
In contrast to the traditional mindâbody problem, which in effect asks about the physical basis for individual mental states such as particular beliefs and desires, the interface problem asks about the physical basis for our practices of commonsense psychological explanation â of explaining why people do what they do in terms of how they take the world to be (the beliefs that they have) and how they want things to turn out (the desires that they have). The interface problem is the problem of determining how commonsense psychological explanation is to be integrated with levels of explanation lower down in the hierarchy.
The interface problem: How does commonsense psychological explanation interface with the explanations of cognition and mental operations given by scientific psychology, cognitive science, cognitive neuroscience and the other levels in the explanatory hierarchy?
Each of the four pictures of the mind that are introduced in this section gives a very different response to the interface problem. The four views together form a spectrum.
At one end of the spectrum is the picture of the autonomous mind, which holds that the interface problem is not really a problem at all. Autonomy theorists make much of the distinction between personal and subpersonal levels of explanation (introduced by Dennett in Chapter 1) and argue that there is a fundamental discontinuity between explanations given at the personal level of commonsense psychology and explanations given at the various subpersonal levels of explanation explored by psychologists and neuroscientists. The picture of the autonomous mind understands the mind in terms of an autonomous and independent type of explanation that has no application to the non-psychological world and that interfaces only indirectly with the types of explanation applicable in the non-psychological realm.
According to the picture of the functional mind, however, there is no such fundamental distinction between personal and subpersonal levels of explanation. Commonsense psychological explanations are a species of causal explanation, no more and no less mysterious than the various types of causal explanation with which we are familiar both from science and from our everyday experience of the physical world. Mental states have associated with them a determinate causal role, specifying what normally gives rise to them and how they themselves typically give rise to other mental states and to behavior. This allows us to understand the mental states that feature in commonsense psychological explanation in terms of their causal dimension.
The picture of the representational mind shares with the functional picture the view that the essence of the mind is given by the causal dimension of mental states, but takes a rather different approach to the interface problem. The key idea behind the representational picture is that psychological states should be understood as relations to sentences in an internal language of thought, where the language of thought is a physically realized medium of thought that has many of the properties of a natural language. This allows representational theorists to think about thinking in terms of operations that act directly only on the physical properties of those inner sentences, but in a way that is sensitive to the semantic relations between those inner sentences (to the relations that hold between their meanings). The causal transitions between states of the representational mind are purely formal in a way that exactly mirrors the transitions between states of a digital computer. In fact, representationalists effectively claim that the mind can best be modeled as a digital computer.
At the other end of the spectrum from the conception of the autonomous mind lies the picture of the neurocomputational mind. Like representational theorists, proponents of the neurocomputational mind are deeply influenced by the requirements of modeling the mind, but they are inspired by a fundamentally different paradigm. Whereas the picture of the representational mind is motivated by the idea that the mind is a digital computer and can be studied as a piece of software, in complete independence of the hardware in which it is implemented, neurocomputational theorists are inspired by research into artificial neural networks, which are computer models of different types of cognitive ability explicitly designed to reflect certain features of how the brain is thought to process information. Neural networks do not seem to accommodate many of the features of commonsense psychological explanations, and this inspires proponents of the neurocomputational mind to stress the discontinuities between personal-level explanation and the neuroscientific explanations occurring at the bottom of the hierarchy. They propose a co-evolutionary model, on which our understanding of commonsense psychology coevolves with our understanding of the neural basis of cognition.
IA The autonomous mind
Theorists of the autonomous mind, such as Donald Davidson, John McDowell, Jennifer Hornsby, and (in some of his incarnations) Daniel Dennett, seek to make a principled distinction between the activities and practices of commonsense psychological explanation, on the one hand, and the many different explanatory projects in the social, behavioral, and neural sciences. Although they each do so for very different reasons, these theorists all think that the enterprise of making sense of the thoughts and behavior of other people is a fundamentally different type of explanatory project from the enterprise of trying to understand the neural and psychological basis of cognition and behavior. The three papers in this section each offer a different way of motivating this distinction between different explanatory projects.
Daniel Dennettâs âPersonal and Subpersonal Levels of Explanationâ, drawn from his first book Content and Consciousness, gives a clear formulation of this basic distinction between two types of explanation, which he terms the distinction between personal and subpersonal levels of explanation. There is, Dennett argues, a fundamental difference between the activities and states characteristic of persons, on the one hand, and the activities and states characteristic of parts of persons, on the other. He claims that it is a mistake to think that questions about what a person does (questions formulated at the personal level) can be answered by appeal to operations and mechanisms at the subpersonal level. The example he gives is our ability to locate sensations â to scratch an itch, for example, or to reach to the location of a pain. Although there are of course afferent pathways that carry information to the brain about the location of painful stimuli on the body, it would be a mistake, Dennett thinks, to appeal to these afferent pathways and associated neural mechanisms in answer to the question of how we locate our sensations. Locating sensations is something that we as persons just do. At the personal level there is nothing more to be said. Moving to the subpersonal level does not provide further explanation. It merely changes the subject.
The second paper in this section is Donald Davidsonâs âPsychology as Philosophyâ. Davidsonâs sophisticated version of the picture of the autonomous mind is derived from his well-known doctrine of anomalous monism. Anomalous monism combines the token-identity of mental and physical events (the idea that every mental event is identical to some physical event) with the denial that there can be any strict laws defined over mental states (where Davidsonâs paradigm of a strict law are the laws to be found in fundamental physics). It is this second claim (the so-called anomalism of the mental) that makes Davidson an autonomy theorist and that is explored and defended in this paper. The impossibility of strict laws involving mental states derives, according to Davidson, from the role of considerations of rationality in psychological explanation. When we try to make sense of other peopleâs behavior, we are constrained to interpret them so that their behavior comes out as rational. We have to credit them with a psychological profile that rationalizes the way they behave, which forces us to see human behavior as taking place within a coherent and consistent psychological frame. Often enough, Davidson notes, âwe must warp the evidence to fit this frameâ (see page). There is, Davidson thinks, nothing corresponding to these requirements of rationality, coherence, and consistency when we move outside the realm of commonsense psychological explanation. This is what makes psychological explanation at the person level so fundamentally different from, and irreducible to, the various forms of explanation at the subpersonal level.1
In âPhysicalist Thinking and Conceptions of Behaviour,â Jennifer Hornsby, who has been much influenced by Dennett and Davidson, offers a different perspective on the distinction between personal and subpersonal levels of explanation. Whereas both Dennett and Davidson stress differences in the style of the two different explanatory projects and the constraints that govern them, Hornsby emphasizes what she takes to be the very different explananda (things to be explained) of the two explanatory projects. In very broad terms, both personal and subpersonal explanation can be seen as exploring the processes by which information arriving at the sensory periphery is processed and transformed in a manner that leads ultimately to behavior. This apparent similarity, however, obscures a fundamental point of difference. According to Hornsby, the concept of behavior is understood very differently in the two different explanatory projects. At the personal level, the behavior that is being explained is understood in a broad sense, in terms of actions that are carried out intentionally for a purpose. At the subpersonal level, in contrast, the behavior that is being explained is understood narrowly, in terms of bodily movements/motor responses. Only if we equivocate between these two ways of understanding behavior does it seem possible to assimilate the two types of explanation. But, Hornsby argues, when we see how different they are, and that it is impossible to assimilate one to the other, we will appreciate that it is a mistake to try to explain facts at the personal level in terms of facts at the subpersonal level.
IB The functional mind
Whereas supporters of the autonomous mind argue that, from the perspective of personal-level explanation, the connections between mental states and between mental states and behavior have to be understood in fundamentally rational terms, and hence as very different from the âordinaryâ relations that hold between non-minded physical objects, the picture of the functional mind sees the connections between mental states and between mental states and behavior as being primarily causal. We understand behavior in terms of its causal antecedents, just as we understand everyday physical occurrences in terms of their causal antecedents. There is no basic difference of kind between explanation at the personal level and explanation at the subpersonal level. From this perspective, the key to understanding the mind is understanding the complex network of causal relations that hold between perceptual inputs and mental states; between different mental states; and between mental states and behavior. Commonsense psychology is itself a (partial) model of this complex network. The generalizations of commonsense psychology to which we appeal, either explicitly or implicitly, in psychological explanation are generalizations about causal relations within the network.
David Lewisâs âReduction of Mindâ illustrates one way of resolving the interface problem on the picture of the functional mind, in the context of a global reductionism set out in the first few pages of the paper. Lewis envisages commonsense (or folk) psychology as a theory about the causal relations between perceptual stimuli, mental states, and motor responses. This theory allows us to think of mental states in terms of their causal role, where the causal role of a mental state is given by the causal generalizations in which it features. Causal roles are discoverable at the personal level, by processes of conceptual analysis. Truths about causal roles are, Lewis thinks, analytic truths. Once we have an understanding of these analytic truths about causal roles we resolve the interface problem by looking to psychology and neuroscience to tell us about the subpersonal states that realize or implement these causal roles. According to Lewis, empirical investigations at the subpersonal level will uncover a causal network of physical states that is isomorphic to the commonsense psychological theory arrived at on the basis of conceptual analysis.
In âFunctionalism, Information and Contentâ Robert Van Gulick explores one key aspect of the functional picture. If we are to think of mental states as the occupiers of causal roles, then we have to explain how every aspect of a mental state can be derived from its causal role. In this context, it is helpful to bear in mind the distinction, originating with Frege, between the force and the content of a particular mental state. The force of my belief that, say, the cat is on the mat is the fact that it is a belief (as opposed to a desire, or a hope). Its content is the proposition that the cat is on the mat â a proposition that could equally be the content of the desire that the cat be on the mat. It is relatively straightforward to see how the force of a mental state can be understood in causal terms. The causal role of a belief is fundamentally different from the causal role of a desire in ways that are relatively easy to specify. If I believe that p, then I will act on the basis that p is the case, while if I desire that p, I will act in ways that I think likely to bring p about. But it is far more difficult to see how the content of a belief can be understood in causal terms. This is the problem that Van Gulick addresses. Van Gulick sets out to explain what makes a particular state an informational state. He offers an account in terms of notions of goal directedness and adaptiveness of the causal role that a mental state must play in order to qualify as a processing information for the organism. The basic idea is that states are informational just if they operate in ways that lead to adaptive modification of the organismâs behavior in appropriate circumstances. The paper spells out how âadaptive modificationâ and âappropriate circumstancesâ are to be understood.
The picture of the functional mind, as presented by Lewis and other âfunctionalistâ philosophers of mind, makes two very strong claims about the nature of psychological explanation. The first, which is one of the claims sharply distinguishing the functional mind from the autonomous mind, is the claim that personal level psychological explanation is ineliminably law-governed. As we might expect, given that the functional roles of mental states are fixed by the causal laws holding over those states, psychological explanation is essentially a matter of subsuming behavior under the relevant causal generalizations. This brings with it a second claim, this time about the nature of psychological explanation at the subpersonal level. It is a key tenet of the functional picture that we will be able to identify realizers at the subpersonal level for the functional rol...
