Page 79 - Encyclopedia of Philosophy of Language

states' semantic values. But the fact that the computer models reality is not relevant to how it works. The transitions from one state to another are determined by the physical structure of the machine and its program, interacting causally with the physical aspects of the states. These processes would be exactly the same even if the states had not been assigned any 'meanings.' Computers are physical engines that realize syntactic engines. They seem to be semantic engines but really are not; for their semantic properties do no causal work (Searle 1980; Dennett 1982). If the intentionality of a mental state is like the semanticity of a computational state, it too is causally irrelevant to the production of subsequent mental states and bodily movements. Contents ride along on top of the physical properties which 'encode' them.
This powerful line of argument relies on quite a few assumptions. Reactions to it take two main forms. One is to query the analogy between computer sem- antics and mental semantics. After all, mental phenomena are not assigned meanings, they have meanings intrinsically (but see Dennett 1987 for a dissenting opinion). The second reaction has been a revival of interest in how reason-giving explanations work, and in the notions of causal efficacy, causal relevance, and explanatory relevance.
5. Varieties of Intentional Contents
5.7 Information Content
The distinction between information content and mental content parallels Grice's distinction between natural meaning ('indicating') and non-natural mean- ing. A natural sign (the signal) carries information about another event or state in virtue of a nomic dependence of the former upon the latter (Dretske 1981:198-99). For example, the firing of a neural unit in the frog's brain indicates that there is a bug in the frog's visual field if, and only if, under normal conditions the unit fires just when a bug flies in front of the frog. Dretske (1988) has refined the basic notion of'indicating' in a number of ways. He constructs the notion of 'functional meaning' out of the idea of a structure's acquiring the biological function of indi- cating something. He thereby gives sense to the idea that a structure can mean that p on occasions when, because p is false, the structure fails to indicate that p. By this move Dretske diminishes the conceptual distance between the naturalistic notion and the men- talistic notion, for it is a key feature of judgments and beliefs that they can misrepresent as well as correctly represent. Many philosophers join with Dretske in hoping that the marriage of information theory with biological teleology will help to explicate mental inten- tionality. However, there is nothing psychological about the basic notion of information; a footprint in the sand can carry information. Even if there is an organism for whom an internal state is supposed to indicate that/?, it is not necessary that the whole organ-
ism be capable of cognizing that p. Subpersonal states do carry information, so this kind of intentionality may legitimately figure in theorizing about how the brain processes information, but some writers have doubted that the notion of mental content can be fully analyzed in informational terms. The sheer range of types of mental contents makes the reductive task a daunting one.
5.2 MentalContent
Mental contents are multifaceted. They have semantic properties (truth-conditions or satisfaction con- ditions, reference, truth-values); explanatory roles; in- ternal structures (concepts and other content-elements are combined in various ways); and they are integrally related to other contents in networks governed by minimal rationality constraints. It is not surprising, therefore, that they can be typed according to various principles. The main ways of classifying, expanded upon below, are by manner of presentation, by Fregean modes of presentation contained within them, and by real-world referents.
5.2.7 Manner of Presentation
Seeing that there is an apple in front of you is a different experience from thinking (without seeing) that there is an apple in front of you, even when the object is the same. Not only are seeing and thinking different attitudes, but also the contents are different. Whether an object is presented perceptually or con- ceptually (or perhaps volitionally) makes a difference to how the content is individuated. Peacocke (1986a) has argued that certain perceptual contents, which he calls 'analogue contents,' cannot be adequately indi- viduated by the rules for individuating conceptual modes of presentation.
5.2.2 Modes of Presentation
Some thought-contents are purely descriptive; they involve the exercise of general concepts only. On the other hand, singular thoughts may incorporate either an individual concept or 'mental name' of an indi- vidual thing, or they may contain indexical elements: the subject thinks about an object demonstratively as 'this' or 'that,' or thinks about a place as 'here' or 'there,' or about a time as 'now' or 'then'. Also some thoughts are anaphoric. These are completed by being thought against the background of other thoughts and memories to which they hark back. Frege called these various ways of thinking 'modes of presentation.'
Content-classification must take account of modes of presentation in order to capture the roles that thoughts play in rational inferences and in the control of behavior. Frege's 'intuitive criterion of difference for thoughts' (Evans 1982) is that if a rational person
judges an object to have property P and at the same time judges it not to have property P, it must be the case that he conceives the object under distinct modes
Intentionality
57