2 Informational semantics
Informational, or causal co-variance, semantics is one version of naturalis-
tic semantics. Semantic theories of this type claim that meaning is carried
by the causal connections between the mind and the world. Roughly, the
idea is that for a mental term ‘S’ to meanS, is for tokenings of ‘S’ to
causally co-vary with Ss – that is, Ss, and only Ss, cause tokenings of ‘S’.
So, the idea is that for the term ‘mouse’ (or its Mentalese equivalent, which
we shall henceforward write asmouse) to meanmouse, is for tokenings of
the termmousein belief to be reliably caused by the presence of mice, and
only by the presence of mice. Such an account is plainly naturalistic, since
the only terms whichWgure in it are ‘cause’, together with terms referring to
worldly properties on the one hand and physical word-tokens and sen-
tence-tokens on the other.
Informational theories of mental content are modelled on the sense of
‘represent’ which is appropriate whenever there are causal co-variance
relations in the natural world, and so whenever one state of the world
carries information aboutanother (see Dretske, 1981). Thus we say, ‘Seven
tree ringsmeans(represents) that the tree was seven years old’, ‘Heavy
clouds mean rain’, ‘Those spots mean measles’ (that is, ‘Spots of that type
causally co-vary with the presence of measles’), and so on. But why would
anyone want to begin asemantictheory here? Since it is obvious that there
is no realintentionality,oraboutness, present in these examples, why
should we take them as our model? There are at least two distinct lines of
attraction.
One comes from noticing that this very same causal co-variance sense
of ‘represent’ is apparently employed by neuropsychologists studying the
brain. They will say, for example, ‘TheWring of this cellrepresentsthe
presence of an upright line in the visualWeld’, on the grounds that the cell
is caused toWre when, and only when, an upright line is present. The
hope is then that we may be able to build up to a full-blown notion of
mental representation from this simple starting point, just as the neuro-
psychologist hopes to build up an account of the visual system from such
simple materials.
The second source of attraction is particularly emphasised by Fodor
(1987, 1990). It is that the account isatomistic, as opposed toholistic,in
form. That is, it attempts to deliver the meaning of each mental term one
by one, without mentioning the meaning of any other mental state of the
thinker. Fodor believes that any acceptable naturalistic, realistic, account
of intentional content should be atomistic. His reasoning is that, once any
degree of holism is admitted into the account, then there is no principled
way of stopping short of saying thatallthe thinker’s actual beliefs partly
Informational semantics 163