Kent Bach, Routledge Encyclopedia of Philosophy
Many of our thoughts are about particular individuals
(persons, things, places, etc.). For example, one can spot a certain Ferrari
and think that it is red. What enables this thought to latch onto
that particular object? It cannot be how the Ferrari looks, for this could
not distinguish one Ferrari from another just like it. In general, how
a thought represents something cannot determine which thing it represents.
What a singular thought latches onto seems to depend also on features of
the context in which the thought occurs. This suggests that its content
is essentially indexical, contextually variable much as the content of an
utterance like 'I am hungry' depends on who utters it and when (see DEMONSTRATIVES AND INDEXICALS). The indexical model of singular
thought is not limited to thoughts about individuals one perceives, like
the Ferrari driving by, but applies also to thoughts about individuals one
remembers or has been informed of, like an old bicycle or Christopher Columbus.
In each case, a certain contextual relation, based on perception, memory,
or communication, connects thought to object.
Our thoughts do not merely represent what the world is like. Many of our thoughts manage to latch onto particular individuals (persons, things, places, etc.), and how this is accomplished needs to be explained. If, for example, one thinks of the bird on the window sill that it is a sparrow (the thought is expressible by 'That bird is a sparrow'), there must be something about the thought which makes it about that bird. It might seem that the object of a thought must be uniquely determined by some component of the thought. This view is analogous to Frege's view that the 'sense' of a linguistic expression determines its reference (see REFERENCE). The idea is that how something is represented determines which thing is represented. That is, the subject component of the thought, expressible in this example by the demonstrative description 'that bird', imposes a certain condition whose satisfaction by a certain individual makes it the object of the thought. This condition is the same no matter who entertains the thought and no matter where or when it is entertained. The trouble with this view is that one might not represent the thing in question with enough specificity to single it out from all other birds. Indeed, one might even misrepresent it--perhaps what one takes to be a sparrow is a porcelain figure.
Here is an alternative view. An individual is singled out as the object of the thought not because it satisfies a condition represented in the thought but because the thought is causally connected to it. That is, as Tyler Burge has suggested, a thought latches on to something by way of bearing a certain contextual relation to that particular individual. If this is correct, then facts about the circumstances in which the thought occurs, and not just atemporal facts about its character, are relevant to the determination of its object. Two widely separated people with similar perceptual experiences (or the same person at different times) could have thoughts of exactly the same character, expressible by 'That bird is a sparrow', and yet be having these thoughts about different birds. One might be correct, the other mistaken. This suggests that singular thoughts are true or false only relative to a context. As John Perry describes them, they are "essentially indexical."
To explain how this can be, we need to distinguish the type of thought that both people have from the particular occurrence (or token) of that type which each has (see TYPE/TOKEN DISTINCTION). Tokens of the same thought type can be about distinct objects because singular thoughts are 'token-reflexive,' much like indexical expressions. For example, the reference of the indexical 'I' depends on who utters it, of 'here' on where it is uttered, and of 'tomorrow' on when it is uttered. As a result, the sentence 'I will be here tomorrow' is neither true nor false in itself but only relative to a particular tokening of it. That is, a given utterance of it is true if the maker of that utterance will be at the place of the utterance on the day after the utterance. Different people, or the same person at a different place or time, would say different things with these words. Notice that there is also a sense in which everyone who utters 'I will be here tomorrow' says the same thing-utterances of the sentence are type-identical, but their truth-conditional contents differ in the way just described. Singular thoughts are similarly token-reflexive and indexical in content, because their objects are determined only relative to their context of occurrence. The object is not thought of as the unique object of a certain sort, that is, under a definite description (see DESCRIPTIONS), but in some contextually sensitive way. On this indexical view of singular thought, a thought represents a certain object in a context, but without representing the context itself. The truth condition of a particular token of the thought-type, such as the one expressible as 'That bird is a sparrow', is not absolute but is contextually variable, in this case varying with the context of perception.
Not only can singular thoughts be about objects of
perception, like the bird on the window sill, they can also be about individuals
one remembers or has been informed of, such as a certain childhood friend
or a certain famous historical figure. François Recanati and Kent
Bach both suggest that these cases involve uses of "mental indexicals."
One suggestive idea, developed most fully by Graeme Forbes, is that when
we are introduced to (or otherwise learn of) an individual by name, we form
a 'mental file' labeled with the name. We can later call up the individual
by name. These mental files on persons and things are relatively permanent,
but, as Recanati suggests, we seem to operate also with temporary files,
labeled with mental counterparts of pronouns, on individuals currently singled
out in perception or in conversation.
References and further reading
* Bach, K. (1987) Thought and Reference, Oxford: Oxford University Press, chs. 1 and 2. (Develops a relational account of perception-, memory-, and communication-based singular thought, according to which percepts, memories, and names (as labels on mental files) function as mental indexicals.)
* Burge, T. (1977) 'Belief de re', Journal of Philosophy 74: 338-62. (Contends that singular thoughts are not reducible to descriptional thoughts and are true or false only relative to contexts.)
Burge, T. (1979) 'Sinning against Frege', Philosophical Review 88: 398-432. (Argues that Frege's theory of context-insensitive senses, if properly understood and suitably refined, is not vulnerable to certain objections based on indexical thoughts.)
Evans, G. (1982) The Varieties of Reference, Oxford: Oxford University Press, chs. 5 and 6. (Proposes an information-based account of singular thought.)
* Forbes, G. (1989) Cognitive Architecture and the semantics of belief', in P. French et al., eds., Midwest Studies in Philosophy, vol. 14, Minneapolis: University of Minnesota Press. (Proposes an account of singular thought, as well as thoughts about natural kinds, according to which the subject term takes the explicitly indexical form, 'the subject of this dossier'.)
* Perry, J. (1993) The Problem of the Essential Indexical and Other Essays, Oxford: Oxford University Press. (Collects the author's influential papers on irreducibly indexical aspects of content.)
* Recanati, F. (1993), Direct Reference: From Language
to Thought, Oxford: Blackwell, chs. 5 and 6. (Develops an account of
non-descriptive modes of presentation, on which names function as labels
on permanent mental files and indexicals as labels on temporary files.)