Page 169 - Encyclopedia of Philosophy of Language

ception of ideas as, in effect, mental images is unten- able. Complex thoughts are certainly not just complicated images; indeed, the details of imagery are usually irrelevant to a thought's identity. Further- more, as Berkeley observed, one cannot abstract from one's perceptions of triangles a single image which will fit all triangles. Finally, as Wittgenstein showed (Wittgenstein 1934), tempting though it is to say that what makes an image an image of red is that it is itself red, in truth the subjectively manifest qualities of an image do not determine which objective qualities it is an image of. Given different contexts or causal histor- ies, images with the same subjective qualities can be images of different things.
It is clear from all this that improved versions of the psychologistic thesis will not rely on introspective psychology and will need to provide an account of communication which explains more clearly than Locke does why meaning is fundamentally psycho- logical. However, before looking at proposals to this effect, alternative approaches to the concept of mean- ing need to be considered.
2. GottlobFrege
The characteristic feature of Frege's approach is his concept of 'sense' (Sinn). He distinguished the sense of an expression from its 'reference' (Bedeutung), the object or property which it would normally be taken to stand for, in order to account for differences of meaning among expressions with the same reference— for example, different names of the same planet (Frege 1892). Since Frege also held, most emphatically, that senses are not in any way psychological, he concluded that senses belong to an abstract, but objective, 'third realm' (Frege 1918): they are depsychologized ideas and thoughts, abstract representations whose associ- ation with a linguistic expression determines that expression's reference. Among senses, the senses of sentences are fundamental; words have sense only insofar as they contribute to the sense of sentences in which they occur. This thesis of Frege's contrasts with the priority assigned by Locke to names of simple ideas and with the autonomous significance of terms within Aristotelian semantics. Yet Frege also recog- nized that people understand a sentence by under- standing its constituent phrases—this is his compositionality thesis. This does not conflict with the priority of sentence meaning; for in understanding a phrase, one grasps in abstraction its contribution to the sense of sentences in which it occurs.
Frege's sense/reference distinction, his com- positionality thesis, and his emphasis on the semantic priority of sentences are fundamental advances in our understanding of meaning. But many philosophers disliked the 'Platonism' of his abstract conception of sense. In some cases, the proposed cure was worse than the disease. Russell, who was as hostile to psy- chologism as Frege, rejected Frege's sense/reference
distinction and tried to base a conception of meaning upon referenceto the objects of immediateexperience. Quite apart from the difficulties in explaining how shared meaning is possible if meaning is founded only upon reference to objects of immediate experience, Russell discovered that without a sense/reference dis- tinction he could not give a satisfactory account of the possibility of meaningful, but false, sentences. What was needed was not a repudiation of Frege's categories, but a way of fitting them into a broader framework that harmonized with a tenable phil- osophy of mind. Wittgenstein seemed to offer a way forward here: rejecting as illusory the Fregean thesis that 'what must be added to the dead signs in order to make a live proposition is something immaterial,' he proposed that 'if we had to name anything which is the life of the sign, we should have to say that it was its use' (Wittgenstein 1934: 4).
3. Meaning and Use
The question is then how the injunction to ground an account of the meaning of language on its use is to be developed. Empiricists interpreted it in the light of their emphasis on ostensive definitions. As mentioned in connection with Locke, this approach provides a way in which attributions of meaning can be rendered public, by defining meanings in terms of the types of evidence which would verify or falsify sentences. Yet this seems to imply that the meaning of all sentences, however theoretical, should be definable in terms of observable evidence alone; and though some logical positivists embraced this conclusion, its reductionist conclusions were recognized as unacceptable—for example, that the meaning of sentences concerning the past or the future should be given in terms of present evidence for their truth or falsehood.
This, however, is not the end for empiricist con- ceptions of meaning. Quine argued that reductionist implications can be avoided by taking account of the holistic structure of evidence (the 'Duhem-Quine the- sis') and incorporating this into a holistic conception of meaning (Quine 1960). Quine's position rests on an empiricist, indeed behaviorist, insistence that attri- butions of meaning be grounded on the dispositions of speakers to assent or dissent to stimuli; but although he allows that the meaning of 'observation- sentences' can be thus defined, he denies that this is generally the case, since for most sentences, there is no determinate package of stimuli that determines assent or dissent. The dispositions of speakers depend upon their other beliefs, and one can indefinitely per- mute attributions of meaning and ascriptions of belief while remaining faithful to all the behavioral evidence. However Quine does not then hold that it makes no sense to attribute meaning to individual sentences (other than observation sentences); his position is articulated through the fiction of the radical trans- lator, the ideal anthropologist who aims to translate
Meaning: Philosophical Theories
147