semantic holism a metaphysical thesis about the nature of representation on which the meaning of a symbol is relative to the entire system of representations containing it. Thus, a linguistic expression can have meaning only in the context of a language; a hypothesis can have significance only in the context of a theory; a concept can have intentionality only in the context of the belief system. Holism about content has profoundly influenced virtually every aspect of contemporary theorizing about language and mind, not only in philosophy, but in linguistics, literary theory, artificial intelligence, psychology, and cognitive science. Contemporary semantic holists include Davidson, Quine, Gilbert Harman, Hartry Field, and Searle. Because semantic holism is a metaphysical and not a semantic thesis, two theorists might agree about the semantic facts but disagree about semantic holism. So, e.g., nothing in Tarski’s writings determines whether the semantic facts expressed by the theorems of an absolute truth theory are holistic or not. Yet Davidson, a semantic holist, argued that the correct form for a semantic theory for a natural language L is an absolute truth theory for L. Semantic theories, like other theories, need not wear their metaphysical commitments on their sleeves.
Holism has some startling consequences. Consider this. Franklin D. Roosevelt (who died when the United States still had just forty-eight states) did not believe there were fifty states, but I do; semantic holism says that what ‘state’ means in our mouths depends on the totality of our beliefs about states, including, therefore, our beliefs about how many states there are. It seems to follow that he and I must mean different things by ‘state’; hence, if he says ‘Alaska is not a state’ and I say ‘Alaska is a state’ we are not disagreeing. This line of argument leads to such surprising declarations as that natural langauges are not, in general, intertranslatable (Quine, Saussure); that there may be no fact of the matter about the meanings of texts (Putnam, Derrida); and that scientific theories that differ in their basic postulates are ’empirically incommensurable’ (Paul Feyerabend, Kuhn).
For those who find these consequences of semantic holism unpalatable, there are three mutually exclusive responses: semantic atomism, semantic molecularism, or semantic nihilism.
Semantic atomists hold that the meaning of any representation (linguistic, mental, or otherwise) is not determined by the meaning of any other representation. Historically, Anglo-American philosophers in the eighteenth and nineteenth centuries thought that an idea of an X was about X’s in virtue of this idea’s physically resembling X’s. Resemblance theories are no longer thought viable, but a number of contemporary semantic atomists still believe that the basic semantic relation is between a concept and the things to which it applies, and not one among concepts themselves. These philosophers include Dretske, Dennis Stampe, Fodor, and Ruth Millikan.
Semantic molecularism, like semantic holism, holds that the meaning of a representation in a language L is determined by its relationships to the meanings of other expressions in L, but, unlike holism, not by its relationships to every other expression in L. Semantic molecularists are committed to the view, contrary to Quine, that for any expression e in a language L there is an in-principle way of distinguishing between those representations in L the meanings of which determine the meaning of e and those representations in L the meanings of which do not determine the meaning of e. Traditionally, this inprinciple delimitation is supported by an analytic/synthetic distinction. Those representations in L that are meaning-constituting of e are analytically connected to e and those that are not meaning-constituting are synthetically connected to e. Meaning molecularism seems to be the most common position among those philosophers who reject holism. Contemporary meaning molecularists include Michael Devitt, Dummett, Ned Block, and John Perry. Semantic nihilism is perhaps the most radical response to the consequences of holism. It is the view that, strictly speaking, there are no semantic properties. Strictly speaking, there are no mental states; words lack meanings. At least for scientific purposes (and perhaps for other purposes as well) we must abandon the notion that people are moral or rational agents and that they act out of their beliefs and desires. Semantic nihilists include among their ranks Patricia and Paul Churchland, Stephen Stich, Dennett, and, sometimes, Quine. See also ANALYTIC – SYNTHETIC DISTINC — TION , MEANING , PHILOSOPHY OF MIN. E.L.