The Scope and the Subtleties of the Contextualism-Literalism-Relativism Debate

In recent years, a number of new trends have seen light at the intersection of semantics and philosophy of language. They are meant to address puzzles raised by the context-sensitivity of a variety of natural language constructions, such as knowledge ascriptions, belief reports, epistemic modals, indicative conditionals, quantifier phrases, gradable adjectives, temporal constructions, vague predicates, moral predicates, predicates of personal taste, etc. A diversity of labels have consequently emerged, such as ‘contextualism’, ‘indexicalism’, ‘invariantism’, ‘literalism’, ‘minimalism’, ‘relativism’, variously qualified. The goal of this essay is to pinpoint the issues that lie at the heart of the recent debates, clarify what is at stake, and provide a snapshot of the current theoretical landscape. the contextualism/literalism/relativism debate 1 1. Literalism vs. contextualism Though it is customary to talk of the contextualism/literalism debate, or of the contextualism/relativism debate, the use of the definite article is misleading in both cases, since there appears to be a family of debates, on different fronts, that seem to fit these labels equally well. We will see, for example, that a position that counts as “relativist” in the light of one debate may well count as “contextualist” from the point of view of another debate. Throughout the paper, I will be striving to disentangle the terminological web that covers the field and threatens to blur the issues under discussion. For the sake of simplicity, I will describe the differences among the various positions in relation to a specific topic, that of quantifier domain restriction. Even though domain-dependence has not been of major interest to relativism, the choice is methodological, since the parameters at stake are easy to isolate, and the range of possible positions fairly easy to delineate. In this section, I will discuss the contextualism/literalism debate. Once the different contextualist positions have become more clear, the goal of Section 2 will be to disentangle those from the recently emerged relativist positions. Section 3 offers a broad survey of the range of topics that have been the concern of the debates, with references to the literature. Suppose that at a symposium on semantics, the following dialogue takes place between Josh and Marsha: (1) Josh: Are there any philosophers? the contextualism/literalism/relativism debate 2 (2) Marsha: Yes, though most people are linguists. According to the literalist, what Josh (literally) asked in (1) is whether there are any philosophers tout court, and Marsha’s answer in (2) is made true by the mere existence of some philosopher somewhere, regardless of how things stand at the symposium. And while her answer to Josh’s question is almost trivially true, what she further says is almost trivially false, since, according to the literalist, it means that most people in the universe are linguists. In other words, literalism holds that quantifier words such as ‘there are’, ‘most’, ‘every’, etc. always behave the way we would take them to behave in e.g. “There are no unicorns” or “Most people have an IQ over 80”; that is, that they quantify over everything. Minimalism is, so to speak, literalism generalized: it holds that the only cases in which context can have an impact on truth value are those of resolving lexical and syntactic ambiguities and of determining the reference of indexical pronouns such as ‘I’, ‘this’, or ‘today’. Contextualism (broadly understood) is the view that Marsha’s answer in (2) isn’t made true by the mere existence of some philosopher somewhere, but only by there being some philosophers in the domain relevant in the context in which the sentence is uttered and/or evaluated for truth. In our case, it would typically be the domain consisting of the people attending the symposium. Similarly, for her sentence in (2) to be true, it is only required that most people in the contextually restricted domain are linguists. The literalism/contextualism debate directly concerns semantics, in the sense that there is no agreement on what the truth values of given sentences the contextualism/literalism/relativism debate 3 (in given contexts) are. In a situation in which 90% of the people at the symposium are linguists, the literalist predicts that what Marsha says in (2) is false, while the contextualist predicts that it is true. The data gathered in the literature on quantifiers appear to show that ordinary speakers’ intuitions on truth value are more in line with the contextualist predictions than with the literalist. Literalism thus faces the problem of explaining away the accuracy of ordinary speakers’ intuitions, and its proponents typically argue that what speakers’ intuitions actually track is some level of pragmatically conveyed content, rather than semantic content itself.1 While contextualism is the mainstream view with respect to quantifier domain restriction, there remains considerable disagreement on the issue of how context is involved in determining the truth values of sentences containing quantifiers. The first branching point is on the question of whether there is something either in the meaning or in the syntax of the expressions used that calls for domain restriction, the alternative being that the context somehow in and by itself restricts the domain of quantification. Hence: (a) mainstream contextualism: there is some lexical, syntactic or semantic level at which the contextual restriction over the domain of quantification is represented. (b) radical contextualism: there is nothing, at any linguistic level, that calls for domain restriction – rather, it is the result of optional or “free” pragmatic 1 In other words, literalists hold that speakers often mistake what is conveyed for what is said (on this distinction, see Grice 1989). For a critical discussion of the role of speakers’ intuitions for semantics, see e.g. (Bach 1994). the contextualism/literalism/relativism debate 4