The Scope of Context-Dependence

I argue, in this thesis, that proper name reference is a wholly pragmatic phenomenon. The reference of a proper name is neither constitutive of, nor determined by, the semantic content of that name, but is determined, on an occasion of use, by pragmatic factors. The majority of views in the literature on proper name reference claim that reference is in some way determined by the semantics of the name, either because their reference simply constitutes their semantics (which generally requires a (...) very fine-grained individuation of names), or because names have an indexical-like semantics that returns a referent given certain specific contextual parameters. I discuss and criticize these views in detail, arguing, essentially, in both cases, that there can be no determinate criteria for reference determination—a claim required by both types of semantic view. I also consider a less common view on proper name reference: that it is determined wholly by speakers’ intentions. I argue that the most plausible version of this view—a strong neo-Gricean position whereby all utterance content is determined by the communicative intentions of the speaker—is implausible in light of psychological data. In the positive part of my thesis, I develop a pragmatic view of proper name reference that is influenced primarily by the work of Charles Travis. I argue that the reference of proper names can only be satisfactorily accounted for by claiming that reference occurs not at the level of word meaning, but at the pragmatic level, on an occasion of utterance. I claim that the contextual mechanisms that determine the reference of a name on an occasion are the same kinds of thing that determine the truth-values of utterances according to Travis. Thus, names are, effectively, occasion sensitive in the way that Travis claims predicates and sentences (amongst other expressions) are. Finally, I discuss how further research might address how my pragmatic view of reference affects traditional issues in the literature on names, and the consequences of the view for the semantics of names. (shrink)

I argue that not all context dependent expressions are alike. Pure (or ordinary) indexicals behave more or less as Kaplan thought. But quasi indexicals behave in some ways like indexicals and in other ways not like indexicals. A quasi indexical sentence φ allows for cases in which one party utters φ and the other its negation, and neither party’s claim has to be false. In this sense, quasi indexicals are like pure indexicals (think: “I am a doctor”/“I am not a (...) doctor” as uttered by different individuals). In such cases involving a pure indexical sentence, it is not appropriate for the two parties to reject each other’s claims by saying, “No.” However, in such cases involving a quasi indexical sentence, it is appropriate for the par- ties to reject each other’s claims. In this sense, quasi indexicals are not like pure indexicals. Drawing on experimental evidence, I argue that gradable adjectives like “rich” are quasi indexicals in this sense. e existence of quasi indexicals raises trouble for many existing theories of context dependence, including standard contextualist and relativist theories. I propose an alternative semantic and pragmatic theory of quasi indexicals, negotiated contextualism, that combines insights from Kaplan 1989 and Lewis 1979. On my theory, rejection is licensed with quasi indexicals (even when neither of the claims involved has to be false) because the two utterances involve conflicting proposals about how to update the conversational score. I also adduce evidence that conflicting truth value assessments of a single quasi indexical utterance exhibit the same behavior. I argue that negotiated contextualism can account for this puzzling property of quasi indexicals as well. (shrink)

Nonindexical Context-Dependence and the Interpretation as Abduction Approach Inclusive nonindexical context-dependence occurs when the preferred interpretation of an utterance implies its lexically-derived meaning. It is argued that the corresponding processes of free or lexically mandated enrichment can be modeled as abductive inference. A form of abduction is implemented in Simple Type Theory on the basis of a notion of plausibility, which is in turn regarded a preference relation over possible worlds. Since a preordering of doxastic alternatives taken for itself only (...) amounts to a relatively vacuous ad hoc model, it needs to be combined with a rational way of learning from new evidence. Lexicographic upgrade is implemented as an example of how an agent might revise his plausibility ordering in light of new evidence. Various examples are given how this apparatus may be used to model the contextual resolution of context-dependent or semantically incomplete utterances. The described form of abduction is limited and merely serves as a proof of concept, but the idea in general has good potential as one among many ways to build a bridge between semantics and pragmatics since inclusive context-dependence is ubiquitous. (shrink)

It is sometimes argued that certain sentences of natural language fail to express truth conditional contents. Standard examples include e.g. Tipper is ready and Steel is strong enough. In this paper, we provide a novel analysis of truth conditional meaning using the notion of a question under discussion. This account explains why these types of sentences are not, in fact, semantically underdetermined, provides a principled analysis of the process by which natural language sentences can come to have enriched meanings in (...) context, and shows why various alternative views, e.g. so-called Radical Contextualism, Moderate Contextualism, and Semantic Minimalism, are partially right in their respective analyses of the problem, but also all ultimately wrong. Our analysis achieves this result using a standard truth conditional and compositional semantics and without making any assumptions about enriched logical forms, i.e. logical forms containing phonologically null expressions. (shrink)

A recent quarrel over faultless disagreements assumes that disputes over evaluative sentences should be understood as regular, factual disagreements. Instead, I propose that evaluative disagreements should be understood in Lewisian terms. Language use works like a rule-governed game. In it, the assertion of an evaluative sentence is an attempt to establish one value as default in the conversation; its rejection, in turn, is in most cases the refusal to accept this move.

According to telling based views of testimony (TBVs), B has reason to believe that p when A tells B that p because A thereby takes public responsibility for B's subsequent belief that p. Andrew Peet presents a new argument against TBVs. He argues that insofar as A uses context-sensitive expressions to express p, A doesn't take public responsibility for B's belief that p. Since context-sensitivity is widespread, the kind of reason TBVs say we have to believe what we're told, is (...) not widespread. Peet doesn't identify any problem with his own argument though he does attempt to limit its sceptical potential by identifying special contexts in which TBVs stand a chance of success. A more general defence of TBVs can be provided by showing Peet's argument to be unsound. I argue that Peet's argument is unsound because it requires us to wrongly suppose that speakers do far less labour than their audiences in context-sensitive linguistic communication. I aim to show why—in the context of the epistemology of testimony and the philosophy of language—it's important to recognize the labour that speakers can do, and so can can be held responsible for not doing, in episodes of context-sensitive linguistic communication. (shrink)

This paper attempts to define indexicality so as to semantically distinguish indexicals from proper names and definite descriptions. The widely-accepted approach that says that indexical reference is distinctive in being dependent on context of use is criticized. A reductive approach is proposed and defended that takes an indexical to be (roughly) an expression that either is or is equivalent to ‘here’ or ‘now’, or is such that a tokening of it refers by relating something to the place and/or time that (...) would have been referred to had ‘here’ and ‘now’ been tokened instead. Alternative reductive approaches are criticized. (shrink)

We think and speak in figures. This is key to our creativity. We re-imagine one thing as another, pretend ourself to be another, do one thing in order to achieve another, or say one thing to mean another. This comes easily because of our abilities both to work out meaning in context and re-purpose words. Figures of speech are tools for this re-purposing. Whether we use metaphor, simile, irony, hyperbole, and litotes individually, or as compound figures, the uses are all (...) rooted in literal meanings. These uses invite us to explore the context to find new meanings, new purposes, beyond the literal. -/- This special issue of Philosophical Studies brings together eight papers on various aspects and applications of figurative speech including: David Hills, Mitch Green, Liz Camp, Ofra Magidor, Larry Horn, Ken Walton, Stephen Barker, and Mihaela Popa-Wyatt. (shrink)

Epistemic Contextualism is the view that “knows that” is semantically context-sensitive and that properly accommodating this fact into our philosophical theory promises to solve various puzzles concerning knowledge. Yet Epistemic Contextualism faces a big—some would say fatal—problem: The Semantic Error Problem. In its prominent form, this runs thus: speakers just don’t seem to recognise that “knows that” is context-sensitive; so, if “knows that” really is context-sensitive then such speakers are systematically in error about what is said by, or how to (...) evaluate, ordinary uses of “S knows that p”; but since it's wildly implausible that ordinary speakers should exhibit such systematic error, the expression “knows that” isn't context-sensitive. We are interested in whether, and in what ways, there is such semantic error; if there is such error, how it arises and is made manifest; and, again, if there is such error to what extent it is a problem for Epistemic Contextualism. The upshot is that some forms of The Semantic Error Problem turn out to be largely unproblematic. Those that remain troublesome have analogue error problems for various competitor conceptions of knowledge. So, if error is any sort of problem, then there is a problem for every extant competitor view. (shrink)

It's sometimes thought that context-invariant linguistic meaning must be a character (a function from context types to contents) i.e. that linguistic meaning must determine how the content of an expression is fixed in context. This is thought because if context-invariant linguistic meaning were not a character then communication would not be possible. In this paper, I explain how communication could proceed even if context-invariant linguistic meaning were not a character.

Mark Richard argues for truth-relativism about claims made using gradable adjectives. He argues that truth-relativism is the best explanation of two kinds of linguistic data, which I call: true cross-contextual reports and infelicitous denials of conflict. Richard claims that such data are generated by an example that he discusses at length. However, the consensus is that these linguistic data are illusory because they vanish when elaborations are added to examples of the same kind as Richard’s original. In this paper I (...) defend the reality of Richard’s data. I show that, in trying to make their point, Richard’s critics have focused upon examples that are similar in some respects to Richard’s original but which lack a crucial feature of that original. When we ensure that this feature is in place, elaborations which make the data vanish are not possible. Richard’s critics therefore fail to show that the data generated by Richard’s original example are illusory. (shrink)

Several philosophers have recently claimed that if a proposition is cancellable from an uttered sentence then that proposition is not entailed by that uttered sentence. The claim should be a familiar one. It has become a standard device in the philosopher's tool-kit. I argue that this claim is false. There is a kind of entailment—which I call “modal entailment”—that is context-sensitive and, because of this, cancellable. So cancellability does not show that a proposition is not entailed by an uttered sentence. (...) I close the paper by describing an implication this has for a disagreement between J. L. Austin and Grice concerning the relation between felicity and truth. (shrink)

This paper considers a now familiar argument that the ubiquity of context -dependence threatens the project of natural language semantics, at least as that project has usually been conceived: as concerning itself with `what is said' by an utterance of a given sentence. I argue in response that the `anti-semantic' argument equivocates at a crucial point and, therefore, that we need not choose between semantic minimalism, truth-conditional pragmatism, and the like. Rather, we must abandon the idea, familiar from Kaplan and (...) others, that utterances express propositions `relative to contexts' and replace it with the Strawonian idea that speakers express propositions by making utterances in contexts. The argument for this claim consists in a detailed investigation of the particular case of demonstratives, which I argue demand such a Strawsonian treatment. I then respond to several objections, the most important of which allege that the Strawsonian account somehow undermines the project of natural language semantics, or threatens the semantics -pragmatics distinction. Please note that the paper posted here is an extended version of what was published. (shrink)

It is often assumed that when one party felicitously rejects an assertion made by an- other party, the first party thinks that the proposition asserted by the second is false. This assumption underlies various disagreement arguments used to challenge contex- tualism about some class of expressions. As such, many contextualists have resisted these arguments on the grounds that the disagreements in question may not be over the proposition literally asserted. The result appears to be a dialectical stalemate, with no independent (...) method of determining whether any particular instance of disagreement is over the proposition literally asserted. In this paper, I propose an independent method for assessing whether a disagreement is about what’s literally asserted. Focusing on epistemic modals throughout, I argue that this method provides evidence that some epistemic modal disagreements are in fact not over the proposition literally asserted by the utterance of the epistemic modal sentence. This method provides a way to break the stalemate, and reveals a new data point for theories of epistemic modals to predict—that is, how there can be such modal disagreements. In the rest of the paper, I motivate a general theory of how to predict these kinds of disagreements, and then offer some brief remarks about how contextualist, relativist, and expressivist theories of epistemic modals might accommodate this new data point. (shrink)

This paper discusses and relates two puzzles for indicative conditionals: a puzzle about indeterminacy and a puzzle about triviality. Both puzzles arise because of Ramsey's Observation, which states that the probability of a conditional is equal to the conditional probability of its consequent given its antecedent. The puzzle of indeterminacy is the problem of reconciling this fact about conditionals with the fact that they seem to lack truth values at worlds where their antecedents are false. The puzzle of triviality is (...) the problem of reconciling Ramsey's Observation with various triviality proofs which establish that Ramsey's Observation cannot hold in full generality. In the paper, I argue for a solution to the indeterminacy puzzle and then apply the resulting theory to the triviality puzzle. On the theory I defend, the truth conditions of indicative conditionals are highly context dependent and such that an indicative conditional may be indeterminate in truth value at each possible world throughout some region of logical space and yet still have a nonzero probability throughout that region. (shrink)

In this paper I offer a defence of a Russellian analysis of the referential uses of incomplete (mis)descriptions, in a contextual setting. With regard to the debate between a unificationist and an ambiguity approach to the formal treatment of definite descriptions (introduction), I will support the former against the latter. In 1. I explain what I mean by "essentially" incomplete descriptions: incomplete descriptions are context dependent descriptions. In 2. I examine one of the best versions of the unificationist “explicit” approach (...) given by Buchanan and Ostertag. I then show that this proposal seems unable to treat the normal uses of misdescriptions. I then accept the challenge of treating misdescriptions as a key to solving the problem of context dependent descriptions. In 3. I briefly discuss Michael Devitt’s and Joseph Almog’s treatments of referential descriptions, showing that they find it difficult to explain misdescriptions. In 4. I suggest an alternative approach to DD as contextuals, under a normative epistemic stance. Definite descriptions express (i) what a speaker should have in mind in using certain words in a certain context and (ii) what a normal speaker is justified in saying in a context, given a common basic knowledge of the lexicon. In 5. I define a procedure running on contextual parameters (partiality, perspective and approximation) as a means of representing the role of pragmatics as a filter for semantic interpretation. In 6. I defend my procedural approach against possible objections concerning the problem of the boundaries between semantics and pragmatics, relying on the distinction between semantics and theory of meaning. (shrink)

one hand, it raises fundamental doubts about the Davidsonian project, which seems to involve isolating specifically semantic knowledge from any other knowledge or skill in a way reflected by the ideal of homophony. Indexicality forces a departure from this ideal, and so from the aspiration of deriving the truth conditions of an arbitrary utterance on the basis simply of axioms which could hope to represent purely semantic knowledge. In defence of Davidson, I argue that once his original idea for dealing (...) with the familiar indexical expressions is suitably implemented, in terms of theorems expressing conditional truth-conditions, the contrast between semantic and nonsemantic knowledge occurs in an appropriate place: between knowledge of the conditional truth conditions themselves (semantic), and knowledge of their antecedents (non-semantic). For example, a simplified version of a conditional truth condition for an utterance u of “That is a cat” will say that for all x, if the utterer in uttering “that” in u thereby referred to x, u is true iff x is a cat. The conditional fact belongs to semantic knowledge; non-semantic knowledge is required to work out what the utterer referred to; the two kinds of knowledge conspire to enable a truth condition to be detached. The other kind of worry relates to difficult technical questions, relating to specific idioms, raised by context-dependence. I consider whether an argument by Travis (2006), designed to establish that there can be no correct truth conditional semantics, can be turned on its head, and used to establish a general method for bringing all forms of context-dependence within Davidsonʼs framework (using quantification into the qualifier “on understanding U”). I argue that this will not work, and that we are committed to a piecemeal examination of various cases, of which I give some examples. One of these relates directly to Travisʼs argument concerning the context-dependence of “grunt”. In addition, Davidsonians need to have a sharp eye for the distinction between pragmatic and semantic content, to be specified by a test akin to Griceʼs notion of cancelability; they should treat sentences like “Itʼs raining” by the conditional truth-condition method; and they should treat possessives as semantically very unspecific. It is part of the thesis of the paper that this list of examples by no means exhausts the cases a Davidsonian needs to address. (shrink)

Contextualism is a view about meaning, semantic content and truth-conditions, bearing significant consequences for the characterisation of explicit and implicit content, the decoding/inferring distinction and the semantics/pragmatics interface. According to the traditional perspective in semantics (called "literalism" or "semantic minimalism"), it is possible to attribute truth-conditions to a sentence independently of any context of utterance, i.e. in virtue of its meaning alone. We must then distinguish between the proposition literally expressed by a sentence ("what is said" by the sentence, its (...) literal truth-conditions) and the implicit meaning of the sentence ("what is implicated" by a speaker uttering the sentence). Over the past forty years, however, an increasing number of linguists and philosophers have begun to underline the phenomenon of semantic underdetermination: the encoded meaning of the sentence employed by a speaker underdetermines the proposition explicitly expressed by an utterance of that sentence. According to the extreme version of this perspective – labelled "radical contextualism" - no sentence of a natural language expresses a complete proposition, or has fixed truth-conditions, even when unambiguous and devoid of indexicals. A sentence expresses a proposition only when completed and enriched with pragmatic constituents that do not correspond to any syntactic element of the sentence and yet are part of its semantic interpretation. More broadly, "contextualism" may be used to refer to a family of views which includes moderate contextualism (also called "indexicalism"), radical contextualism and non-indexical contextualism – and which contrasts with semantic minimalism. (shrink)

The paper emphasizes the inadequacy of formal semantics, the classical paradigm in semantics, in treating contextual dependence. Some phenomena of contextual dependence threaten one central assumption of the classical paradigm, namely the idea that linguistic expressions have a fixed meaning, and utterances have truth conditions well defined. It is possible to individuate three forms of contextual dependence: the one affecting pure indexicals, the one affecting demonstratives and "contextual expressions", and the one affecting all linguistic expressions. The third type of dependence (...) is top-down: context, and not only linguistic material, shows which variables must be instantiated, relying on context itself. The generalization of underdetermination to all linguistic expressions is in fact a kind of metadependence: the mode of dependence itself depends on context. (shrink)

The main goal in this paper is to outline and defend a form of Relativism, under which truth is absolute but assertibility is not. I dub such a view Norm-Relativism in contrast to the more familiar forms of Truth-Relativism. The key feature of this view is that just what norm of assertion, belief, and action is in play in some context is itself relative to a perspective. In slogan form: there is no fixed, single norm for assertion, belief, and action. (...) Upshot: 'knows' is neither context-sensitive nor perspectival. (shrink)