Pragmatics James Slotta University of Texas at Austin jslotta@utexas.edu Abstract: Pragmatics is a branch of linguistics that investigates the ways language is tied to the contexts in which it is used. Pragmatics thus coalesces as a distinct and coherent domain of inquiry only in relation to the study of language abstracted from its use in context, which has been the prime focus of both twentieth century linguistics and philosophy of language. Investigation of standard pragmatic issues such as deixis, presupposition, speech acts, implicatures, politeness, and information structure has been motivated by a variety of difficulties and impasses encountered in the analysis of language in a significantly de-contextualized form. This entry focuses on the ways such pragmatic phenomena have complicated propositional and lexical-grammatical abstractions of language, and some of the prominent pragmatic frameworks developed to address these complications. Keywords: pragmatics, speech acts, implicatures, semantics, indexicality Pragmatics is a branch of linguistics that investigates the ways language is tied to the contexts in which it is used. As this definition indicates, pragmatics coalesces as a distinct and coherent domain of inquiry only in relation to the study of language abstracted from its use in context <iela0078>, which has been the prime focus of both twentieth century linguistics and philosophy of language. The topics typically discussed under the heading of pragmatics arise from a variety of difficulties and impasses encountered in the analysis of language extracted from context; and as a result, they compose a motley collection, including deixis, presupposition, speech acts, implicatures, politeness, information structure, and so on. As an organizing framework, this survey highlights two of the principal ways in which twentieth century philosophy and linguistics have abstracted language from context in constituting their objects of study. These de-contextualized objects of study form the backdrop against which many of the diverse issues and phenomena that make up the field of pragmatics come into view. As such, they provide a useful vantage point from which to survey the field. The first mode of abstraction considered here is one that privileges a single function of speech—the propositional function—over others (e.g., expressive, performative). This unifunctional abstraction common in twentieth century analytic philosophy <iela0312> takes language to be composed of a set of sentences that describe states of affairs in ways that may be judged true or false. As such, language becomes unmoored from context. After all, the truth of the sentence “There are an odd number of planets in the solar system” can be evaluated independently of who, when, where, or why it is uttered. This proposition-centered view of language has been challenged by a number of philosophers, linguists, and anthropologists who point up a variety of other functions of speech. Plurifunctional views of language have not only enriched the limited propositioncentered view, but also increased attention to contexts of language use, and even cast doubt on the context-independence of propositions and their constituents. With increased attention to the non-propositional functions of language, a new set of questions 2 arise: How are the variety of non-propositional functions of language to be identified and analyzed? How are non-propositional functions signaled in language form? Are propositional and non-propositional functions of language independent of one another? Or, are they intertwined in ways that necessitate a more holistic account of language that does not limit considerations of context to the “peripheral” functions of language? The second mode of abstraction considered here, prominent in twentieth century linguistics, may be termed the lexical-grammatical abstraction of language, in which language is taken to be composed of a set of linguistic form-types (e.g., lexemes, morphemes, syntactic configurations), each with distinctive meanings or functions that recur across their varied token <iela0391> occurrences. In other words, invariance across contexts—a kind of context independence—is a fundamental feature of the form-types that serve as units of analysis in the sort of distributional <iela0105> analysis common to descriptivist, structuralist <iela0371>, and generativist <iela0049> linguistics alike. How this lexical-grammatical idealization of language relates to both propositional and non-propositional functions of language remains a persistent issue, giving rise to recurring questions about the relationship of language form to language function: is the wellformed sentence <iela0356>—the maximal domain in the lexical-grammatical analysis of language—the proper domain of analysis for non-propositional functions of language? Is it even sufficient for an analysis of the propositional function? Are additional frameworks beyond grammar and lexicon needed to account for the connection of linguistic forms and (non-)propositional functions in context? The aim here is not to review all of the answers to the questions introduced so far, but to provide a sense of the ways pragmatics as a domain of inquiry has been constituted through some of the more influential responses that have been offered, focusing in particular on the Anglo-American linguistic and philosophical traditions. Pragmatics, Propositionality, and the Plurifunctionality of Language The field of pragmatics—or at least the term “pragmatics”—emerges at the confluence of two philosophical schools: American pragmatism <iela0324> and logical positivism. The term was coined by Charles Morris <iela0279>, who derived it from the name of the American philosophical school associated with Charles Sanders Peirce <iela0305>, William James, and John Dewey. Morris’ account of the functioning of signs draws heavily on the semiotic framework of Peirce, to which he adds an additional trichotomy that distinguishes three branches of the study of signs (i.e. semiotics <iela0354>): 1) syntactics, which considers the relationship of sign forms to one another; 2) semantics, which concerns the relation of sign forms to what they stand for (i.e. their designata); and 3) pragmatics, which is to be the study of the effects of signs on interpreters, and therefore concerned with “all the psychological, biological, and sociological phenomena which occur in the functioning of signs” (Morris 1971: 43). In addition to American pragmatism, Morris’ approach to semiotics also reflects the influence of the then regnant logical positivist view of language. In this view, and that of much early 20th century analytic philosophy, natural language is composed of a host of “systematically misleading expressions” that give rise to philosophical errors and confusions. Clarity was to be achieved through the logical analysis of natural language expressions and, when that proved insufficient, the wholesale reconstruction of language on a logical basis. 3 As with analytic philosophy more generally, the logical positivist program centered on language in its propositional capacity — viz., its capacity to describe reality in ways that may be judged true or false. More specifically, the role of philosophy in the logical positivists’ view was to ensure the logical rigor and clarity of scientific statements. Morris’ colleague Rudolf Carnap sought to advance this project by developing logical “syntaxes” of science, consisting of rules for the ways the sentences of scientific observation and theory may be formed and transformed into one another in truth preserving fashion. As part of this program, pragmatics was largely irrelevant, given that it concerned neither empirically verifiable representations of denotata (the domain of semantics) nor the logical formation and transformation of propositions (the domain of syntax). As a result, the domain of pragmatics was given little consideration as part of efforts to clarify natural language or formulate ideal languages for science, the heart of the logical positivist endeavor. But for those interested in taking languages as they are and not as they should be, pragmatic considerations of contextualized usage are unavoidable even in a highly circumscribed, propositional view of language. Sentences containing indexical expressions, as in (1), cannot be judged true or false without information about their context of use: (1) Your age is now an odd number. Indexical expressions <iela0092>, including the possessive pronoun “your,” the “present” tense of the verb, and the temporal adverb “now,” tie the sentence to its context of use. That is to say, the propositional content of (1) cannot be resolved and a judgement of the proposition’s truth made without specifying the context in which (1) is spoken. In keeping with the spirit of logical positivism, indexical expressions were generally excluded from logical reconstructions of language. But, notably, they are found in every natural language, displaying diverse ways of tying the propositional content of utterances to their context of use (e.g., tenses, pronouns, moods, temporal adverbs, spatial demonstratives, evidentials, and many others). Indexical expressions are thus perhaps the most conspicuous way in which pragmatic considerations have a bearing on propositionality. But more recent philosophical discussions of meaning have put forward a number of other ways in which propositionally-relevant meanings of words and expressions are contextdependent (see the final section for further discussion of this point). In contrast to logical positivist “ideal language philosophers,” so-called ordinary language philosophers directed attention to everyday uses of language as a way to resolve or dissolve philosophical issues. This shift in focus prompted increasing attention to the nonpropositional functioning of language, with JL Austin’s <iela0021> account of “performativity” <iela0307> among the most influential exemplars (Austin 1962). Austin begins his attack on the propositional view of language with an account of what he terms performative utterances— “systemically misleading expressions” that appear to be propositions but are not actually evaluable as true or false. For instance, the expression (2) I apologize for acting oddly. appears identical in form to indicative sentences that can be judged true or false, such as: (3) She apologized for acting oddly. 4 Unlike (3), though, (2) is not a sentence that would normally be judged true or false, according to Austin. Rather, it appears to be the performance of an act, namely the act of apologizing. As an act and not a propositional use of language, its occurrence in context may be judged to be successful or unsuccessful, but not true or false. Moreover, context is essential to evaluating a performative utterance’s success. To be successful, a performative utterance requires that certain contextual conditions (i.e. felicity conditions) be met; and if successful, it gives rise to a particular transformation of the context. At first, Austin appears to be carving out a domain of performative utterances that is distinct from the sort of propositional uses of language that made up the bread and butter of the philosophical interest in language (i.e. “constative utterances,” in Austin’s parlance). But Austin goes on to undermine his own distinction of constative and performative utterances. In contrast to explicit performative utterances as in (2)—which have a regular form (first person subject, verb in the unmarked “present” tense, verbs that name the very act being performed)—Austin points out that there are a variety of what he calls primary performative utterances, which take an unruly, even limitless, variety of forms: (4) I’m sorry for acting oddly. (5) I was being an idiot. (4) and (5) may also function as apologies but they do not announce themselves as such. Moreover, they prove hard to classify as either a performative or constative utterance. Is (4) a description of how one feels and so subject to evaluation as true or false (“No you aren’t!”); or is it an act that is to be judged felicitous or infelicitous (“I accept your apology”)? It functions like an apology (performative), but the success of the apology depends on the truth of what is said (constative), insofar as actually being sorry is a felicity condition for a successful apology. In the process of undermining his own distinction between constative and performative utterances, Austin undermines the possibility of saving the propositional view of language by quarantining it within the domain of constative utterances. Many performative utterances have a constative dimension, as in examples (4) and (5). And, constative utterances that describe the world all have a performative dimension; they occur as part of acts of asserting, declaring, and the like, and so they too have felicity conditions. All utterances, Austin concludes, are uttered as part of what he calls an illocutionary act—an action performed with words (e.g., inquiring, ordering, requesting, promising, asserting, declaring). In Austin’s plurifunctional view, pragmatic considerations of contextualized use cannot be avoided, regardless the type of utterance. The Grammaticalization of Pragmatics and the Pragmaticization of Grammar Though Austin’s conception of illocutionary acts offered an expanded view of the functioning of language, Austin took a formally conservative approach to identifying and analyzing these functions. The sentence-length illocutionary acts that are the prime domain of Austin’s account look very much like the sentence-length propositions examined by analytic philosophers and, significantly, the well-formed sentences that have been the maximal domain of analysis of grammarians. 5 Even more, Austin presents abstract sentence-types as the bearers of illocutionary forces in much the same way that words or lexemes are the bearers of the relatively context independent “meanings” listed in a dictionary. For instance, an explicit performative sentence-type (as in example (2)) appears to bear the same illocutionary force and have the same felicity conditions whenever it is uttered. Illocutionary acts thus appear amenable to a sort of lexical-grammatical analysis that links linguistic form-types to their recurring and distinctive meaning or function across different occurrences. Indeed, John Searle refined Austin’s account by identifying a range of “illocutionary force indicating devices,” including mood markings, word order, intonation, special morphology, and even the propositional content of sentences themselves, as loci where illocutionary forces are coded (1969). From this perspective, the lexical and grammatical features of sentencetypes encode illocutionary forces (e.g., the interrogative “Did he act like an idiot?” typically has the force of an inquiry, the imperative “Don’t act like an idiot” the force of a directive). And, as a result, illocutionary forces can be incorporated as components of sentencegrammar. This grammaticalization of illocutionary forces reached an extreme with the Abstract Performativity Hypothesis (APH), which proposed that all sentences have an explicit performative clause in their deep structure (e.g., “[I DECLARE that]…”). Though it appears absurd, perhaps, the hypothesis was supported by some quite subtle observations about the distribution of lexemes in sentences that gave the hypothesis plausibility. For instance, John Robert Ross (1970) noted that the difference in grammaticality between sentences like (6) and (7) could be explained assuming the APH. On this account, the first- and second-person reflexive pronoun in (6) is licensed by the first- and second-person pronouns in the abstract performative clause (“[I DECLARE to you that] Fools like myself/yourself...”). The third person reflexive pronoun is not similarly licensed by anything in the performative clause, and so (7) is ungrammatical (“[I DECLARE to you that] Fools like himself…”): (6) (7) Fools like myself/yourself often fall in love. *Fools like himself often fall in love. In the course of a sentence’s derivation, the deep structure explicit performative clause could be deleted, thus generating inexplicit, primary performative sentences (as in (6)). Illocutionary acts were thus rendered a lexically and grammatically coded feature of a wide variety of sentence-types in their deep structure. Ross also proposed an alternative “pragmatic” analysis of (6) and (7) that nicely illustrates the inverse approach; instead of grammaticalizing speech acts as the APH does, Ross’ alternative analysis pragmaticized grammar. In this analysis, the first- and secondperson reflexive pronouns in sentences like (6) are licensed not by a deep structure performative clause, but by the co-presence of speaker and addressee in the context of utterance. In effect, pragmatic information about contextual conditions would be introduced into the syntactic component of a grammar. Though Ross was rather dismissive of this alternative, functional theories of grammar have continued to explore the interface of pragmatics and grammar (see Nichols 1984 for a useful if dated review). In a similar vein, discourse analysis <iela0103> has explored aspects of sentence grammar (e.g., topicalization constructions, clefts of various sorts, pronoun distribution) that appear to be 6 influenced by the larger discourse context and distinctions relevant to it (e.g., given and new information, topic and focus, theme and rheme). Even as Generative Semanticists like Ross and functional grammarians blur the boundaries between syntax <iela0380>, semantics, and pragmatics, in many cases the relationship of pragmatic properties to syntactic ones is not clear cut. Presupposition <iela0325> provides a good example. Some linguistic forms, termed presupposition triggers, indicate that particular background assumptions are held to be true. For instance, definite noun phrases like “the king of France” in the sentence “The king of France is a fool” presuppose that there is a King of France at present. Similarly, factive verbs like “regret” occurring with clausal complements as in “She regrets that he acted oddly” presuppose that he acted oddly. Because these presuppositions can be cancelled in some contexts (“I don’t regret having acted oddly, because I didn’t” does not presuppose that I acted oddly), they appear to be context-sensitive, pragmatic properties of utterances and not semantic properties of abstract lexical-grammatical forms. Moreover, they are pragmatic properties that are regularly, though not invariably, associated with particular forms that have distinctive syntactic properties (see (8)). (8a) Factive verbs occur with a wide variety of gerunds: I regret/resent/deplore your falling in love. (8b) While non-factive verbs do not: *I believe/conclude/claim your falling in love. (9) Some factive verbs, though, do not occur with a wide variety of gerunds: *I know/realize your falling in love. This suggests that grammatical analysis should take account of presuppositions as properties of lexical items and the syntactic configurations they occur in. But, in the case of presupposition triggers, as in the case of many pragmatic phenomena, pragmatic distinctions do not map neatly onto syntactic ones (see the factive verbs in (9)). The significance of pragmatic considerations for grammatical analysis is often not straightforward and remains an area of continuing inquiry. While the importance of pragmatic considerations in grammatical analysis is an open question, the attempt to grammaticalize speech acts represented most strikingly in the Abstract Performativity Hypothesis proved untenable. The problem with assimilating speech acts to the lexical-grammatical meaning of sentence-types is evident in the case of indirect speech acts, i.e., utterances whose illocutionary force does not match the force suggested by their lexical-grammatical form. (10a) I want you to bury the turtle. (10b) The turtle is beginning to smell. Though the utterances in (10) might be deemed assertions that can be judged true or false due to the use of the indicative mood, they can also be used with the force of a request. They thus appear to be an indirect means of conveying an illocutionary force that would be more directly conveyed with an imperative or an explicit performative: (11a) Bury the turtle! 7 (11b) I request that you bury the turtle. The illocutionary force of indirect speech acts, then, is not simply a product of their semantic and grammatical properties as sentence-types; consideration of token uses in context appears necessary to explain how indicative sentences can take on the force of requests. Sentence Meaning, Speaker Meaning, and Pragmatic Rationality Indirect speech acts point up the fact that illocutionary forces are not (always) encoded by the grammatical and lexical features of sentence types. An influential alternative to this coding view of speech acts is one that involves a kind of inferential reasoning about linguistic forms in context. In such a view, participants recognize that the utterance of (10b) in context is not in fact functioning (merely) as an assertion about the odor emitted by the turtle. Assuming that the addressee has some responsibility for the disposition of the turtle, they could then infer that there is a covert point to the statement: namely, that (10b) is actually a request that the addressee do something about the turtle. Indirect speech acts point up the difference between what has been termed “speaker meaning” and “sentence meaning” in much of the pragmatics literature. While we might say the “sentence meaning” of (10b) has the illocutionary force of an assertion, the speaker “means” (or intends) something different by it; namely, they mean to issue a request or directive, as in the “sentence meaning” of (11b). This distinction is pertinent to a variety of different situations in which meaning in context is seen as not reflecting the “literal” meaning of an utterance (e.g., hints, irony, metaphor, and a range of other conventional and conversational implicatures). The distinction between speaker meaning and sentence meaning, developed notably in the work of Paul Grice <iela0147>, allows for a neat distinction of semantics (sentence meaning) from pragmatics (speaker meaning). Gricean pragmatics distinguishes itself as a branch of linguistics concerned with 1) a kind of meaning based on speakers’ intentions <iela0173> (as opposed to the context-independent meaning of lexical and grammatical forms) and 2) a kind of inferential reasoning that allows participants to connect what is said to what is meant (as opposed to the rule- and convention-based accounts offered in grammatical and semantic analyses). Among Grice’s most influential contributions is a framework for conversational implicatures <iela0081>, according to which the speaker meaning of utterances in a cooperative conversation is largely a function of the meaning of the underlying propositions (sentence meaning), the context in which they occur, relevant background information, and several purportedly universal, rational maxims of conversation (1975). Drawing on these elements, participants can imply meanings—implicatures—that go beyond the propositional meaning of the utterances. For instance, a letter of recommendation for a professorship in pragmatics that reads in its entirety: (12) The applicant has excellent penmanship. flouts what Grice terms the “maxim of quantity”: the speaker offers much too little information to assess the merits of the job candidate. But, this violation of the maxim can be made to conform with the assumption that the letter writer is being cooperative if one 8 infers that the letter writer has nothing else good to say about the candidate. The implicature of the too brief utterance in (12), then, is that the candidate is not well qualified for the job. As this example illustrates, getting from the sentence meaning of “he has excellent penmanship” to the speaker meaning that “he is not well qualified for this job” involves a great deal of background knowledge about the norms that govern language use in this context (e.g., what is the accepted purpose of the exchange, what counts as being as informative as required, what counts as being relevant to the communicative activity, and so on.; see Keenan 1976 for discussion of this issue). In fact, given all of the accessory information about norms that is necessary to arrive at the implicature, it is not entirely clear what the cooperative principle and its maxims add. Once one knows what a letter of recommendation is supposed to look like and that people try not to say negative things in them, it is fairly clear that (12) is a violation of one genre specific norm (i.e., the expected length and informativeness of a letter of recommendation). And one can infer (among other possibilities) that the violation is part of an effort to uphold another genre specific norm (i.e., not saying negative things in a letter of recommendation). In short, the cooperative principle and its maxims are so general—capable of generating so many possible inferences—that they must be restricted through reference to norms of language use and interpretation. Yet, once these norms are spelled out, it is often not clear what the maxims themselves add to the analysis. Grice’s account of implicatures has been influential in defining the domain pragmatics (relating speaker meaning to sentence/utterance meaning) and singling out the tools particular to it (implicatures and inferences, rather than rules). The approaches to pragmatics that have grown out of Grice’s framework preserve these general features while seeking to constrain the power of Grice’s maxims, which in Grice’s formulation are capable of transforming virtually any sentence meaning into any speaker meaning. The study of generalized conversational implicatures (GCIs)—as opposed to the particularized conversational implicatures that Grice focused on—has been an area where linguists have had some success in formalizing and constraining Gricean implicatures. GCIs are nearly automatic implicatures that occur unless the context indicates otherwise—that is, they occur in virtually any context as something like “default interpretations,” to use Stephen Levinson’s expression (2000). They are thus much less context-sensitive than particularized conversational implicatures and, therefore, more like semantic meanings in terms of their regular association with linguistic forms. A canonical example is “scalar implicatures,” which occur with quantifiers (none, some, all), numbers, and other linguistic scales (e.g., warm/hot, possible/certain). An implicature-based account of the meaning of quantifiers begins with the somewhat unintuitive claim that a sentence like (13) has the same sentence meaning as (14): (13) Some pragmatists have excellent penmanship. (14) At least some pragmatists have excellent penmanship. (15) Some pragmatists have excellent penmanship—in fact they all do. (16) **Some pragmatists have excellent penmanship—in fact none do. (17) Some but not all pragmatists have excellent penmanship. This account of the sentence meaning is supported by the fact that a sentence like (15) does not involve a contradiction: the sentence in (13) is consistent with the statement 9 that “all pragmatists have excellent penmanship.” ((16), on the other hand, does involve a contradiction.) Of course, (13) typically is used with same meaning as (17). This meaning can be derived as a scalar implicature grounded in something like Grice’s maxim of quantity (e.g., “say as much as you can,” “be as informative as possible”): had the stronger claim that all pragmatists have excellent penmanship been warranted, it would have been incumbent on the speaker to say that, so the speaker meaning in (17) is implicated. Scalar implicatures and other GCIs are thus pragmatic meanings, but ones that depend in a limited and negative way on contextual factors. Moreover, they are identified with limited and structured domains of the grammar and lexicon (e.g., scalar implicatures apply to linguistic scales of various sorts). With such limitations, it has proven significantly easier to formalize and constrain the implicatures generated by utterances. But even then, context and cultural knowledge can complicate efforts to provide readily generalizable, formal accounts of them. For instance, one needs to account for the fact that “The apartment is big enough for families with 3 children” implicates “at most 3 children” and not “exactly 3 children.” Relevance theory is an alternative attempt to streamline and constrain the Gricean framework, reworking Grice’s maxim of relevance as a cognitive principle. In effect, relevance is taken to be the optimization of 1) the cognitive effort involved in deriving inferences from what is said and 2) the positive cognitive outcomes of those inferences. How cognitive effort and cognitive outcome are to be measured remain issues to be worked out. In addition to Gricean implicatures, relevance theorists have sought to account for the inferential work that goes into interpreting what they term “explicatures,” a range of pragmatic enrichment processes (e.g., disambiguation, reference resolution) that go into making up the propositional content of utterances. For instance, the context in which (18) is uttered may allow a hearer to determine whether “my book” means “the book that I lost” or “the book that I wrote”: (18) Have you seen my book? Explicatures, like GCIs, are thus an instance of what might be called “truth conditional pragmatics,” to use Francois Recanti’s expression, an approach to propositional meaning in which context-sensitive pragmatic meanings play a part. Truth conditional pragmatics argues that pragmatic reasoning about language use in context enriches the literal, semantic meaning of utterances and therefore shapes the propositional content expressed by those utterances. Propositions thus become the concern of pragmatics. Grice-inspired implicature- and inference-based approach to pragmatics have also been developed to account for non-propositional functions of utterances. In Brown and Levinson’s influential account of politeness <iela0317> (1987), polite ways of speaking arise at the intersection of two conflicting aims: 1) the speaker’s desire to preserve the public self-imagine, or “face” <iela0126> of the hearer while, at the same time, 2) engaging in a face threatening act that undermines the hearer’s self-image. The indirectness <iela0102> made possible by Gricean implicatures provides an optimal means of accomplishing these two goals at once. Indirect speech acts, such as those in (10), allow the speaker 1) to issue a directive that, in imposing limits on the autonomy of the addressee, constitutes a face threatening act while 2) doing so indirectly, thus mitigating the face threat and, in consequence, being polite. As a framework that spells out indirect ways of 10 conveying speaker meanings, the Gricean account of implicatures opens the way for indirectness itself to serve social pragmatic functions like politeness. Brown and Levinson’s theory of politeness highlights the versatility of the Gricean framework, which was initially developed to explain the communication of information. Building on Grice’s insights, neo-Gricean pragmatics has sought to account for everything from lexical-grammatical phenomena like intrasentential anaphora and gaps in the lexicon to social pragmatic phenomena like politeness using a limited set of pragmatic principles. Laurence Horn, for instance, posits only two: the q-principle (“Say as much as you can”) and the r-principle (“Say no more than you must”) (1984). But the hope for a formalized and universally applicable pragmatics—akin to a universal grammar—inevitably runs up against the need to account for cultural particularities of language use, raising a number of questions that neo-Gricean pragmatists are currently attempting to address: can cultural particularities be accounted for through universal principles and maxims that concern the transmission of information? Or, will other functions of language need to be incorporated into the Gricean framework through the addition of other principles, as some theories of politeness have argued? And, crucially, how are additional maxims, principles, or constraints to be identified and incorporated so as to avoid blending cultural particularities and universal tendencies in an ad hoc way that would undermine the universalizing and formalizing aspirations of this program? Sociocultural Pragmatics: Cultural Norms of Social Action In contrast to the universal aspirations of much linguistic pragmatics, socially, culturally, and historically variable norms and regularities of language use in context have been a prime object of study in both sociolinguistics <iela0362> and linguistic anthropology <iela0228>. Variationist sociolinguistics, for instance, has considered the way speech forms vary based on the social <iela0208> and stylistic <iela0372> context in which they occur. The ethnography of communication highlighted the ways in which culturally variable norms shape language use as a socially significant activity. And symbolic interactionism <iela0377>, ethnomethodology <iela0120>, interactional sociolinguistcs, and conversation analysis <iela0080> have in different ways attended to the diverse norms that structure linguistic interaction, viewed as a site where participants manage their identities and relationships, and coordinate their interactional activity through the use of language <iela0175>. The Prague Linguistic Circle’s plurifunctional perspective on language was an important forerunner of these approaches, particularly the influential work of Roman Jakobson <iela0183>. Going beyond the traditional focus on sentence-grammar, these linguists explored the multiplicity of functions performed through the use of language <iela0068>; that is, not only stating propositions about the world but also indicating the attitudes and identities of speakers, affecting addressees, constructing cohesive discourse, and so on. And, importantly, they regarded language form as a medium through which these varied functions are realized. Building on this plurifuctional view, sociolinguistic and anthropological attention to language form has not been restricted to abstract sentencetypes and their constituents <iela0076>. Rather they have regarded everything from sociophonetic <iela0363> variation to positions in conversational sequences to the “social meanings” associated with the use of different languages in interaction as units relevant to the functioning of language in context. Indeed, formal variation that appears to be 11 unconstrained from a lexical-grammatical perspective (i.e., “free variation”) has been found to be structured—an “orderly heterogeneity” in Weinreich, Labov, and Herzog’s terms (1968)—when norms and regularities of language use in context are considered. Integrating this plurifunctional view of language with the sensibility of American cultural anthropology, Dell Hymes <iela0160> proposed a research program he dubbed “the ethnography of communcation” <iela0119>, which considered how language functions varied across different speech communities. The ethnography of communication joined the study of language use with ethnographically rich accounts of social and cultural life as the contextual framework needed to understand the functioning of language in context. In concert with sympathetic folklorists and sociolinguists, ethnographers of communication investigated a host of genres <iela0144> of language use, emphasizing the performance <iela0306> in context of many, such as myth <iela0281>, narrative <iela0283>, and poetry <iela0316>, that had long been analyzed apart from their use in context. Their work highlighted the structure and poetics of verbal artistry and mundane verbal interaction alike, with an eye toward the social and cultural context in which they played a part. Michael Silverstein brought a number of the concerns of these socially, culturally, and interactionally attuned perspectives on language together under the heading of “pragmatics,” conceptualized as the study of what CS Peirce termed “indexical” signs <iela0307>— signs that have a real (e.g., causal, copresent) connection to their object (1976). Indexical signs tie language to its context of use in different ways. “Non-referential indexes” encompass the wide range of speech forms that point to features of the context without employing the referential or propositional machinery of language. Viewed as non-referential indexes, the use of sociophonetic variants, honorifics, gendered “languages” <iela0144>, codes-switches, and the like all more or less consciously point to features of the context—the identity of speakers, the relationship of speakers to addressees, and much else. “Referential indexes”—or deictics, as they are more widely known—tie the propositional meaning of utterances to the moment in which they are uttered (see example (2)); in other words, they conspicuously tie what is said propositionally to the act of saying it. Referential indexes thus serve as important resources for acting in the world by talking about it. For instance, Austin’s explicit performative utterances involve the use of deictics in a particular configuration (first person subject, second person (in)direct object, “present” tense) that maps the propositional content of explicit performatives (i.e. what is talked about) onto the very moment in which the proposition is uttered. In this way, the propositional capacity of language is mobilized to transformative pragmatic effect by describing what is happening in that instance of speech (e.g., “I promise to bury the turtle” describes the act that it performs). Self-referential speech acts like explicit performative utterances are merely one way in which referential indexes can be used as part of language-mediated social action. The study of ritual <iela0347> and oratory (Stasch 2011) provides a particularly rich site for exploring alternative ways in which the power of speech derives from the mapping of what is said onto the act of speaking itself. Non-referential indexes too can play an important role in transforming the world through speech. A non-referential index occurring in interaction may either conform with the context as already constituted up to that point in the interaction; or it may transform the context by pointing to new elements that thence become part of the context. Such transformative uses of non-referential indexes—e.g., shifting into a more formal speech 12 style or using an “informal” pronoun with someone for the first time—are in some respects very much like speech acts. Like the “illocutionary acts” Austin identified, they are verbal acts that transform the context in which they occur. But unlike Austin’s account of speech acts, these are verbal acts accomplished without the use of the propositional or referential capacity of language. From this indexical-centered perspective, speech acts arise from the dynamic interrelation of speech forms indexing a changeable context, regardless of whether those forms involve the propositional or non-propositional capacities of language, whether they are sentence-length, shorter, or longer. A prime analytic domain of sociocultural pragmatics is thus the communicative event and the wide range of signs, linguistic and nonlinguistic, contained therein. The multimodal <iela0276> array of signs produced in communicative events may be considered together in terms of their textuality <iela0382>, i.e., the way signs cohere with other signs that make up their co-text <iela0445>. Cohesion is forged through the use of linguistic elements like anaphors and pronouns, but textuality also encompasses the cohesive relationship of words and expressions as part of culturally structured domains of knowledge as well as the cohesiveness of indexes of identity, social relationship, and so on, comprising linguistic and non-linguistic signs alike. Through such textual arrays of linguistic and non-linguistic signaling in communicative events, participants are seen as taking on social roles, identities <iela0164>, and relations, in effect constituting themselves as social beings in a social world (Bucholtz and Hall 2005). Unusual and unexpected configurations of signs provide for the possibility for new, “metaphorical” or “tropic” <iela0417> meanings, understandings, actions, identities, and relationships to emerge in interaction as well, as when non-Spanish speaking Anglo-Americans use bits of Spanish to present themselves as playful and easy-going, and simultaneously devalue Spanish and those associated with it (Hill 1998). This perspective thus shares with performativist approaches to gender (Butler 1990) and social constructivist theories of identity (Hacking 2002) the view that the social world is performatively constituted in communicative events. The iteration of signs across events of use and the intertextual <iela0456> links that tie one communicative encounter <iela0423> to another thus become an important area for investigating broader institutional processes and political projects through which social identities and relationships are stabilized and transformed over time and space. Attention to reported speech <iela0345> and other ways speech is de- and re-contextualized in new contexts has been central to this approach as a way of tracking how speech events and interactants are linked to one another in complex participation frameworks <iela0428>. Significantly, the sort of pragmatic frameworks that normatively and/or recurrently link signs to contexts in contemporary sociocultural pragmatics are themselves generally taken to be context-bound. Pragmatic considerations of contextualized use of language thus extend to the level of metapragmatic discourse and norms <iela0292> (i.e. discourse and norms that concern language use in context), whose historical emergence in particular social, cultural, economic, and political environments has itself become an object of study. Research on language ideologies <iela0217> and processes of enregisterment <iela0425> considers how socioculturally situated, reflexive <iela0338> models of language use arise from and give shape to linguistic practices and language form. In this approach, the universal principles analysts look to uncover are not metapragmatic norms and regularities themselves, which are held to vary across time and space; rather, attention is given to systematic constraints on people’s awareness <iela0025> of language structure and use, 13 as well as recurring processes through which language ideologies or metapragmatic frameworks are formed (Irvine and Gal 2000). (Though, recent work on pragmatic typology has counterbalanced the cultural and historical particularism of the pragmatics that arise from this approach by seeking to identify empirically grounded, cross-cultural regularities in pragmatic phenomena, e.g., Dingemanse and Enfield 2015). Viewed as a dialectic process, language ideologies and registers emerge historically from conditions of language use in context and in turn (re)shape language use in context. In this way, the context-bound practice of interpretation that is so central to Peirce's pragmatic conception of semiosis has reemerged in discussions of the sociocultural pragmatics of language. The result is a pragmatist (in the philosophical sense) view of pragmatic signification, grounded in processes of use and inference that are themselves context-bound, historically contingent, and in part reflexive in nature. The Pragmatic Makings of Propositionality As sociolinguists, linguistic anthropologists, and others have explored the multiple functions of language that center on indexical relations of speech to its contexts of use, philosophers have probed ways in which the propositional capacity of language is itself tied inextricably to contextualized uses of language. As noted earlier, Austin’s discussion of speech acts was radical in the context of the logical positivism of its day in arguing that the propositional capacity of language is embedded in acts of using language. Determinations of truth and falsity, then, are bound to speech act considerations like felicity conditions. For instance, (19) The woman burying the tortoise over there is acting oddly. may be judged true or false even if the woman happens to be burying a turtle, not a tortoise. Though nothing is denoted by the noun phrase “the woman burying a tortoise over there,” what renders (19) a functioning proposition is that the noun phrase successfully refers to someone for participants in the interaction. Propositionality, then, depends in part on the success of a speech act—reference—in context and not entirely on a decontextualized sense of the correct denotation of terms. Causal theories of reference propose an even more thorough-going relationship between speech acts, contextual use, and propositional meaning. According to this account of reference, proper names <iela0282> come to refer to an individual through an illocutionary act of naming (i.e. a baptism) that fixes a name to its referent. Future uses of the name are parasitic on that originary act, causally connected to it through a history of usage that links later acts of reference to the initial baptism. The meaning of a name is not dependent on the descriptive associations that might be tied to the name, but on a history of use. Hilary Putnam has offered a similar account of natural kind terms, i.e., terms denoting groups of objects that share “natural” properties. Their denotation <iela0093>, he argued, is grounded in a real relation between a natural kind term and paradigmatic examples of it (1975). On this account, a term like water denotes all things sufficiently similar to the paradigm examples. There is an indexical connection between the stuff denoted by natural kind terms and the term itself. In Putnam’s account, this indexical connection is mediated by a social division of linguistic labor in which experts determine what entities are 14 sufficiently similar to the paradigm examples to be included in the set of things denoted by the term (i.e. its denotata). While non-experts may associate certain properties with natural kind terms, these are neither necessary nor sufficient to identify its denotata. Rather, meaning is grounded in a contextual relationship of terms to the objects they denote and a social organization of expert knowledge and authority. Denotation, in this view, is rendered pragmatic to a considerable degree. Putnam’s arguments were developed as an alternative to holistic accounts of the meaning of scientific terms, which viewed meaning as a product of the theories in which they are embedded. The theory dependent or theory relative nature of meaning holds, for instance, that the meaning of the term gold was fundamentally different after it was determined to be an element with an atomic number of 79. When John Locke was writing in 1689, he did not understand gold in these terms, but as a “metal” characterized by “yellowness, great weight, ductility, fusibility and solubility in aqua regia, etc.” For a theory dependent view of meaning, the development of atomic theory radically transformed the meaning of gold. But such a view of meaning as fundamentally context-dependent and radically changeable flies in the face of the intuition that gold denotes (roughly) the same things before and after the emergence of atomic theory. Putnam’s externalist semantics was an effort to defend that intuition by grounding meaning in the external world rather than in theories, so that terms can retain their denotation even as theories change. The causal theory taken alone, though, makes it difficult to account for reference change and failures of reference. For instance, terms like phlogiston and witch lack denotata (at least from the perspective of many philosophers) despite their causal connections with real world paradigm exemplars at one time in the past. A hybrid descriptive-causal theory of reference that remedies this defect but retains some key elements of externalist semantics posits an array of pragmatic considerations that go into the denotation of natural kinds, none of which alone is decisive: 1) there are real (indexical) connections between paradigm examples of denotata and the application of natural kind terms; 2) there is a social division of linguistic labor with experts who adjudicate the denotational meaning of the terms for a community; and 3) terms have an intensional meaning informed by historically changeable theories, beliefs, and interests. In some cases, the collapse of a theory renders terms without an extension (e.g., witches and phlogiston); in other cases, terms retain their connection to paradigm examples despite changes in theory (e.g., gold). Such a view of meaning shares a great deal with performativist or social constructivist views of denotation that stress the role of social and cultural context, power relations and interests, and even particular languages themselves in shaping the denotational meaning of words and expressions. In contrast to a reflectionist view of meaning, in which words are seen as a reflection of reality, the constructivist view holds that the world is constituted by words in significant ways. This constructivist perspective is often associated with poststructuralism, evident for instance in Foucault-inspired accounts of the way institutionally situated discursive formations “make up people” (Hacking 2002). But it is also a key element in the structuralist theories of Benjamin Lee Whorf <iela0405> and Ferdinand Saussure, which do not regard language as reflecting the structure of reality but as imposing structure on it. Though linguistic meaning is shaped by context in all of these approaches, the contexts are quite different, ranging from state institutions to the structure of language itself. But in each case, these constructivist views render linguistic meaning ultimately pragmatic in nature, bound to some contextually variable feature of the social, cultural, or linguistic environment. 15 References and Further Reading Austin, J. L. 1962. How to Do Things with Words. London: Oxford University Press. Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Bucholtz, Mary, and Kira Hall. 2005. “Identity and Interaction: A Sociocultural Linguistic Approach.” Discourse Studies 7: 585–614. Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. Dingemanse, Mark and N.J. Enfield. 2015. “Other-initiated Repair across Languages: Towards a Typology of Conversational Structures.” Open Linguistics 1(1): 96-118. Grice, H. Paul. 1975. “Logic and Conversation.” In Syntax and Semantics, Vol. 3: Speech Arts, edited by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press. Hacking, Ian. 2002. “Making up People.” In Historical Ontology, 99–114. Cambridge, MA: Harvard University Press. Hill, Jane H. 1998. “Language, Race, and White Public Space.” American Anthropologist 100: 680–89. Horn, Laurence. 1984. “Toward a New Taxonomy for Pragmatic Inference: Q-Based and R-Based Implicature.” In Meaning, Form, and Use in Context: Linguistic Applications, edited by Deborah Schiffrin, 11–42. Washington, DC: Georgetown University Press. Irvine, Judith T., and Susan Gal. 2000. “Language Ideology and Linguistic Differentiation.” In Regimes of Language: Ideologies, Polities, and Identities, edited by Paul V. Kroskrity, 35–83. Santa Fe: SAR Press. Jakobson, Roman. 1960. “Closing Statement: Linguistics and Poetics.” In Style in Language, edited by Thomas A. Sebeok, 350–77. Cambridge, MA: The MIT Press. Keenan, Elinor Ochs. 1976. “The Universality of Conversational Postulates.” Language in Society 5: 67–80. Levinson, Stephen C. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Morris, Charles. 1971 [1938]. “Foundations of the Theory of Signs.” In Writings on the General Theory of Signs, 13–71. The Hague: Mouton & Co. Nichols, J. 1984. “Functional Theories of Grammar.” Annual Review of Anthropology13: 97– 117. Putnam, Hilary. 1975. “The Meaning of ‘Meaning.’” In Mind, Language and Reality: Philosophical Papers, Volume 2, 215–71. Cambridge: Cambridge University Press. Ross, John Robert. 1970. “On Declarative Sentences.” In Readings in English Transformational Grammar, edited by Roderick A. Jacobs and Peter S. Rosenbaum, 222–77. Waltham, MA: Ginn and Company. Searle, John R. 1969. Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Silverstein, Michael. 1976. “Shifters, Linguistic Categories, and Cultural Description.” In Meaning in Anthropology, edited by Keith Basso and Henry Selby, 11–55. Albuquerque: University of New Mexico Press. Stasch, Rupert. 2011. “Ritual and Oratory Revisited: The Semiotics of Effective Action.” Annual Review of Anthropology 40: 159–74. Weinreich, Uriel, William Labov, and Marvin I. Herzog. 1968. “Empirical Foundations for a Theory of Language Change.” In Directions for Historical Linguistics, edited by Winfred P. Lehmann, 95–195. Austin: University of Texas Press. 

Pragmatics JAMES SLOTTA University of Texas at Austin, USA Pragmatics is a branch of linguistics that investigates the ways language is tied to the contexts in which it is used. As this definition indicates, pragmatics coalesces as a distinct and coherent domain of inquiry only in relation to the study of language abstracted from its use in context, which has been the prime focus of both twentieth-century linguistics and philosophy of language. The topics typically discussed under the heading of pragmatics arise from a variety of difficulties and impasses encountered in the analysis of language extracted from context; and as a result, they compose a motley collection, including deixis, presupposition, speech acts, implicatures, politeness, information structure, and so on. As an organizing framework, this survey highlights two of the principal ways in which twentieth-century philosophy and linguistics have constituted objects of study that are context-independent. These decontextualized objects of study form the backdrop against which many of the diverse issues and phenomena that make up the field of pragmatics come into view. As such, they provide a useful vantage point from which to survey the field. The first mode of abstraction considered here is one that privileges a single function of speech – the propositional function – over others (e.g. expressive, performative). This unifunctional abstraction common in twentieth-century analytic philosophy takes language to be composed of a set of sentences that describe states of affairs in ways that may be judged true or false. As such, language becomes unmoored from context. After all, the truth of the sentence “There are an odd number of planets in the solar system” can be evaluated independently of who, when, where, or why it is uttered. This proposition-centered view of language has been challenged by a number of philosophers, linguists, and anthropologists who point up a variety of other functions of speech that tie language more closely to its context of use. With increased attention to the non-propositional functions of language, a new set of questions arise: How are the variety of non-propositional functions of language to be identified and analyzed? How are non-propositional functions signaled in language form? Are propositional and non-propositional functions of language independent of one another? Or, are they intertwined in ways that necessitate a more holistic account of language that does not limit considerations of context to the “peripheral” functions of language? The second mode of abstraction considered here, prominent in twentieth-century linguistics, may be termed the lexical-grammatical abstraction of language. In this mode of abstraction, a language is composed of a set of linguistic form-types (e.g. lexemes, morphemes, syntactic configurations), each with distinctive meanings or functions that recur across their varied token occurrences. In other words, invariance The International Encyclopedia of Linguistic Anthropology. Edited by James Stanlaw. © 2021 John Wiley & Sons, Inc. Published 2021 by John Wiley & Sons, Inc. DOI: 10.1002/9781118786093.iela0323 2 P R A G M AT I CS across contexts – a kind of context independence – is a fundamental feature of the form-types that serve as units of analysis in the sort of distributional analysis common to descriptivist, structuralist, and generativist linguistics alike. How this lexical-grammatical idealization of language relates to both propositional and non-propositional functions of language remains a persistent issue, giving rise to recurring questions about the relationship of language form to language function: is the well-formed sentence – the maximal domain in the lexical-grammatical analysis of language – the proper domain of analysis for non-propositional functions of language? Is it even sufficient for an analysis of the propositional function? Are additional frameworks beyond grammar and lexicon needed to account for the connection of linguistic forms and (non-)propositional functions in context? The aim here is not to review all of the answers to the questions introduced so far, but to provide a sense of the ways pragmatics as a domain of inquiry has been constituted through some of the more influential responses that have been offered, focusing in particular on the Anglo-American linguistic and philosophical traditions. Pragmatics, propositionality, and the plurifunctionality of language The field of pragmatics – or at least the term “pragmatics” – emerges at the confluence of two philosophical schools: American pragmatism and logical positivism. The term was coined by Charles Morris, who derived it from the name of the American philosophical school associated with Charles Sanders Peirce, William James, and John Dewey. Morris’s account of the functioning of signs draws heavily on the semiotic framework of Peirce, to which he adds an additional trichotomy distinguishing three branches of the study of signs (i.e. semiotics): (i) syntactics, which considers the relationship of sign forms to one another; (ii) semantics, which concerns the relation of sign forms to what they stand for (i.e. their designata); and (iii) pragmatics, which is to be the study of the effects of signs on interpreters, and therefore concerned with “all the psychological, biological, and sociological phenomena which occur in the functioning of signs” (Morris [1938] 1971, 43). In addition to American pragmatism, Morris’s approach to semiotics also reflects the influence of the then regnant logical positivist view of language. In this view, and that of much early twentieth-century analytic philosophy, natural language is composed of a host of “systematically misleading expressions” that give rise to philosophical errors and confusions. Clarity was to be achieved through the logical analysis of natural language expressions and, when that proved insufficient, the wholesale reconstruction of language on a logical basis. As with analytic philosophy more generally, the logical positivist program centered on language in its propositional capacity – viz. its capacity to describe reality in ways that may be judged true or false. More specifically, the role of philosophy in the logical positivists’ view was to ensure the logical rigor and clarity of scientific statements. Morris’s colleague Rudolf Carnap sought to advance this project by developing logical P R A G M AT I CS 3 “syntaxes” of science, consisting of rules for the ways the sentences of scientific observation and theory may be formed and transformed into one another in truth-preserving fashion. As part of this program, pragmatics was largely irrelevant, given that it concerned neither empirically verifiable representations of denotata (the domain of semantics) nor the logical formation and transformation of propositions (the domain of syntax). As a result, the domain of pragmatics was given little consideration as part of efforts to clarify natural language or formulate ideal languages for science. But for those interested in taking languages as they are and not as they should be, pragmatic considerations of contextualized usage are unavoidable even in a highly circumscribed, propositional view of language. Sentences containing indexical expressions, as in (1), cannot be judged true or false without information about their context of use: 1. Your age is now an odd number. Indexical expressions, including the possessive pronoun “your,” the “present” tense of the verb, and the temporal adverb “now,” tie the sentence to its context of use. That is to say, the propositional content of (1) cannot be resolved and a judgment of the proposition’s truth made without specifying the context in which (1) is spoken. In keeping with the spirit of logical positivism, indexical expressions were generally excluded from logical reconstructions of language. But, notably, they are found in every natural language, displaying diverse ways of tying the propositional content of utterances to their context of use (e.g. tenses, pronouns, moods, temporal adverbs, spatial demonstratives, evidentials, and many others). Indexical expressions are thus perhaps the most conspicuous way in which pragmatic considerations have a bearing on propositionality. But more recent philosophical discussions of meaning have put forward a number of other ways in which propositionally relevant meanings of words and expressions are context dependent (see the final section for further discussion of this point). In contrast to logical positivist “ideal language philosophers,” so-called ordinary language philosophers directed attention to everyday uses of language as a way to resolve or dissolve philosophical issues. This shift in focus prompted increasing attention to the non-propositional functioning of language, with J.L. Austin’s account of “performativity” among the most influential exemplars (Austin 1962). Austin begins his attack on the propositional view of language with an account of what he terms performative utterances – “systemically misleading expressions” that appear to be propositions but are not actually evaluable as true or false. For instance, the expression 2. I apologize for acting oddly. appears identical in form to indicative sentences that can be judged true or false, such as: 3. She apologized for acting oddly. Unlike (3), though, (2) is not a sentence that would normally be judged true or false, according to Austin. Rather, it appears to be the performance of an act, namely the 4 P R A G M AT I CS act of apologizing. As an act and not a propositional use of language, its occurrence in context may be judged to be successful or unsuccessful, but not true or false. Indeed, performative utterances are closely linked to the context in which they occur. To be successful, they require that certain contextual conditions (i.e. felicity conditions) be met; and if successful, they give rise to a particular transformation of the context. At first, Austin appears to be carving out a domain of performative utterances that is distinct from the sort of propositional uses of language that made up the bread and butter of the philosophical interest in language (i.e. “constative utterances,” in Austin’s parlance). But Austin goes on to undermine his own distinction of constative and performative utterances. In contrast to explicit performative utterances as in (2) – which have a regular form (first-person subject, verb in the unmarked “present” tense, verbs that name the very act being performed) – Austin points out that there are a variety of what he calls primary performative utterances, which take an unruly, even limitless, variety of forms: 4. I’m sorry for acting oddly. 5. I was being an idiot. (4) and (5) may also function as apologies but they do not announce themselves as such. Moreover, they prove hard to classify as either a performative or constative utterance. Is (4) a description of how one feels and so subject to evaluation as true or false (“No you aren’t!”); or is it an act that is to be judged felicitous or infelicitous (“I accept your apology”)? It functions like an apology (performative), but the success of the apology depends on the truth of what is said (constative), insofar as actually being sorry is a felicity condition for a successful apology. In the process of undermining his own distinction between constative and performative utterances, Austin undermines the possibility of saving the propositional view of language by quarantining it within the domain of constative utterances. Many performative utterances have a constative dimension, as in examples (4) and (5). And, constative utterances that describe the world all have a performative dimension; they occur as part of acts of asserting, declaring, and the like, and so they too have felicity conditions. All utterances, Austin concludes, are uttered as part of what he calls an illocutionary act – an action performed with words (e.g. inquiring, ordering, requesting, promising, asserting, declaring). In the plurifunctional view of language Austin winds up espousing, pragmatic considerations of contextualized use cannot be avoided, regardless the type of utterance. The grammaticalization of pragmatics and the pragmaticization of grammar Though Austin’s conception of illocutionary acts offered an expanded view of the functioning of language, Austin took a formally conservative approach to identifying and analyzing these functions. The sentence-length illocutionary acts that are the prime P R A G M AT I CS 5 domain of Austin’s account look very much like the sentence-length propositions examined by analytic philosophers and, significantly, the well-formed sentences that have been the maximal domain of analysis of grammarians. Even more, Austin presents abstract sentence-types as the bearers of illocutionary forces in much the same way that words or lexemes are the bearers of the relatively context-independent “meanings” listed in a dictionary. For instance, an explicit performative sentence-type (as in example [2]) appears to bear the same illocutionary force and have the same felicity conditions whenever it is uttered. Illocutionary acts thus appear amenable to a sort of lexical-grammatical analysis that links linguistic form-types to their recurring and distinctive meaning or function across different occurrences. Indeed, John Searle refined Austin’s account by identifying a range of “illocutionary force indicating devices,” including mood markings, word order, intonation, special morphology, and even the propositional content of sentences themselves, as loci where illocutionary forces are coded (1969). From this perspective, the lexical and grammatical features of sentence-types encode illocutionary forces (e.g. the interrogative “Did he act like an idiot?” typically has the force of an inquiry, the imperative “Don’t act like an idiot” the force of a directive). And, as a result, illocutionary forces can be incorporated as components of sentence grammar. This grammaticalization of illocutionary forces reached an extreme with the Abstract Performativity Hypothesis (APH), which proposed that all sentences have an explicit performative clause in their deep structure (e.g. “[I DECLARE that] … ”). Though it appears absurd, perhaps, the hypothesis was supported by some quite subtle observations about the distribution of lexemes in sentences that gave the hypothesis plausibility. For instance, John Robert Ross (1970) noted that the difference in grammaticality between sentences like (6) and (7) could be explained assuming the APH. On this account, the first- and second-person reflexive pronoun in (6) is licensed by the firstand second-person pronouns in the abstract performative clause (“[I DECLARE to you that] Fools like myself/yourself … ”). The third-person reflexive pronoun is not similarly licensed by anything in the performative clause, and so (7) is ungrammatical (“[I DECLARE to you that] Fools like himself … ”): 6. Fools like myself/yourself often fall in love. 7. *Fools like himself often fall in love. In the course of a sentence’s derivation, the deep-structure performative clause may be deleted, thus generating inexplicit, primary performative sentences (as in [6]). Illocutionary acts were thus rendered a feature grammatically coded in the deep structure of sentences. Ross also proposed an alternative “pragmatic” analysis of (6) and (7) that nicely illustrates the inverse approach; instead of grammaticalizing speech acts as the APH does, Ross’s alternative analysis pragmaticized grammar. In this analysis, the firstand second-person reflexive pronouns in sentences like (6) are licensed not by a deep structure performative clause, but by the co-presence of speaker and addressee in the context of utterance. In effect, pragmatic information about contextual conditions would be introduced into the syntactic component of a grammar. Though Ross was rather dismissive of this alternative, functional theories of grammar have continued to 6 P R A G M AT I CS explore the interface of pragmatics and grammar (see Nichols 1984 for a useful if dated review). In a similar vein, discourse analysis has explored aspects of sentence grammar (e.g. topicalization constructions, clefts of various sorts, pronoun distribution) that appear to be influenced by the larger discourse context and distinctions relevant to it (e.g. given and new information, topic and focus, theme and rheme). The efforts of Generative Semanticists like Ross and functional grammarians to blur the boundaries between syntax, semantics, and pragmatics have highlighted how complex the relationship of syntax and pragmatics is. Presupposition provides a good example. Some linguistic forms, termed presupposition triggers, indicate that particular background assumptions are held to be true. For instance, definite noun phrases like “the king of France” in the sentence “The king of France is a fool” presuppose that there is a king of France at present. The presupposition remains even when the sentence is negated: “The king of France is not a fool,” presupposes that there is a king of France. Similarly, factive verbs like “regret” presuppose the truth of their complement clause: “She regrets that he acted oddly” and “She does not regret that he acted oddly” both presuppose that he acted oddly. These presuppositions are considered pragmatic properties of utterances because they can be canceled in some contexts. For instance, “I don’t regret having acted oddly, because I didn’t” does not presuppose that I acted oddly. Moreover, they are pragmatic properties that are regularly, though not invariably, associated with particular linguistic forms that have distinctive syntactic properties (see [8]). 8a. Factive verbs occur with a wide variety of gerunds: I regret/resent/deplore your falling in love. 8b. While non-factive verbs do not: *I believe/conclude/claim your falling in love. 9. Some factive verbs, though, do not occur with a wide variety of gerunds: *I know/realize your falling in love. This suggests that grammatical analysis should take account of presuppositions as properties of lexical items and the syntactic configurations they occur in. But, in the case of presupposition triggers, as in the case of many pragmatic phenomena, pragmatic distinctions do not map neatly onto syntactic ones (see the factive verbs in [9]). The significance of pragmatic considerations for grammatical analysis is often not straightforward and remains an area of continuing inquiry. While the importance of pragmatic considerations in grammatical analysis is an open question, the attempt to grammaticalize speech acts represented most strikingly in the Abstract Performativity Hypothesis proved untenable. The problem with assimilating speech acts to the lexical-grammatical meaning of sentence types is evident in the case of indirect speech acts, i.e. utterances whose illocutionary force does not match the force suggested by their lexical-grammatical form. 10a. I want you to bury the turtle. 10b. The turtle is beginning to smell. P R A G M AT I CS 7 Though the utterances in (10) might be deemed assertions that can be judged true or false due to the use of the indicative mood, they can also be used with the force of a request. They thus appear to be an indirect means of conveying an illocutionary force that would be more directly conveyed with an imperative or an explicit performative: 11a. Bury the turtle! 11b. I request that you bury the turtle. The illocutionary force of indirect speech acts, then, is not simply a product of their semantic and grammatical properties as sentence types; consideration of the uses of tokens in context appears necessary to explain how indicative sentences can take on the force of requests. Sentence meaning, speaker meaning, and pragmatic rationality Indirect speech acts point up the fact that illocutionary forces are not (always) encoded by the grammatical and lexical features of sentence types. An influential alternative to this coding view of speech acts is one that involves a kind of inferential reasoning about linguistic forms in context. In such a view, participants recognize that the utterance of (10b) in context is not in fact functioning (merely) as an assertion about the odor emitted by the turtle. Assuming that the addressee has some responsibility for the turtle, they could then infer that there is a covert point to the statement: namely, that (10b) is actually a request that the addressee do something about the turtle. Indirect speech acts point up the difference between what has been termed “speaker meaning” and “sentence meaning.” While we might say the “sentence meaning” of (10b) has the illocutionary force of an assertion, the speaker “means” (or intends) something different by it; namely, they mean to issue a request or directive, as in the “sentence meaning” of (11b). This distinction is pertinent to a variety of different situations in which meaning in context is seen as not reflecting the “literal” meaning of an utterance (e.g. hints, irony, metaphor, and a range of other conventional and conversational implicatures). The distinction between speaker meaning and sentence meaning, developed notably in the work of Paul Grice, allows for a neat distinction of semantics (sentence meaning) from pragmatics (speaker meaning). Gricean pragmatics distinguishes itself as a branch of linguistics concerned with (i) a kind of meaning based on speakers’ intentions (as opposed to the context-independent meaning of lexical and grammatical forms) and (ii) a kind of inferential reasoning that allows participants to connect what is said to what is meant (as opposed to the rule- and convention-based accounts offered in grammatical and semantic analyses). Among Grice’s most influential contributions is a framework for conversational implicatures, according to which the speaker meaning of utterances in a cooperative conversation is largely a function of the meaning of the underlying propositions (sentence meaning), the context in which they occur, relevant background information, 8 P R A G M AT I CS and several purportedly universal, rational maxims of conversation (Grice 1975). Drawing on these elements, participants can imply meanings – implicatures – that go beyond the propositional meaning of the utterances. For instance, a letter of recommendation for a professorship in pragmatics that reads in its entirety: 12. The applicant has excellent penmanship. flouts what Grice terms the “maxim of quantity”: the speaker offers much too little information to assess the merits of the job candidate. But, this violation of the maxim can be made to conform with the assumption that the letter writer is being cooperative if one infers that the letter writer has nothing else good to say about the candidate. The implicature of the too-brief utterance in (12), then, is that the candidate is not well qualified for the job. As this example illustrates, getting from the sentence meaning of “they have excellent penmanship” to the speaker meaning that “they are not well qualified for this job” involves a great deal of background knowledge about the norms that govern language use in this context (e.g. what is the accepted purpose of the exchange, what counts as being as informative as required, what counts as being relevant to the communicative activity, and so on; see Keenan 1976 for discussion of this issue). In fact, given all of the accessory information about norms that is necessary to arrive at the implicature, it is not entirely clear what the cooperative principle and its maxims add. Once one knows what a letter of recommendation is supposed to look like and that people try not to say negative things in them, it is fairly clear that (12) is a violation of one genre-specific norm (i.e. the expected length and informativeness of a letter of recommendation). And one can infer (among other possibilities) that the violation is part of an effort to uphold another genre-specific norm (i.e. not saying negative things in a letter of recommendation). In short, the cooperative principle and its maxims are so general – capable of generating so many possible inferences – that they must be restricted through reference to norms of language use and interpretation. Yet, once these norms are spelled out, it is often not clear what the maxims themselves add to the analysis. Grice’s account of implicatures has been influential in defining the domain pragmatics (relating speaker meaning to sentence/utterance meaning) and singling out the tools particular to it (implicatures and inferences, rather than rules). The approaches to pragmatics that have grown out of Grice’s framework preserve these general features while seeking to constrain the power of Grice’s maxims, which in Grice’s formulation are capable of transforming virtually any sentence meaning into any speaker meaning. The study of generalized conversational implicatures (GCIs) – as opposed to the particularized conversational implicatures that Grice focused on – has been an area where linguists have had some success in formalizing and constraining Gricean implicatures. GCIs are nearly automatic implicatures that occur unless the context indicates otherwise – that is, they occur in virtually any context as something like “default interpretations,” to use Stephen Levinson’s expression (2000). They are thus much less context sensitive than particularized conversational implicatures and, therefore, more like semantic meanings in terms of their regular association with linguistic forms. P R A G M AT I CS 9 A canonical example is “scalar implicatures,” which occur with quantifiers (none, some, all), numbers, and other linguistic scales (e.g. warm/hot, possible/certain). An implicature-based account of the meaning of quantifiers begins with the somewhat unintuitive claim that a sentence like (13) has the same sentence meaning as (14): 13. 14. 15. 16. 17. Some pragmatists have excellent penmanship. At least some pragmatists have excellent penmanship. Some pragmatists have excellent penmanship – in fact they all do. **Some pragmatists have excellent penmanship – in fact none do. Some but not all pragmatists have excellent penmanship. This account of the sentence meaning is supported by the fact that a sentence like (15) does not involve a contradiction: the sentence in (13) is consistent with the statement that “all pragmatists have excellent penmanship.” ([16], on the other hand, does involve a contradiction.) Of course, (13) typically is used with the same meaning as (17). This meaning can be derived as a scalar implicature grounded in something like Grice’s maxim of quantity (e.g. “say as much as you can,” “be as informative as possible”): had the stronger claim that all pragmatists have excellent penmanship been warranted, it would have been incumbent on the speaker to say that, so the speaker meaning in (17) is implicated by the statement of (13). Scalar implicatures and other GCIs are thus pragmatic meanings, but ones that depend in a limited and negative way on contextual factors. Moreover, they are identified with limited and structured domains of the grammar and lexicon (e.g. scalar implicatures apply to linguistic scales of various sorts). With such limitations, it has proven significantly easier to formalize and constrain the implicatures generated by utterances. But even then, context and cultural knowledge can complicate efforts to provide readily generalizable, formal accounts of them. For instance, one needs to account for the fact that “The apartment is big enough for families with three children” implicates “at most three children” and not “exactly three children.” Relevance theory is an alternative attempt to streamline and constrain the Gricean framework, reworking Grice’s maxim of relevance as a cognitive principle. In effect, relevance is taken to be the optimization of (i) the cognitive effort involved in deriving inferences from what is said and (ii) the positive cognitive outcomes of those inferences. How cognitive effort and cognitive outcome are to be measured remain issues to be worked out. In addition to Gricean implicatures, relevance theorists have sought to account for the inferential work that goes into interpreting what they term “explicatures,” a range of pragmatic enrichment processes (e.g. disambiguation, reference resolution) that go into making up the propositional content of utterances. For instance, the context in which (18) is uttered may allow a hearer to determine whether “my book” means “the book that I lost” or “the book that I wrote”: 18. Have you seen my book? Explicatures, like GCIs, are thus an instance of what might be called “truth conditional pragmatics,” to use Francois Recanti’s expression, an approach to propositional meaning in which context-sensitive pragmatic meanings play a part. Truth conditional 10 P R A G M AT I CS pragmatics argues that pragmatic reasoning about language use in context enriches the literal, semantic meaning of utterances and therefore shapes the propositional content expressed by those utterances. Propositions thus become the concern of pragmatics. Grice-inspired implicature- and inference-based approaches to pragmatics have also been developed to account for non-propositional functions of utterances. In Brown and Levinson’s influential account of politeness (1987), polite ways of speaking arise at the intersection of two conflicting aims: (i) the speaker’s desire to preserve the public self-imagine, or “face” of the hearer while, at the same time, (ii) engaging in a face-threatening act that undermines the hearer’s self-image. The indirectness made possible by Gricean implicatures provides an optimal means of accomplishing these two goals at once. Indirect speech acts, such as those in (10), allow the speaker (i) to issue a directive that, in imposing limits on the autonomy of the addressee, constitutes a face-threatening act while (ii) doing so indirectly, thus mitigating the face threat and, in consequence, being polite. As a framework that spells out indirect ways of conveying speaker meanings, the Gricean account of implicatures opens the way for indirectness itself to serve social pragmatic functions like politeness. Brown and Levinson’s theory of politeness highlights the versatility of the Gricean framework, which was initially developed to explain the communication of information. Building on Grice’s insights, neo-Gricean pragmatics has sought to account for everything from lexical-grammatical phenomena like intrasentential anaphora and gaps in the lexicon to social-pragmatic phenomena like politeness using a limited set of pragmatic principles. Laurence Horn, for instance, posits only two: the q-principle (“Say as much as you can”) and the r-principle (“Say no more than you must”) (1984). But the hope for a formalized and universally applicable pragmatics – akin to a universal grammar – inevitably runs up against the need to account for cultural particularities of language use, raising a number of questions that neo-Gricean pragmatists are currently attempting to address: can cultural particularities be accounted for through universal principles and maxims that concern the transmission of information? Or, will other functions of language need to be incorporated into the Gricean framework through the addition of other principles, as some theories of politeness have argued? And, crucially, how are additional maxims, principles, or constraints to be identified and incorporated so as to avoid blending cultural particularities and universal tendencies in an ad hoc way that would undermine the universalizing and formalizing aspirations of this program? Sociocultural pragmatics: Cultural norms of social action In contrast to the universal aspirations of much linguistic pragmatics, socially, culturally, and historically variable norms and regularities of language use in context have been a prime object of study in both sociolinguistics and linguistic anthropology. Variationist sociolinguistics, for instance, has considered the way speech forms vary based on the social and stylistic context in which they occur. The ethnography of communication highlighted the ways in which culturally variable norms shape language use as a socially significant activity. And symbolic interactionism, ethnomethodology, P R A G M AT I CS 11 interactional sociolinguistcs, and conversation analysis have in different ways attended to the diverse norms that structure linguistic interaction, viewed as a site where participants manage their identities and relationships, and coordinate their interactional activity through the use of language. The Prague Linguistic Circle’s plurifunctional perspective on language was an important forerunner of these approaches, particularly the influential work of Roman Jakobson. Going beyond the traditional focus on sentence grammar, these linguists explored the multiplicity of functions performed through the use of language; that is, not only stating propositions about the world but also indicating the attitudes and identities of speakers, affecting addressees, constructing cohesive discourse, and so on. And, importantly, they regarded language form as a medium through which these varied functions are realized. Building on this plurifunctional view, sociolinguistic and anthropological attention to language form has not been restricted to abstract sentence types and their constituents. Rather they have regarded everything from sociophonetic variation to positions in conversational sequences to the “social meanings” associated with the use of different languages in interaction as units relevant to the functioning of language in context. Indeed, formal variation that appears to be unconstrained from a lexical-grammatical perspective (i.e. “free variation”) has been found to be structured – an “orderly heterogeneity” in Weinreich, Labov, and Herzog’s terms (1968) – when norms and regularities of language use in context are considered. Integrating this plurifunctional view of language with the sensibility of American cultural anthropology, Dell Hymes proposed a research program he dubbed “the ethnography of communication”, which investigated the diverse functions of language across different speech communities. The ethnography of communication joined the study of language use with ethnographically rich accounts of social and cultural life, which offered the kind of contextual framework deemed necessary to understand the functioning of language in context. In concert with sympathetic folklorists and sociolinguists, ethnographers of communication investigated a host of genres of language use, emphasizing the performance in context of many, such as myth narrative, and poetry, that had long been analyzed apart from their use in context. Their work highlighted the structure and poetics of verbal artistry and mundane verbal interaction alike, with an eye toward the social and cultural context in which they played a part. Michael Silverstein brought a number of the concerns of these socially, culturally, and interactionally attuned perspectives on language together under the heading of “pragmatics,” conceptualized as the study of what C.S. Peirce termed “indexical” signs – signs that have a real (e.g. causal, co-present) connection to their object (Silverstein 1976). Indexical signs tie language to its context of use in different ways. “Non-referential indexes” encompass the wide range of speech forms that point to features of the context without employing the referential or propositional machinery of language. Viewed as non-referential indexes, the use of sociophonetic variants, honorifics, gendered “languages”, codes-switches, and the like all more or less consciously point to features of the context – the identity of speakers, the relationship of speakers to addressees, and much else. 12 P R A G M AT I CS “Referential indexes” – or deictics, as they are more widely known – tie the propositional meaning of utterances to the moment in which they are uttered (see example [2]); in other words, they conspicuously tie what is said propositionally to the act of saying it. Referential indexes thus serve as important resources for acting in the world by talking about it. For instance, Austin’s explicit performative utterances involve the use of deictics in a particular configuration (first-person subject, second-person (in)direct object, “present” tense) that maps the propositional content of explicit performatives (i.e. what is talked about) onto the very moment in which the proposition is uttered. In this way, the propositional capacity of language is mobilized to transformative pragmatic effect by describing what is happening in that instance of speech (e.g. “I promise to bury the turtle” describes the act that it performs). Self-referential speech acts like explicit performative utterances are merely one way in which referential indexes can be used as part of language-mediated social action. The study of ritual and oratory (Stasch 2011) provides a particularly rich site for exploring alternative ways in which the power of speech derives from the mapping of what is said onto the act of speaking itself. Non-referential indexes too can play an important role in transforming the world through speech. A non-referential index occurring in interaction may either conform with the context as already constituted up to that point in the interaction; or it may transform the context by pointing to new elements that thence become part of the context. Such transformative uses of non-referential indexes – e.g. shifting into a more formal speech style or using an “informal” pronoun with someone for the first time – are in some respects very much like speech acts. Like the “illocutionary acts” Austin identified, they are verbal acts that transform the context in which they occur. But unlike Austin’s account of speech acts, these are verbal acts accomplished without the use of the propositional or referential capacity of language. From this indexical-centered perspective, speech acts arise from the dynamic interrelation of speech forms indexing a changeable context, regardless of whether those forms involve the propositional or non-propositional capacities of language, and regardless of whether they are sentence-length, shorter, or longer. A prime analytic domain of sociocultural pragmatics is thus the communicative event and the wide range of signs, linguistic and non-linguistic, contained therein. The multimodal array of signs produced in communicative events may be considered together in terms of their textuality, i.e. the way signs cohere with other signs that make up their co-text. Cohesion is forged through the use of linguistic elements like anaphors and pronouns. But textuality also encompasses the cohesive relationship of indexes of identity, social relationship, and so on, comprising linguistic and non-linguistic signs alike. Through such textual arrays of linguistic and non-linguistic signaling in communicative events, participants are seen as taking on social roles, identities, and relations, in effect constituting themselves as social beings in a social world (Bucholtz and Hall 2005). Unusual and unexpected configurations of signs provide for the possibility for new, “metaphorical” or “tropic” meanings, understandings, actions, identities, and relationships to emerge in interaction as well, as when non-Spanish-speaking Anglo-Americans use bits of Spanish to present themselves as playful and easy-going, and simultaneously devalue Spanish and those associated with it (Hill 1998). P R A G M AT I CS 13 This perspective thus shares with performativist approaches to gender (Butler 1990) and social constructivist theories of identity (Hacking 2002) the view that the social world is performatively constituted in communicative events. The iteration of signs across events of use and the intertextual links that tie one communicative encounter to another thus become an important area for investigating broader institutional processes and political projects through which social identities and relationships are stabilized and transformed over time and space. Attention to reported speech and other ways speech is de- and re-contextualized in new contexts has been central to this approach as a way of tracking how speech events and interactants are linked to one another in complex participation frameworks. Significantly, the sort of pragmatic frameworks that normatively and/or recurrently link signs to contexts in contemporary sociocultural pragmatics are themselves generally taken to be context-bound. Pragmatic considerations of contextualized use of language thus extend to the level of metapragmatic discourse and norms (i.e. discourse and norms that concern language use in context), whose historical emergence in particular social, cultural, economic, and political environments has itself become an object of study. Research on language ideologies and processes of enregisterment considers how socioculturally situated, reflexive models of language use arise from and give shape to linguistic practices and language form. In this approach, the universal principles analysts look to uncover are not metapragmatic norms and regularities themselves, which are held to vary across time and space; rather, attention is given to systematic constraints on people’s awareness of language structure and use, as well as recurring processes through which language ideologies or metapragmatic frameworks are formed (Irvine and Gal 2000). Viewed as a dialectic process, language ideologies and registers emerge historically from conditions of language use in context and in turn (re)shape language use in context. In this way, the context-bound practice of interpretation that is so central to Peirce’s pragmatic conception of semiosis has reemerged in discussions of the sociocultural pragmatics of language. The result is a pragmatist (in the philosophical sense) view of pragmatic signification, grounded in processes of use and inference that are themselves context-bound, historically contingent, and in part reflexive in nature. The pragmatic makings of propositionality As sociolinguists, linguistic anthropologists, and others have explored the multiple functions of language that center on indexical relations of speech to its contexts of use, philosophers have probed ways in which the propositional capacity of language is itself tied inextricably to contextualized uses of language. As noted earlier, Austin’s discussion of speech acts was radical in the context of the logical positivism of its day in arguing that the propositional capacity of language is embedded in acts of using language. Determinations of truth and falsity, then, are bound to speech-act considerations like felicity conditions. For instance, 19. The woman burying a tortoise over there is acting oddly. 14 P R A G M AT I CS may be judged true or false even if the woman happens to be burying a turtle, not a tortoise. Though nothing is denoted by the noun phrase “the woman burying a tortoise over there,” what renders (19) a functioning proposition is that the noun phrase successfully refers to someone for participants in the interaction. Propositionality, then, depends in part on the success of a speech act – reference – in context and not entirely on a decontextualized sense of the correct denotation of terms. Causal theories of reference propose an even more thorough-going relationship between speech acts, contextual use, and propositional meaning. According to this account of reference, proper names come to refer to an individual through an illocutionary act of naming (i.e. a baptism) that fixes a name to its referent. Future uses of the name are parasitic on that originary act, causally connected to it through a history of usage that links later acts of reference to the initial baptism. The meaning of a name is not dependent on the descriptive associations that might be tied to the name, but on a history of use. Hilary Putnam has offered a similar account of natural kind terms, i.e. terms denoting groups of objects that share “natural” properties. Their denotation, he argued, is grounded in a real relation between a natural kind term and paradigmatic examples of it (1975). On this account, a term like “water” denotes all things sufficiently similar to the paradigm examples. There is an indexical connection between the stuff denoted by natural kind terms and the term itself. In Putnam’s account, this indexical connection is mediated by a social division of linguistic labor in which experts determine what entities are sufficiently similar to the paradigm examples to be included in the set of things denoted by the term (i.e. its denotata). While non-experts may associate certain properties with natural kind terms, these are neither necessary nor sufficient to identify its denotata. Rather, meaning is grounded in a contextual relationship of terms to the objects they denote and a social organization of expert knowledge and authority. Denotation, in this view, is rendered pragmatic to a considerable degree. Putnam’s arguments were developed as an alternative to holistic accounts of the meaning of scientific terms, which viewed meaning as a product of the theories in which they are embedded. The theory-dependent or theory-relative nature of meaning holds, for instance, that the meaning of the term “gold” was fundamentally different after it was determined to be an element with an atomic number of 79. When John Locke was writing in 1689, he did not understand the term “gold” in these terms, but as a “metal” characterized by “yellowness, great weight, ductility, fusibility and solubility in aqua regia, etc.” For a theory-dependent view of meaning, the development of atomic theory radically transformed the meaning of “gold.” But such a view of meaning as fundamentally context-dependent and radically changeable flies in the face of the intuition that “gold” denotes (roughly) the same things before and after the emergence of atomic theory. Putnam’s externalist semantics was an effort to defend that intuition by grounding meaning in the external world rather than in theories, so that terms can retain their denotation even as theories change. Putnam’s causal theory taken alone, though, makes it difficult to account for reference change and failures of reference. Terms like “phlogiston” and “witch” lack denotata (at least from the perspective of many philosophers) despite their causal connections with real-world paradigm exemplars at one time in the past. A hybrid descriptive-causal P R A G M AT I CS 15 theory of reference that remedies this defect but retains some key elements of externalist semantics posits an array of pragmatic considerations that go into the denotation of natural kinds, none of which alone is decisive: (i) there are real (indexical) connections between paradigm examples of denotata and the application of natural kind terms; (ii) there is a social division of linguistic labor with experts who adjudicate the denotational meaning of the terms for a community; and (iii) terms have an intensional meaning informed by historically changeable theories, beliefs, and interests. In some cases, the collapse of a theory renders terms without an extension (e.g. “witches” and “phlogiston”); in other cases, terms retain their connection to paradigm examples despite changes in theory (e.g. “gold”). Such a view of meaning shares a great deal with performativist or social constructivist views of denotation that stress the role of social and cultural context, power relations and interests, and even particular languages themselves in shaping the meaning of words and expressions. This constructivist perspective is often associated with poststructuralism, evident for instance in Foucault-inspired accounts of the way institutionally situated discursive formations “make up people” (Hacking 2002). But it is also a key element in the structuralist theories of Benjamin Lee Whorf and Ferdinand de Saussure, which do not regard language as reflecting the structure of reality but as imposing structure on it. Though linguistic meaning is shaped by context in all of these approaches, the contexts are quite different, ranging from state institutions to the structure of language itself. But in each case, these constructivist views render linguistic meaning ultimately pragmatic in nature, bound to some contextually variable feature of the social, cultural, or linguistic environment. SEE ALSO: Austin, J.L. (John Langshaw); Chomsky, and the Chomskyan Tradition vs. Linguistic Anthropology; Classes of Signs; Context and Contextualization; Conversation Analysis (CA); Conversational Maxims and Conversational Implicature; Deixis and Indexicals; Discourse Analysis; Distributional Method; Ethnography of Speaking and Communication; Ethnomethodology; Face; Fishman, Joshua; Genre; Goodenough, Ward; Grice, H. Paul; Harris, Zellig; Hymes, Dell; Jakobson, Roman; Language and Gender; Language and Identity; Language Ideology; Language and Social Class; Linguistic Anthropology, History and Development of; Modality, Multimodality; Morris, Charles W.; Naming; Narrative; Norms in Language and Communication; Peirce, Charles Sanders; Performance, Linguistic and Communicative; Performativity vs. Indexicality; Poetics, Poetry, and Ethnopoetics; Politeness; Practice(s) and Practice-based Approaches; Pragmatism; Presupposition and Entailment; Reported Speech and Represented Speech; Ritual and Forms of Communication; Sentence as Unit of Analysis; Silverstein, Michael; Sociolinguistics; Sociophonetics; Structuralism; Style and Stylization; Symbolic Interactionism; Syntax; Whorf, Benjamin Lee; Wittgenstein, Ludwig REFERENCES AND FURTHER READING Austin, J.L. 1962. How to Do Things with Words. London: Oxford University Press. 16 P R A G M AT I CS Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Bucholtz, Mary, and Kira Hall. 2005. “Identity and Interaction: A Sociocultural Linguistic Approach.” Discourse Studies 7: 585–614. Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. Grice, H. Paul. 1975. “Logic and Conversation.” In Syntax and Semantics, Volume 3: Speech Acts, edited by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press. Hacking, Ian. 2002. “Making up People.” In his Historical Ontology, 99–114. Cambridge, MA: Harvard University Press. Hill, Jane H. 1998. “Language, Race, and White Public Space.” American Anthropologist 100: 680–689. Horn, Laurence. 1984. “Toward a New Taxonomy for Pragmatic Inference: Q-based and R-based Implicature.” In Meaning, Form, and Use in Context: Linguistic Applications, edited by Deborah Schiffrin, 11–42. Washington, DC: Georgetown University Press. Irvine, Judith T., and Susan Gal. 2000. “Language Ideology and Linguistic Differentiation.” In Regimes of Language: Ideologies, Polities, and Identities, edited by Paul V. Kroskrity, 35–83. Santa Fe: SAR Press. Jakobson, Roman. 1960. “Closing Statement: Linguistics and Poetics.” In Style in Language, edited by Thomas A. Sebeok, 350–377. Cambridge, MA: MIT Press. Keenan, Elinor Ochs. 1976. “The Universality of Conversational Postulates.” Language in Society 5: 67–80. https://doi.org/10.1017/S0047404500006850. Levinson, Stephen C. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Morris, Charles. (1938) 1971. “Foundations of the Theory of Signs.” In his Writings on the General Theory of Signs, 13–71. The Hague: Mouton. Nichols, J. 1984. “Functional Theories of Grammar.” Annual Review of Anthropology 13: 97–117. https://doi.org/10.1146/annurev.an.13.100184.000525. Putnam, Hilary. 1975. “The Meaning of ‘Meaning.’” In Mind, Language and Reality: Philosophical Papers, Volume 2, 215–271. Cambridge: Cambridge University Press. Ross, John Robert. 1970. “On Declarative Sentences.” In Readings in English Transformational Grammar, edited by Roderick A. Jacobs and Peter S. Rosenbaum, 222–277. Waltham, MA: Ginn and Company. Searle, John R. 1969. Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Silverstein, Michael. 1976. “Shifters, Linguistic Categories, and Cultural Description.” In Meaning in Anthropology, edited by Keith Basso and Henry Selby, 11–55. Albuquerque: University of New Mexico Press. Stasch, Rupert. 2011. “Ritual and Oratory Revisited: The Semiotics of Effective Action” Annual Review of Anthropology 40: 159–174. Weinreich, Uriel, William Labov, and Marvin I. Herzog. 1968. “Empirical Foundations for a Theory of Language Change.” In Directions for Historical Linguistics, edited by Winfred P. Lehmann, 95–195. Austin: University of Texas Press.