Language Evolution: The Windows Approach addresses the question: "How can we unravel the evolution of language, given that there is no direct evidence about it?"
The LINGUIST List is dedicated to providing information on language and language analysis, and to providing the discipline of linguistics with the infrastructure necessary to function in the digital world. LINGUIST is a free resource, run by linguistics students and faculty, and supported primarily by your donations. Please support LINGUIST List during the 2016 Fund Drive.
Review of The Oxford Handbook of Philosophy of Language
EDITORS: Lepore, Ernest; Smith, Barry C. TITLE: The Oxford Handbook of Philosophy of Language PUBLISHER: Clarendon Press YEAR: 2006
Anne Reboul, L2C2 CNRS-UMR5230, Institute for Cognitive Sciences, Lyons, France, email@example.com
The book is divided into 8 unequal parts, on historical context, the nature of language, the nature of meaning, the nature of reference, semantic theory, linguistic phenomena, varieties of speech acts and the epistemology and metaphysics of language.
The editors defend their approach (i.e., presenting the philosophical debates concerning language by asking the protagonists themselves to express their views) and outline the book organization in a brief ''Preface''. The first part, THE HISTORICAL CONTEXT, opens with a chapter (by Heck and May) devoted to ''Frege's contribution to philosophy of language''. The authors point out Frege's innovations, beginning with the ''notions of scope and binding and how they are used to represent generality'' (4) in his logicist program, with its emphasis on compositionality and on the strong relation between logic and truth. This highlights compositionality not only between predicates and arguments, but between truth-functional (sentential) operators and the truth-values of the sentences on which they apply (the referents of the sentences for Frege). Thus, reference itself is ''essentially compositional'' (19). This leads to the distinction between sense and reference as a Fregean, ''a distinction within content'' (23), sense being taken by Frege to determine reference. This is the main point of contention between Fregeanism and contemporary philosophy of language, few contemporary philosophers accepting that sense determines reference.
Chapter 2, by Beaney, deals with ''Wittgenstein on language: from simples to samples''. The famous ''linguistic turn'' of the first half of the twentieth century can be traced back to Wittgenstein and the Tractatus Logico-Philosophicus. While Wittgenstein first held (in the Tractatus) that the meaningfulness of language is grounded in the existence of simple objects, he came to hold that it rests in fact in use. Propositions are organized in systems (e.g., 'A is green/red/yellow/etc.'), only one of them being usable in a given situation. Thus, the meaning of a word is not its bearer, but its use in different language-games and Wittgenstein extends his conception of language from logic to grammar. A rule is not an ordinary proposition or sentence but a tautological proposition lacking sense. Beaney concludes with a distinction permanently held by Wittgenstein, despite his intellectual evolution, between meaningful propositions and logical propositions (= tautologies).
Chapter 3, ''Philosophy of language in the twentieth century'', by Baldwin, deals with the central place language has occupied in philosophy during the twentieth century, largely replacing epistemology, which had been central since Descartes. Baldwin traces the history of philosophy of language through Frege (emphasis on a ''logically reconstructed language''), Wittgenstein's Tractatus and Russell's theory of description, logical empiricism (Carnap's ''logical syntax of language''), Quine's ''indeterminacy of translation'', Davidson's principle of charity and anomalous monism, and the later Wittgenstein, and his contention that most, if not all, philosophical problems stem from ''misunderstandings of our everyday language'' (87), ultimately leading to Austin's speech act theory, Strawson's notion of presupposition, and Grice's non-natural meaning and logic of conversation. Baldwin then turns to recent times, and to the possibility, proposed by Chomsky, that natural language does not have a semantics, but only a syntax and a pragmatics, a position Baldwin links to contemporary post-Gricean pragmatics, such as Relevance Theory.
Part II, ''THE NATURE OF LANGUAGE'', opens with chapter 4: ''Psychologism'', by Travis, relating the accusation of ''psychologism'' (a negative word in philosophy since Frege) to a concern with ''answerability'', which, in the end, translates to a concern with ''truth''. The correctness of a judgment has to depend on the way things are, INDEPENDENTLY of any thinker. This leads Travis to what he calls ''The Martian Principle'', according to which ''No thinker, or stance, could be answerable to anything to which any other thinker, in principle, could not be'' (114). However, some logical laws may apply to some thoughts but not to others. This leads Travis to the conclusion that ''Frege's picture of logic is a very subtle form of psychology'' (126).
Bezuidenhout, in the 5th chapter, speaks about ''Language as internal'', defining Chomsky's internalist view of language as the idea that ''languages are properties of the mind/brains of individuals and supervene entirely on the internal states of these mind/brains'' (127). Bezuidenhout then reminds us that Chomsky's anti-externalist crusade has two targets, which she dubs ''language externalism'' (denying that languages supervene on mind/brain states) and ''semantic externalism'' (where reference of mental representations depends on the physical environment and not on the brain states of language users). Finally, Bezuidenhout shows that Chomsky's internalism is linked to his scientific individualism (his insistence on the primacy of the individual and his or her idiolect and rejection of the very notion of ''languages'' as socially shared constructs). Idiolects are the only possible objects of scientific investigations.
Higginbotham, in chapter 6 (''Languages and their idiolects: their language and ours''), pursues the same topics, outlining a dialectic between the fact that language is a social phenomenon, subject to variations across space, time and individuals, and the individualist view, emphasizing the individual and his/her idiolect, independently of his/her social surroundings. The social view does not have to deny the existence of idiolects in a given social environment, or the conventionality of language. That, however, does not mean that these social aspects of language are relevant to a scientific investigation of language.
Part III, THE NATURE OF MEANING, opens with a chapter by Wilson on ''Rule-following, meaning, and normativity''. It thus turns back to Wittgenstein and the notion of rule in applying a word, formulating the Wittgensteinian problem as a paradox: ''The paradox says that, since any action can be interpreted as being either in accord or conflict with the rule, it makes no sense to suppose of any action that it is in accord or conflict with the rule, i.e., either the expression of the rule or the substantive rule that is meant to determine the instances of accord and conflict'' (152). This is of course a skeptical argument and Wilson outlines different ways it can be understood. However, non-factualism about meaning can be seen as self-defeating and any account of meaning should ''validate our intuitive conception of the objectivity of judgment'' (168). Anti-realist accounts may fail to satisfy that constraint, though Wilson argues that this basically boils down to the fundamental philosophical dissensions about the notion of ''truth''.
Papineau centers the 8th chapter on the currently very influential ''Naturalist theories of meaning''. These theories try to account for representations in a naturalist framework (as defined by the natural sciences), usually taking linguistic representations as derivative on (basic) mental representations, thus usually subscribing to some version of the language of thought hypothesis. Basically, the naturalist program is based on the possibility of reducing representation to other scientific categories. One approach is ''inferential role semantics'', where representational content of a concept is taken to be its inferential role. This approach meets with several problems and Papineau then turns to causal theories, accounting for the content of mental states in terms of what causes them, and outlines Millikan's teleosemantics. This bottom-up view distinguishes between mechanisms which produce representations from those which ''consume'' them, thereby linking behaviour to pursuit of biological functions, where biological functions are defined in historical terms (of either phylogenetic or ontogenic evolution).
Segal, in the 9th chapter, turns to ''Truth and meaning'', beginning, inevitably, with Frege's establishing an indefeasible link between truth and logic. Tarski introduced a semantics in terms of truth definitions, along the line of Convention T (''s'' is true iff p), using the notion of translation in that the consequent of the conditional is a translation of its antecedent (''s''). Davidson drew from Quine the notion that there may be several equally good translations for a given expression, but eschewed drawing skeptical conclusions from that. He developed a truth theory from Tarski's T-theory (based on Convention T), adding to it an instrumentalist (and social) approach to reference, based on assent and dissent. As Segal points out, this is contradictory with the (highly non-social) linguistic theory developed by Chomsky, and Segal goes on to develop a cognitivist perspective on Davidsonian semantics, approaching the problem of reference from the Chomskian distinction between competence and performance. Finally, Segal defends such a semantic approach against a possible objection based on the context relativity of word meaning (e.g., color words), advocating an indexical account supplementary to the semantics.
Chapter 10, by Pagin, is devoted to ''Meaning holism'', the view that the meaning of a given linguistic expression depends on its relations with the meanings of other linguistic expressions in the same language, and possibly with the meanings of all the expressions in that language. Pagin traces the origins of meaning holism to logical empirism (notably to its verificationism) and its aftermath, noting the central roles in later times of Quine, Sellars and Davidson, in the emergence of various types of meaning holism, inferential role semantics, and belief holism. Pagin then turns to counter-arguments to meaning holism, before giving arguments in its favor.
Chapter 11, by Weir, is devoted to Quine's well-known notion of “Indeterminacy of translation”. Quine sees language as a social art, and indeterminacy of translation is the conclusion of an argument which begins with the thought experiment of ''radical translation'' (i.e., the ''gavagai'' thought experiment). Basically Quine's argument is that two behaviourally equivalent sentences (produced in the same situations) can be intuitively non-synonymous and indeed distinct in truth-value, extending the idea to the ontological relativity or inscrutability of reference for names and predicates. Weir points out that Quine's position, in the end, is a relativist and anti-realist one, in contradiction to his endorsement of realism.
In chapter 12, Borg discusses ''Intention-based semantics'' and begins by distinguishing a mild form of intention-based semantics, which acknowledges the fact that semantics is based on both intention and conventions, dubbed ''A-style intention-based semantics'' and a much stronger form according to which intentional speakers play an active role in determining semantics in that meaning depends on their current state of mind, dubbed ''B-style intention-based semantics''. The main difference, according to Borg, between A-style and B-style semantics is that B-style semantics takes interpretation arrived at without taking into account the speaker's current intentions as yielding non-propositional or non-truth-evaluable items, i.e., propositional schemas or incomplete logical forms. Borg criticizes B-style semantics precisely on the ground that it considers that a theory of literal truth-conditional meaning should be ''simply subsumed within a theory of communication'' (260). She defends the specificity of linguistic meaning on the basis of linguistic convention (i.e., the codic nature of language) and promotes sentences as the bearers of semantic content.
In chapter 13, Schiffer discusses ''Propositional content'', beginning with an approximate definition of propositional content as what that-clauses contribute to what is ascribed in such sentences as ''Ralph believes/says/etc. that Tony Curtis is alive''. Face-value approaches of belief reports (A believes that S) take them to be true when A stands in the belief relation to S and considers that that-clauses (''S'' in the present case) are propositions. Schiffer examines non-propositional alternatives, non-relational accounts of believing (not plausible, according to him), and relational but non-propositional accounts of believing (no more satisfactory than the non-relational accounts). Schiffer concludes that propositional content remains a problematic notion.
Greenberg and Harman tackle ''Conceptual role semantics'' (CRS) in chapter 14, linking it with the Wittgensteinian notion that meaning is determined by use, in this specific case use in thinking. Specifically, meaning can be DETERMINED by conceptual role, without being EQUIVALENT to conceptual role. Though one's meaning is determined by one's uses, one's understanding of one's meaning does not consist in understanding one's uses. CRS is compatible with both translation and compositionality, The authors discuss a few objections to CRS, concluding to the pursuit of the ongoing debate regarding these different objections to CRS.
Farkas deals with ''Semantic internalism and externalism'' in chapter 15, beginning with the tension inherent in recognition of the fact that ''even if the meaning of an expression is determined by social agreement, grasping the meaning of the word is an individual psychological act'' (323). He describes the Twin Earth argument, as highlighting the incompatibility between thinking, on the one hand, that what a speaker means by a word is not determined by circumstances outside of her and thinking, on the other hand, that meaning determines reference as well as that the content of a sentence determines its truth-conditions. This ''externalist'' view, according to which, famously, ''meaning ain't in the head'' (Putnam 1975), was extended to mental contents. Farkas then proceeds to confront externalism with contradictory, ''internalist'' theories, noting that, pace Putnam, there is no incompatibility between ''the two assumptions that meaning determines reference, and that meaning is internal'' (333).
Carston and Powell outline ''Relevance theory -- new directions and developments'' in chapter 16, beginning by an outline of relevance theory, and going on to its view on the semantics-pragmatics distinction, pointing out that relevance theory has advocated a strong pragmatic involvement in the determination of propositional content, based on the notion that both lexical meaning and propositional content are context-sensitive. This is seen in the recent developments of relevance theory, not as deriving implicatures, but as processes of mutual adjustment of the meaning of the words in the utterance. Carston and Powell then quickly describe recent results in experimental pragmatics, notably regarding scalar implicatures, which vindicate the Relevance Theory ''nonce'' approach to implicatures over the neo-Gricean ''default'' approach. Finally, they outline some future directions, having to do with early communication and developmental pathologies (autism, Williams syndrome).
Szabo returns to ''The distinction between semantics and pragmatics'' in chapter 17. He bases his discussion on the very possibility of agreement and disagreement which depends on mutual understanding, which, in turn, depends on a common interpretation of the linguistic expressions used. Noting that mutual understanding does not require agreement about paraphrase, Szabo points out that the distinction between verbal and substantive disagreements is not exhaustive, given the existence of contextual disagreement. Verbal disagreement concerns the semantics of expressions, while contextual disagreement concerns pragmatics. Thus ''semantics and pragmatics [are] subfields within the general study of utterance interpretation'' (379). Szabo argues in favour of the inclusion of some information about speaker's intentions in the context, and advocates abandonment of the dubious notion of what is said.
Part IV, THE NATURE OF REFERENCE, opens with a chapter (the 18th) by Sainsbury on ''The essence of reference'', which begins by noting that reference implies agents, words or concepts and things (the referents). Referring expressions have a number of properties: semantic simplicity, uniqueness of referent, understanding as knowledge of the referent, scopelessness, being rigid designators. Singular referring expressions have at most one referent (uniqueness) and this leads to the notion that understanding a referring expression involves knowing its referent. Sainsbury then goes on to discuss whether definite descriptions are or are not referential, introducing Donnellan's distinction between referential and attributive uses. He notes that the crucial debate is on whether referential uses are to be accounted for in semantic or pragmatic terms, refusing to take a stand on this issue.
In chapter 19, MacBride investigates ''Predicate reference'', asking whether predicates are, or not, referring expressions. Russell viewed predicates as either referring expressions or reflecting the activity of the mind, and, given that the second possibility would be tantamount to abandoning objectivity, predicates should be referring expressions. MacBride notes however that Quine questioned whether Russell's list of possibilities was exhaustive and proposed a disquotational account, though Quine's disquotational theory falls foul of a problem familiar from discussions of Tarski's theory, the difficulty of applying the concept 'predicate' to novel strings (it fails to say what a predicate is). MacBride discusses Frege's insight that intersubstitutability salva veritate is the controlling principle of referring expressions. He however points out that this Reference Principle itself is subject to well-known difficulties having to do with failures of substitution, which leads him to the conclusion that failures of substitution between predicates and names should not be taken as indicating lack of co-reference. This, however, does not establish that predicates refer and his conclusion acknowledges the persistence of a mystery concerning the nature of predication.
Sosa speaks of ''Rigidity'' in chapter 20, defining it as the fact that an expression refers ''to one and the same thing WITH RESPECT TO any POSSIBLE SITUATION'' (476, Sosa's emphasis). His first endeavor is to specify what is meant by the expression 'with respect to', and he says that ''this specification takes place in the actual situation, in which the expression is used''. Proper names are rigid, while definite descriptions are not. From the fact that Kripke extended the notion of rigidity to natural kind terms, and the fact that any general term can indeed be used rigidly, Sosa raises the possibility that one should think of a term as rigid only if it has this feature essentially (rather than as a matter of use).
Braun pursues the same type of questions in chapter 21 on ''Names and natural kind terms''. He begins with the Millian Theory of proper names, according to which what proper names contribute to the proposition expressed are their referents. The four objections to that theory are the Objection from Cognitive Significance (i.e., identity judgments are not trivial), the Objection from Belief Ascription (i.e., one can believe that Mark Twain is Mark Twain and not believe that Mark Twain is Samuel Clemens), the Objection from Meaningful Sentences Containing Non-Referring Names, and the Objection from Negative Existentials. This theory, together with fatal objections against alternative ''descriptive'' accounts, has led to a new proposal, the Causal Theory of Reference, which claims that proper names get their reference from a causal chain beginning at the baptism of the referent. The Millian Theory was extended, by Kripke, from proper names to natural kind terms (e.g., 'water', 'gold', 'tiger', etc.).
Bach, in chapter 22, pursues the topic of reference, asking ''What does it take to refer?''. He begins by restricting reference to singular reference. He then distinguish between speaker reference (what speakers do when they are said to refer) and linguistic reference, listing properties for each of them, which he then develops one by one, beginning with those associated with speaker reference. His basic point is that ''much of what speakers do that passes for reference is really something else, and much of what passes for linguistic reference is really nothing more than speaker reference'' (518). This is because ''it is one thing for a speaker, when using an expression in a certain way, to express a thought about a certain object and quite another for the expression to stand for that object, even relative to the context'' (552).
Part V, SEMANTIC THEORY, opens with a chapter (the 23rd) by King, on ''Formal semantics'', tracing its beginning to Tarski's theory of truth. After a review of Carnap's and Kripke's work, he turns to Montague who ''assigned intensions, functions from world and times to extensions, to English expressions'' (567). Kaplan introduced two indices where Montague had a single one, one representing the context of utterance, while the second represents its circumstance of evaluation, linking them to two kinds of meaning. King then outlines the development of dynamic semantics, through a concern for ''conversational dynamics'', by, notably, Kamp and Heim and its modifications by Groenendijk and Stokhof, through ''dynamic predicate logic'', rather than ''Discourse Representation Theory'' or ''File Change Semantics''.
Chalmers, in chapter 24, introduces ''Two-dimensional semantics'', tracing its source to the Fregean distinction between sense and reference. He reworks these two notions as, respectively, intensions (functions from possible worlds to extensions) and extensions. As he reminds us, two expressions can share their extensions without sharing their intensions, as shown by the old example of 'cordate' and 'renate'. A natural suggestion is that two such expressions would have the same extension in, e.g., the actual world but different extensions in some possible worlds (where not all renate organisms, for instance, would have hearts, contrary to what happens in our world). This is cashed out in a matrix in which the left column represents what turns out in the actual world (roughly, the world of the speaker) while the second column represents what turns out in a counterfactual world (not the world of the speaker). The core of two-dimensional semantics is that every expression token has a primary intension, a secondary intension and a two-dimensional extension. Applications of two-dimensionalism are Fregean senses, contents of thoughts, belief ascriptions, indicative conditionals, and conceivability and possibility.
Bar-On and Simmons deal with ''Deflationism'' in chapter 25. Deflationism is the position according to which truth is not a genuine property and this raises the question of how to ''make sense of our pervasive truth TALK'' (607, authors' emphasis). Bar-On and Simmons begin by enumerating varieties of deflationism, from disquotationalism to illocutionary deflationism, through minimalism, redundancy theory and the prosentential theory of truth, and then discuss a few objections to deflationism: the possibility that it is self-defeating, that it is not stateable, the problem of its scope, the fact that it presupposes some semantic notions. The consequence of this last problem is that truth is not an autonomous notion, contrary to what deflationists pretend, raising the worry that deflationism might fail at deflating all truth.
Part VI, LINGUISTIC PHENOMENA, begins with ''Compositionality'' (chapter 26, by Dever). Compositionality is at the root of generativity (the possibility of producing an infinite number of different — and new — sentences from a limited number of components), and can be simply defined as the fact that the meanings of complex expressions are derived from the meaning of their parts. Dever then reviews two varieties of compositionality theories, compositionality as functionality and compositionality as substitutability, turning then to discuss how compositionality can be non-trivial in a mathematical sense, i.e., how it can be a stringent constraint on the grammar of a language. He concludes that ''compositionality thus enforces a variety of honesty in semantic theory'' (663).
Chapter 27, by Richard, focusing on ''Opacity'', pursues the intersubstitutability problem. Richard notes ''that there seems to be a lot of opacity in our language'' (667), raising the questions of whether this is indeed so, and, if it is, of why it should be so. The term 'Opacity' was introduced by Quine, to characterize some linguistic contexts in which substitution salva veritate of coreferential expressions is not possible, making opacity a case of nonextensionality. Setting aside quotational contexts, Richard asks whether there is such a thing as opacity, concluding that indeed there is, but that it is not clear that it is an homogeneous phenomenon (opacity in quotation and opacity in complements seem quite different).
In chapter 28, Ludlow treats ''Tense'', asking whether tense is a property of our language and thoughts or ''a property of aspects of the external world'' (688). He begins by noting that tense marking seems scarce outside of Indo-European languages, which of course does not mean that tenseless languages do not express past, present or future, suggesting that tense might be in the world, after all. He opposes two different approaches, dubbing the defenders of the first the 'detensers' (they reject tenses from the metalanguage) and the defenders of the second the 'tensers' (they don't). In the first approach, tenses are represented by a semantic value of some kind in the metalinguistic semantic formula corresponding to the utterance; in the second, tenses are eliminated in favour of a ''relative ordering of events on a time line'' (693), on the basis of a primitive temporal relation 'earlier-than/later-than'. Ludlow then reviews arguments for and against each type of approach, showing the advantages and inconveniences of both, and concluding that ''by having a firm enough grasp on BOTH approaches we afford ourselves a deeper insight into the nature of tense itself'' (714, Ludlow's emphasis).
Schein turns to the problem of ''Plurals'' in chapter 29, linking it to the intricate questions raised by the distinction between mass and count terms, as well as by the porosity of that distinction. He extends the question to events (quantifiable in a Davidsonian approach).
Edgington examines ''The pragmatics of the logical constants'' in chapter 30, i.e., 'and', 'or', 'if', 'not', 'all' and 'some'. The question is whether one should, following Grice, consider that the linguistic terms have the same meaning as the corresponding logical constants, or whether the discrepancies between their linguistic use and their logical characterization (in terms of truth tables, for instance) indicates differences in meaning. She points out that this question is linked to that of the semantics-pragmatics distinction: its answer is directly related to an answer to the question of whether discrepancies are matters of use (pragmatics) or of meaning (semantics). Through a discussion, mainly, of conditionals, conjunction and negation, and of different approaches in addition to Grice's, Edgington comes to the conclusion that pragmatic approaches may have been overrated in the specific case of logical constants.
Glanzberg discusses ''Quantifiers'' in chapter 31, introducing Generalized Quantifier Theory, where expressions of generality (e.g., 'everything', 'nothing') are treated as second-level properties (properties of properties: thus 'everything' contributes to sentences ''the second-level property of being a property under which everything falls'' (796)). Properties apply to sets of individuals and this is cashed out inside a universe of discourse. Natural languages have more quantifiers than the universal and the existential (and their negations), including such quantifiers as 'few', 'both', 'enough', etc. These usually correspond to relations between sets inside a universe of discourse. Glazberg lists some constraints which quantifiers satisfy and concludes to the effect that quantifiers have ''significant semantic and grammatical implications'' (818).
In chapter 32, Pietrovski deals with ''Logical form and LF'', beginning with a comparison between patterns of reason and traditional grammar. The main question is whether ''all valid inferences are valid by virtue of propositional structure'' (823), as was traditionally thought from Aristotle's treatment of syllogisms until Frege and Russell, who divorced between logical form and grammatical structure. The emergence of transformational grammar in the seventies, with the distinction between surface and deep structures, and the resultant identification between logical form and deep structure meant that ''many apparent examples of grammar/logic mismatches were rediagnosed as mismatches between different aspects of GRAMMATICAL structure'' (836, Pietroski's emphasis).
Part VII, VARIETIES OF SPEECH ACT, opens with a chapter on ''Metaphor'' by Reimer and Camp. The authors begin by noting that metaphors are currently considered some sort of conceptual phenomena, in that, though intelligible, they produce a conceptual tension. Contemporary accounts share a tendency to downplay differences among tropes as well as differences between figurative and nonfigurative language. The authors conclude that recent research has taken a strong pragmatic turn and that contemporary investigation sees metaphor as a species of context-sensitive communication.
Boisvert and Ludwig concentrate in chapter 34 on ''Semantics for non-declaratives'', and begin by pointing out that mood and force should be dissociated in that there is not a one-to-one correspondence between a sentence's grammatical mood and a corresponding utterance's illocutionary force. Yet a semantics for non-declaratives should explain the connection between mood and force, while giving a uniform and compositional description of the contribution of mood to sentences, explaining why non-declarative sentences are not truth-evaluable, accounting for mixed-mood sentences, for quantification into mood markers, and explaining the distribution patterns in mixed-mood sentences (e.g., the impossibility of imperatives in the antecedents of conditionals). The authors advocate an approach in terms of satisfaction and sincerity conditions, thus obtaining the right measure between a too tight or too loose link between mood and force.
Hornsby dedicates the 35th chapter to ''Speech acts and performatives'', beginning with the classification of speech acts and the relations between force, indirect speech and implicatures, before turning to illocution. She describes the alliance proposed by Strawson between Austin's notion of illocution and Grice's notion of non-natural meaning, closing the section with a look at taxonomies of speech acts. The chapter ends with an analysis of the notion of ''performativity'' and its possible application in politics.
Part VIII, THE EPISTEMOLOGY AND METAPHYSICS OF LANGUAGE, opens with Stainton's 36th chapter on ''Meaning and reference: some Chomskian themes'', supporting Chomsky's claim that a scientific approach of semantic reference is impossible. Having reviewed two major insights of Chomsky, the innateness and modularity of language, he then describes the view he intends to reject, that is, the notion that there are word-world semantic relations, outlining three negative arguments: the ''radical'' argument from ontology (there are no public languages, but only individual idiolects and hence no public words to relate to the world); the ''moderate'' argument from ontology (there are public words, but there can be no science of the relata -- words and world objects -- because neither are recognised as such by science); the failure of compositional referential semantics (sentences fail to express propositions which can be attributed truth values INDEPENDENTLY of -- non-linguistic -- world knowledge). Stainton proposes a positive alternative to referential semantics, in terms of an internal semantics.
In chapter 37, Smith, once again relying on a Chomskian position, tries to clarify ''What I know when I know a language'', intimately linking questions about the nature of linguistic knowledge to questions about the nature of language (this is where Chomsky comes in). Smith begins by noting that language IS a source of knowledge, but that knowledge of language itself remains mysterious, especially given language generativity which implies that we understand myriads of utterances which we have never been confronted with before. Smith then discusses linguistic acquisition, insisting that inductive approaches just will not do, and endorses Chomsky's proposal in terms of an innate language acquisition device. He then comes back to the possibility of knowing public languages and shows that it is difficult to account for such knowledge, because, at best, such public languages are abstractions for individual idiolects, which makes not only speakers' knowledge, but also the very notion of public languages in the sense of speaker's languages, elusive.
In Chapter 38, Miller talks about ''Realism and antirealism'', opening his paper by noting that Dummett, taking the centrality of philosophy of language for granted, replaced the debate between realism and antirealism inside philosophy of language, attacking semantic realism. Semantic realism is the claim that even undecidable sentences can be understood through a grasp of their truth-conditions, and, hence that these truth-conditions are evidence-transcendent. Dummett's two arguments against semantic realism are the acquisition argument (we could not acquire a language whose truth-conditions are evidence-transcendent) and the manifestation argument (we could not manifest our understanding of language through our use of it if truth-conditions are evidence-transcendent). Miller discusses arguments for or against Dummett's view, but notes that all supporting arguments beg the question in as much as they presuppose that realism is an ESSENTIALLY SEMANTIC theory, a dubious view at best.
Glüer deals with ''Triangulation'' in chapter 39, a central notion in Davidsonian semantics. Basically, triangulation has to do with the view that to think, entertain propositions, etc., a creature must have the concept of objective truth, which, in turn, supposes that it has experienced a specific sort of interaction, in which it and other creatures ''sufficiently like itself'' (1006) interact with objects in the world. This is taken as a criterion of correctness of a sort, both in order to attribute thought to a creature, and in order to assess the correctness of our own thought, a requirement which, according to Glüer, leads to infinite regress. As well, she spots an antirealism or idealism in the requirement of triangulation. She notes that triangulation is poorly social in that it does not imply interaction between the creatures, but mere observation. She advocates a stronger requirement, such as mutual knowledge, noting the strong circularity in the triangulation argument.
Cappelen and Lepore coauthored chapter 40 on ''Shared content'', noting a strong tension between the generally acknowledged context-sensitivity of sentences and the obvious fact that the same sentence basically means the same thing ''across a wide range of different contexts'' (1020), using reported speech as the basis for their arguments. They defend their own view, pluralistic minimalism, based on a rejection of speech act monism, the (implicit) constraint that a given utterance corresponds to a single thing (proposition or thought). Pluralistic minimalism is thus a combination of speech act pluralism (a given utterance can express a whole bunch of propositions) and semantic minimalism. Same-saying corresponds to the overlap between utterances.
The final and 41st chapter, written by the late Donald Davidson, bears on ''The perils and pleasures of interpretation''. Davidson begins by noting the difference between ''the difficulties that stand in the way of explaining in detail how we manage to find out what is in other people's minds and the relative ease with which we do it in practice'' (1057). This leads him to a pessimistic view of the project of naturalization of the mind: our interpretation of other minds will not be reduced ''to a branch of the natural sciences'' (1056). Davidson begins with the skeptic's concern about our certainty that others have minds, let alone minds like our own, and remarks that answering that concern is linked to saying how we interpret other minds. The main difficulty regarding that second question is the holism of conceptual thought, ''for what we observe is the product of many cognitive factors'' (1058). However, ''the difficulty is not in the practice, but in the theory'' (1061). Here Davidson introduces the notion of triangulation, as a key factor for a creature's having the notion of error, which helps us attribute thoughts to other minds if we trust others to, on the whole, be not only truthful but correct about current situations.
This mammoth book (1,083 pages including the index) should be read by anyone with an interest not only in philosophy of language, but in semantics and pragmatics, and even, though less centrally, in syntax. Though not introductory in the sense that it could be read by a first year student, it is well worth the effort of reading and, given the overall clarity of the chapters, accessible. The quality of the papers is sustained throughout and is of the highest standard.
Thus, the Oxford Handbook of Philosophy of Language should be in any university library, and, indeed, should be on the bookshelves of anyone with an educated interest in language.
REFERENCES Putnam, Hilary (1975) Mind, Language and Reality: Philosophical Papers, volume 2, Cambridge, Cambridge University Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER:
Anne Reboul is a Senior Researcher at the French Center for Scientific
Research (CNRS) in France. She has a Ph.D. in Linguistics (EHESS, Paris)
and a Ph.D. in philosophy (University of Geneva, Switzerland). She has
written some books, among which are an Encyclopedic Dictionary of
Pragmatics and quite a few papers in French and English, on pragmatics
and/or philosophical subjects. She has developed an interest in recent
years in both language evolution and animal cognition and communication and
has recently published a book on Language and human cognition. A longer
version of this review is available on her web site