Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Philosophy of language is a ‘philosophy of’ discipline, concerned with conceptual issues centring on the nature, origin and purpose of language, in all its multifarious uses. As such, it is a relatively new part of philosophy, unlike, say, metaphysics or ethics. The present chapter focuses on philosophy of language in Australasia, but because philosophy of language is a field with an impressive prehistory/early history, we begin our discussion by looking at central features of that early history. Later sections cover major themes in philosophy of language as it has developed in Australasia.

Language and Analytic Philosophy

As a discipline, philosophy of language is continuous with, and arose out of, the preoccupation with the notion of an ideal language that helped to inaugurate analytic philosophy. That preoccupation, notably in the work of Frege and Russell, was characterised by a suspicion of ordinary language and its failure to conform to various logical norms. This suspicion continued with the work of the logical empiricists and even that of a successor of logical empiricism like Quine, with his insistence that language should be regimented if it was to be of use for the expression of a properly scientific worldview. These philosophers were, to a lesser or greater degree, reformers. Another reaction to the work of Frege and Russell took the view that language was not so much to be reformed as understood for what it was: an instrument of communication and much else that was well suited to its many purposes and whose articulation was not beholden to formal logic. This was the view of Peter Strawson, who attacked Russell on just such grounds and along the way made use of a notion—presupposition—that had played a role in Frege’s account of some of the failures of ordinary language but was now refashioned to support the idea that meaning was use. (The latter doctrine was integral to the later Wittgenstein’s philosophy, although Wittgenstein himself was notoriously unwilling to develop a theory of meaning as use.)

Strawson’s critique of Russell and his defence of ordinary language appeared in 1950s, and it is not unreasonable to trace the emergence (or, if you prefer, the maturing) of philosophy of language as a discipline in its own right to roughly this time—a discipline with its own set of problems and agendas, overseen by a reflective attitude to its own subject matter. This was a decade that saw not only Strawson’s response to Russell on behalf of ordinary language (as well as Russell’s response to Strawson on behalf of ideal language) but also Quine’s sceptical attack on the notion of meaning, including such attempts as Carnap’s Meaning and Necessity (1947) to set up a scientifically rigorous account of meaning, as well as Strawson and Grice’s influential critique of Quine’s arguments. Orthogonal to these developments, the 1950s also saw the first remarkable stirrings of the new linguistics, in the work of Noam Chomsky. From its earliest manifestations, the new linguistics had a very different picture of the role that the notion of meaning might play in a scientific account of language.

In short, this was a time when a number of starkly different, indeed incompatible, views of the nature of language and the relation of language to logic as well as psychology began to appear: the reformist approach, the ordinary language approach and the psychological approach. It is fair to say that the USA headquartered the reformist wing, at least in part because of the influence of the logical empiricists who had moved there prior to the advent of WWII. Quine, while a critic of the logical empiricist ‘dogmas’ of reductionism and the analytic/synthetic distinction, retained the belief that logic was the appropriate vehicle for ‘limning’ the structure of the world—first-order logic, however, not second-order logic, since the latter involved the notion of an attribute, which he thought required the philosophical legitimacy of the notion of meaning. In Britain, things were rather different. Because of the influence of Wittgenstein and the work of Strawson, formal logic was seen as having a more limited role. In addition, the theory of meaning began to occupy even more of the centre stage with the work of Paul Grice, who initiated a bold program that tried to understand the notion of linguistic meaning in terms of a notion of intention based utterance meaning. The decade that followed saw an explosion of work on this and other topics, inspired by a new-found confidence in ordinary language, its manifold uses and our ability to discuss these rigorously. In particular, John Austin inaugurated speech act theory, and this development meant that (apparently) non-descriptive uses of language were also made subject to rigorous study. The successes of such a preoccupation with language helped to foster the movement in British philosophy known as ‘ordinary language philosophy’, with its ideological belief that philosophical problems in general could be solved through a study of the way we actually use language—a movement that should be distinguished from the more sober view that language is something that deserves philosophical investigation in its own right.

The linguist Chomsky had stressed the productivity of natural language in his defence of the psychological conception of language: the way speakers can construct and recognise a potential infinity of grammatical strings. A new phase in the debate about natural language meaning came when Donald Davidson inaugurated a new kind of truth-conditional theory of meaning that appealed to productivity in the context of a theory of meaning. Davidson wanted to explain the capacity of agents to grasp the meanings of a potential infinity of sentences on the basis of a finite learning process but in a way that was not compromised (as far as Quine and his followers were concerned) by its association with Fregean notions. Instead, Davidson appealed to Tarski’s account of truth and, in particular, Tarski’s convention T. More even than Grice, Davidson thereby inaugurated an entire research program for understanding the nature of meaning for natural language, and he did so in a way that respected the role of formal logical frameworks while allowing for modes of speech that were not purely descriptive.

While Davidson was prepared to work within the strictures of Quine’s sceptical perspective on intensional notions, Richard Montague, a student of Tarski, developed a far richer truth-conditional approach to the issue of meaning, one that was based on intensional logic with a rich type theory and a possible-worlds model-theoretic semantics. So Montague used tools that Davidson, following Quine, rejected on ideological grounds. All this was taking place in the 1960s, a particularly rich, innovative period for the burgeoning field of philosophy of language because of the way an emphasis on the theory of meaning was combined with a renewed commitment to formal techniques. The late 1960s saw an integration of some of this work with certain other concerns, such as the question of the meaning of names and descriptions that had initially ushered in the analytic turn in philosophy at the hands of Frege and Russell. Kripke launched his influential attack on descriptivist accounts of names, presenting an alternative ‘causal’ picture of reference and defending the view that names were (de jure) rigid designators, while Keith Donnellan, Hilary Putnam and David Kaplan used related arguments to distinguish names and other ‘directly referential’ terms like demonstratives from descriptions. (Kripke also appealed to rigidity to argue for a category of a priori contingent truths as well as a posteriori necessary ones, and gave an influential argument against the identity of mental and physical states based on the failure of such identities to fit the pattern of a posteriori necessary identities).

All of these authors were committed to the use of modal notions, especially that of possible worlds—notions that had been anathema to Quine. But from the point of view of Australasian philosophy, the most influential American philosopher to wield such notions in the service of philosophy of language as well as other parts of philosophy was David Lewis, who not only connected intensional Montague-style truth-condition theories with Gricean theories of languages as social practices (Lewis 1975) but who had also developed an influential early argument for the mind-brain identity theory that was later connected to a certain elaboration of the Ramsey-Carnap picture of theoretical terms (Lewis 1966, 1970, 1972). For many, the former work succeeded in showing how one might answer the broadly pragmatic question of what it is to use a language whose meaning is described in intensionalist terms (significant, since intensionalist semantics became an important enterprise in Australasian philosophy of language), while the latter work became an important plank in the platform of the Canberra Plan. (Lewis also developed a new logic and semantics of counterfactuals, and much more, all set within an account of possible worlds that he himself construed in famously realist terms).

Early Influences and the Davidsonian Revolution

To understand the history of philosophy of language in Australasia,Footnote 1 it is important to understand the foregoing more general history of the philosophy of language and its antecedents. It would be a mistake to think that there is a distinctive Australasian philosophy of language that arose and developed independently of this history. What is true, however, is that trends in Australian philosophy slowed the reception of philosophy of language as an important field in its own right. On the one hand, Andersonian philosophy (centred on Sydney) treated metaphysics as paramount; the preoccupation with language fostered by the rise of ordinary language philosophy in Britain was regarded with disdain. On the other hand, the Wittgensteinian approach to philosophy was dominant in Melbourne in the 1950s. Practised by such figures as Douglas Gasking, Camo Jackson and George Paul, such an approach favoured ordinary language philosophy but tended to be suspicious of the professionalisation of philosophy, and that included the kind of disciplinary specialisation found in the burgeoning field of philosophy of language. In New Zealand, by contrast, formal logic had been used by Arthur Prior to study the phenomenon of temporal and modal language, but this work was regarded as logic (or perhaps logic in the service of metaphysics) rather than as an attempt to develop a theory of meaning for (parts of) ordinary language.

What changed such attitudes and set philosophy of language on course to becoming an important and respected field of teaching and research in Australasia was the increasing internationalisation of philosophy. Melbourne may have been a centre of Wittgensteinianism, and Sydney an island of Andersonianism, but other universities were not so bound. In particular, the University of Adelaide appointed J. J. C. (‘Jack’) Smart in 1950, and Smart combined a respect for the sciences with a respect for the need to attend to and understand ordinary language. His work on the incoherence of talk of the passage of time, for example, is a model of conceptual clarification through linguistic analysis. Jonathan Bennett, reporting on Smart’s paper on time at the first New Zealand congress of Philosophy, commented that ‘Professor Ryle of Oxford (whose form of linguistic philosophy seems to be that to which Professor Smart owes most) has been accused of doing metaphysics while pretending to do linguistic analysis; Professor Smart goes one better: he does both and disguises neither’ (Bennett 1953, p. 197).

Smart was no closet philosopher of language, however—he was more interested in showing how language can mislead, thus highlighting the need for reform. The development of philosophy of language in Australia came later, and not just as a result of the impact of new books and articles or the views expounded in lectures by visiting philosophers (important though these were) but, increasingly, as a result of the weight added to such views by young lecturers appointed as a result of the expansion of the university system in the 1960s who had gained their postgraduate degrees abroad.Footnote 2

One event that had a considerable impact on Australian philosophy was the 1968 visit of Donald Davidson, who was invited by Jack Smart as Gavin David Young lecturer at the University of Adelaide and later toured the country introducing the truth-theoretic conception of a theory of meaning that he had begun to advance in the 1960s (e.g. Davidson 1967). Interest in Davidson’s work in Melbourne had been fuelled by a paper on adverbs read by Gary Malinas, visiting Melbourne soon after his arrival from the USA. Not surprisingly, perhaps, Davidson’s work proved especially influential in Melbourne—not only did it continue the (Wittgenstein-inspired) emphasis on ordinary language found in British philosophy but it combined this with the rigour that logic could provide, and all in the course of dealing non-sceptically, and in a principled way, with the thorny issue of the nature of meaning. (Sydney was less impressed.) One local Melbourne product influenced by Davidson was Barry Taylor, who subsequently went to Oxford for his DPhil before returning to Melbourne. His arrival was timely since it coincided with the start of the Davidsonian revolution at Oxford, and Taylor’s writings in the 1970s and 1980s reflect his immersion in the Davidsonian framework. In Taylor (1980), for example, he tackles the problem of constructing a truth-theory for a language containing simple and complex demonstratives, opting for a certain hybrid account that invokes relativity to utterer and time as well as points of reference, while Taylor (1985) builds on his DPhil work to provide a sophisticated Davidsonian truth-theory for a language containing adverbs, using a formal account of the notions of change and fact to develop a theory of events that is a reconstruction of Aristotle’s idea of events as changes of various types.

Another Australian deeply influenced by Davidson was Martin Davies, currently Wilde Professor of Mental Philosophy at Oxford, who went to Oxford in the mid-1970s after attending Monash University in Melbourne. Davies (1981) not only presents a detailed and influential interpretation of Davidson’s general position but also considered the technical problems facing a truth-theory able to accommodate quantification, anaphora and modality. Because of the influence of Quine, Davidson himself remained sceptical of modal notions and had not extended his project to languages with modal operators. Davies’ work thereby showed the influence of an Oxford perspective on Davidson’s project (in particular the influence of Gareth Evans and John McDowell). Around this time Davies also studied other aspects of truth-theory, in particular the vexed question of the relation of truth-theory to meaning and semantic competence (cf. Davies 1981).

One of the most intriguing of Davies’ contributions to Davidsonian truth-theory, and one which is discussed in the final chapter of Davies (1981), is based on joint work he did with the Monash logician Lloyd Humberstone on the logic and truth-theory of the modal operators ‘A(ctually)’ and ‘F(ixedly)’ (Davies and Humberstone 1980). Given a sentence s, the effect of applying ‘A’ and then ‘F’ is a sentence ‘FAs’ that says: whichever world had been actual, s would have been true at that world considered as actual. (An example of a sentence s such that FAs is true at an arbitrary world is ‘If anyone invented the zip, then the actual inventor of the zip invented the zip’.) Davies and Humberstone thought that the kind of necessity expressed by ‘FA’ corresponded to Gareth Evans’ ‘deep’ necessity, and discussed the idea that it might be used to explain the phenomenon of the Kripkean contingent a priori (they believed it could; that such statements were superficially contingent but deeply necessary) as well as the Kripkean necessary a posteriori (here they were more tentative—see especially the discussion at pp. 19ff of the suggestion that ‘Water is H2O’ is deeply contingent; for their more recent views, see Humberstone 2004 and Davies 2004). The issue returns with the work of Frank Jackson, to be discussed later.

The Davidsonian revolution provided a truth-theoretic perspective on the notion of meaning. It did not provide a full truth-conditional perspective on meaning, if by the latter is meant an approach that aims to identify, in compositional terms, something—some semantic value—that might be said to be the truth-condition of a sentence. For Davidson, truth-theories provide truth-conditions only in the sense of generating instances of the T-schema’s is true iff p’ as theorems of some metalanguage (making appropriate allowance for indexical expressions and the contribution of context), where the metalanguage is the language of the interpreter. Davidson thus makes it clear that his theory of meaning is a theory of interpretation, and, if so, it seems that there can be no requirement that we identify something extralinguistic that might count as the meaning of a sentence.

This has been one of the most contentious aspects of the Davidsonian program, and it has been rejected by many philosophers of language working on the notion of meaning and truth. In particular, it was rejected by intensionalists who followed Montague in his reformulation of Carnap’s appeal to intensional foundations for the semantics of language.

Intensional Semantics

In contrast to Davidson’s truth-theoretic account of meaning, Montague’s work in the late 1960s and 1970s provides a full truth-conditional perspective, with truth-conditions understood as generated from the semantic values of constituent expressions (see, e.g. Montague 1970). The most widely known development of this alternative perspective in Australasia is due to the New Zealand logician, Max Cresswell, who had studied under Arthur Prior after Prior’s departure for Manchester and then, at the insistence of David Lewis, attended Richard Montague’s lectures on English as a formal language while on leave at UCLA in early 1970. Underscoring how little the debates were seen as debates within a discipline of philosophy of language, Cresswell has this to say [personal communication]:

Until then we all knew that there were two kinds of philosophers, logicians and ordinary language philosophers. And we all knew that the ‘languages’ of formal logic and ordinary language were quite different. We disagreed about which kind of language was best for addressing philosophical problems. Then came the Chomsky revolution, and Montague showed us that the linguists were on the side of the logicians rather than on the side of the ordinary language philosophers.

Where Montague had considered a range of formal representations (universal grammar being the most abstract), Cresswell took the underlying structure of a natural language to be that of a lambda-categorical language. Sentences in such languages are represented as being composed out of functors and terms. At the level of semantics, these expressions are assigned appropriate semantic values (for instance, a function from entities to propositions, if the functor is a one-place predicate), with the semantic values of entire sentences determined by appropriate functional composition of the values of component expressions. The values of sentences at contexts of utterance are propositions, taken as corresponding to sets of possible worlds (the role played by possible worlds is less direct than their role in Montague’s framework; for a recent defence of this understanding of possible worlds). Cresswell (1973) sets out the basic framework, while later works discuss the proper treatment in this framework of such natural linguistic phenomena as anaphora, tense, prepositions and points of view, adverbial modification (Cresswell 1985a) and the propositional attitudes (Cresswell 1985b). The latter work defends the view that the meanings of propositional attitude verbs like ‘believes’ are sensitive not only to the proposition expressed by a whole sentence but to the meanings of its separate parts; the ensuing logic and semantics he calls ‘hyperintensional’ since such contexts do not respect intensional identity. (The theory admits other complications; for one thing, attitudes can be iterated, as in ‘Mortimer believes that Natasha believes that the earth is flat’. But to let the meaning/intension of (the first occurrence of) ‘believes’ operate on an agent plus something that includes the meaning of (the second occurrence of) ‘believes’ seems to involve a kind of vicious regress, and to handle these cases Cresswell allows the complement-forming word ‘that’ to be category-ambiguous; see Chap. 10, esp. pp. 89–92.) Cresswell defends his framework and analyses against rival accounts in Cresswell (1988), while Cresswell (1990) is an extended argument for the claim that possible worlds are more than theoretical posits of formal semantics—like times, they belong to the things that natural language assumes. (In this and other work Cresswell makes it clear that natural language is his starting point and destination and that he invokes an intensional formal semantics because natural language can and should be represented as an intensional language. A less committed approach was taken by Malcolm Rennie, a young philosopher and logician whose premature death in 1974 robbed Australasia of one of its rising stars and whose important monograph, Rennie (1974), sets out to show ‘that Church’s formulation of the simple theory of types provides a comprehensive and workable framework in which to deal with the logic of predicate modifiers and various aspects of intensional logics’ (p. 156), without, however, arguing that natural language should be represented as an intensional language.)

While Cresswell has been the most influential Australasian contributor to the Montague-inspired program of an intensional formal semantics for natural language, there have been others, among them John Bigelow, who was a colleague of Cresswell’s at Victoria University of Wellington. Bigelow (1975) argues that the best way to handle the contribution of context in formal semantics—for example, non-shiftable indices such as the utterer-index in sentences employing ‘I’—is by means of a certain ‘quotation device’ (work by David Kaplan on the notion of character has famously taken mainstream intensional semantics in a very different direction), while Bigelow (1978) is an influential paper on the semantics of propositional attitudes. This paper, published around the same time as Cresswell’s first attempt at a ‘structured meanings’ approach to hyperintensionality, lets the embedded clause in a propositional attitude ascription refer to what Bigelow calls its ‘semantic structure’, a fine-grained abstract representation containing ‘markers’ and symbols that, he argues, is nonetheless sensitive to the phenomenon of quantifying into hyperintensional contexts while allowing for the representation of the iteration of attitudes.

Bigelow’s research following his departure for Australia (first La Trobe, then Monash) took him in the direction of metaphysics rather than the philosophy of language. The problem of the logic and semantics of propositional attitude ascriptions, however, has been the subject of a number of papers by some of his current Monash colleagues, including Karen Green while still at Melbourne University following her Sydney PhD under Michael Devitt. In Green (1985), she raised the sceptical worry that the very possibility of a logic of propositional attitude ascriptions was under threat if we adopted either the Quinean or Davidsonian conception of logical form. Graham Oppy, who had been a student of Green’s at Melbourne before completing his PhD on the semantics of propositional attitudes at Princeton University, returned to the topic in a number of papers in the early 1990s. While suspicious of the argument for semantic innocence made by neo-Russellians (Oppy 1992a), Oppy claimed that neo-Fregean and neo-Russellian approaches to propositional attitude ascriptions were structurally alike at least to the extent that neither could avoid dealing with the role, content and compositional structure of the modes of presentation that featured in the (neo-Fregean) semantics or (neo-Russellian) pragmatics of such ascriptions. Using a neo-Fregean theory as his working model, Oppy then provided an account of the way context furnished such modes of presentation (Oppy 1992b). It should be noted that this working model is a kind of ‘hidden-indexical’ theory. Like neo-Russellianism, such an approach is very different from the intensionalist approach favoured by Cresswell and Bigelow.

Another New Zealand philosopher who advocated an intensionalist semantic framework was Pavel Tichý, a Czech philosopher who spent the second half of his life in New Zealand as a political refugee, where he taught at the University of Otago. Early in his career he invented a version of intensional logic, simultaneously with Richard Montague’s invention of such a logic, but published slightly after Montague’s papers of 1970. Tichý’s theory, which he called ‘Transparent Intensional Logic’, is a logical semantics for the analysis of sizeable fragments of both natural and artificial languages. Like Montague’s approach, it belongs squarely to the research paradigm of possible-worlds semantics but differs from it by ignoring pragmatic and other contextual features of language. And unlike Montague’s approach or Cresswell’s development of this approach, the lambda-expressions of Tichý’s language do not stand for functions and the results of functional application but for the very procedures of forming functions and applying functions, respectively. Such procedures or constructions constitute a third semantic tier in Tichý’s system (after the two tiers of extensions and intensions). Indeed, construction constitutes the single most important notion of Transparent Intensional Logic, playing a crucial role in his account of hyperintensionality and much else. (Tichý 1988 provides a mature statement of the foundations together with applications, while Tichý 2005 is a posthumous collection of papers that traces the development of his ideas.)

Probably the philosophically most radical version of intensionalism advanced by an Australasian philosopher (or by any other philosopher, for that matter) is that of the New Zealand-born logician Richard Routley, most of whose professional life was spent as a Research Fellow at the ANU. But before turning to Routley’s work, I want to describe the distinctive contribution to intensionalism of another Australasian philosopher, Charles Hamblin, who taught philosophy at the University of New South Wales and during his career was responsible for a number of important innovations in computer science (such as reverse Polish notation and the notion of a stack). In his 1957 London School of Economics PhD, Hamblin had presented a critique of Shannon’s theory of information from a semantic perspective and developed a possible-worlds semantics for question-response exchanges. He returned to the topic of questions in Hamblin (1973) but this time adopted Montague’s framework, showing how one might give a complete syntactic-semantic set of rules for the kind of fragment of English discussed by Montague, but supplemented with questions and using the fact that the basic interrogative words fit more or less neatly into Montague’s categories. (According to Hamblin’s influential account, the intension of an interrogative is a function from possible worlds to sets of answers, where answers are propositions). This preoccupation with the logic and semantics of non-declarative sentences continued with his work on imperatives (Hamblin 1987, published posthumously). Hamblin assigned to each imperative a set of possible worlds in which it was satisfied, with worlds construed as chains of states connected by deeds and happenings, able to accommodate both physical and agent causation; this rich underlying model allowed Hamblin to develop a contrasting notion of ‘wholehearted’ satisfaction that characterised an agent’s involvement and responsibility in fulfilling an imperative. Like much else of Hamblin’s work, Hamblin’s theory of imperatives has had consequences for computer science, in this case for developing protocols governing the delegation of tasks between software agents (McBurney 2003).

The contribution of Richard Routley, later ‘Sylvan’, to philosophy in Australasia is hard to overestimate. Routley/Sylvan was one of Australasia’s most prolific and systematic philosophers, and his work on relevance and paraconsistent logics has influenced numerous logicians, inside and outside of Australasia. His work inevitably impacted on the philosophy of language, even though his logical and metaphysical ideas really hold centre stage (see Hyde 2001). Following his collaboration with Len Goddard in the 1960s and early 1970s on the logic of meaningfulness or significance (Goddard and Routley 1973), he began to explore some themes thrown up by that early work, in particular the phenomena of intensionality, semantic paradox and failure of reference. On the basis of joint work with Val Routley, he came to think that much work in philosophy was the subject of a pervasive error, the reference theory of meaning, which holds that all truth-valued discourse is referential (i.e. is true or false depending on the reference of constituent expressions, with reference construed as an existence-entailing relation). In its place, Routley urged the adoption of what he termed a noneist framework, a version of Meinongianism that holds that every term whatsoever is about, or designates, something, in many cases something that has a range of properties but is incomplete (for some properties, it lacks both the property and the complement of the property) and doesn’t exist (only things occupying actual regions of space and time exist). The description ‘the non-square square’ (i.e. ‘the x such that x is square and ∼[x is square]’), for example, designates something that we can think and talk about, and that is both square and not square—but it doesn’t exist and is incomplete. Not only that, it is also an impossible object: it infringes the law of non-contradiction. The truth-theoretic semantics of ordinary language is thus very different on Routley’s view from the standard view. Some sentences are true in impossible worlds, not just possible worlds. Not surprisingly, this has an impact on the theory of meaning. Routley’s work belongs to the general intensionalist camp—he takes his work on universal semantics to show that every logic has a designative theory of meaning (Routley 1980, p. 335), but in particular the notion of meaning for a rich lambda-categorical language is taken to be designative in a highly intensional sense. For Routley, expressions in a language of this kind are interpreted by means of functions on the class of situations and contexts, but such a meaning-specifying interpretation for the language is furnished with a domain of individuals and a class of situations and contexts that are far richer than those admitted by other intensionalist frameworks. They comprise individuals and worlds that can be complete or incomplete, possible or impossible. In addition, Routley rejects the reduction of functions to set-theoretic entities, and he argues that this allows him to avoid some common problems facing intensionalist theories of meaning, including the problem of propositional attitude contexts (for Routley, the semantics of the verb ‘believes’ relates agents to propositions—no need to invoke hyperintensionality; Routley 1980, pp. 343–345).

Routley tended to present his ideas in a rather combative fashion. A more recent, and rather more reader-friendly, attempt to apply a version of Routley’s liberal semantic machinery to the analysis of language occurs in Priest (2005), the first part of which is concerned with the semantics of intentionality (including both intentional predicates and operators) and which accepts a noneist framework of possible/impossible/open worlds and existent/non-existent objects. This account is then applied to a range of topics, including the topic of fictional and extra-fictional discourse (the latter including claims like ‘Holmes is admired by readers of the Doyle stories’ and ‘Holmes doesn’t really exist’). Contrary to Routley’s version of noneism, Priest holds that Holmes is not a detective at the actual world but only at worlds that are the way Doyle’s stories represent them as being: in particular, at worlds, including impossible worlds, where he exists. At the actual world, however, Holmes has such properties as being non-existent and being admired by readers of the Doyle stories.

By contrast, a more traditional intensionalist like Cresswell rejects any appeal to impossible objects and impossible worlds (Cresswell 1994, pp. 61–62) and takes fictional and extra-fictional discourse to be about merely possible objects. But there are other Australasian philosophers who have taken rather different approaches to Meinongianism or intensionalism when writing on the topic of the semantics of fiction. The British-born philosopher Greg Currie, for example, who was a colleague of Pavel Tichý at Otago University before moving back to the UK, has advanced an influential Gricean theory on which the author of a work of fiction intends her audience to make believe the content of the work and to recognise that she intends them to do this (Currie 1990). On Currie’s account, fictional names should be understood as bound variables when they occur within a work, as abbreviated descriptions (generated Ramsey-style from the work in question) in statements about the work (in which case they occur inside the scope of an ‘in the fiction’ operator), and as names of roles in extra-fictional contexts, where roles are partial functions from worlds to the individuals in those worlds who satisfy the abbreviated descriptions (Currie takes these to be versions of Pavel Tichý’s ‘offices’). Other authors have taken a different approach. Stuart Brock, for example, is a fictionalist about fictional characters. He eschews realism about fictional characters (even Currie’s role-realism), and he thinks that we can defuse realism about such characters by interpreting extra-fictional statements (the most difficult category of fiction-involving statements) as shorthand for statements that certain things are so according to the realist’s theory of fictional characters (Brock 2002). As Brock acknowledges, such a fictionalism still does not make room for negative existence statements like ‘Holmes doesn’t exist’, and Brock has argued that there are serious difficulties confronting the analysis of such statements even for those who think that fictional and other empty names are to be understood descriptively (Brock 2004). Kroon (2004) is an attempt to solve these difficulties by combining a pretence account of fictional language with a descriptivist account of the reference of names in general. (Note that these critics of intensionalist attempts to deal with fictional language tend to think that fictional language is a rather special phenomenon; they do not see their challenges as challenges to intensionalism as such).

Naturalising Semantics

The intensional approach to formal semantics of natural language is a way of doing theory of meaning. Proponents of such an approach were usually not particularly concerned with the epistemological question of how knowledge of meaning was possible (although this question played a prominent role in Tichý’s work); still less did they worry whether its theoretical articulation of the notion of meaning could be meshed with a broadly naturalistic philosophy. The attempt to understand semantic notions in broadly naturalistic terms became an important part of the motivation of another well-known Australasian philosopher of language, Michael Devitt. As a student in the mid-1960s at the University of Sydney, Devitt’s interest in the philosophy of language was kindled by the work of C. B. (‘Charlie’) Martin, whom Devitt recalls as urging a thesis for names and demonstratives that was, in effect, like the rigidity theses to be made famous by Saul Kripke and David Kaplan, and a view of definite descriptions that was similar to the view that Donnellan was about to publish. In his Harvard PhD and later works such as Devitt (1974) and (1981a), Devitt developed a version of Kripke’s causal account of reference for proper names and natural kind terms (Kripke 1972) but one that was far more specific about the kinds of grounding events and causal connections that determine this kind of reference (or, as Devitt prefers, designation). He also argued for a semantic account of the role played by Donnellan’s ‘referential’ descriptions (Devitt 1981b) and gave a causal theory of the way their reference was determined. What was distinctive about these causal theories was that, unlike Kripke’s much more tentative account, they were placed squarely within a naturalistic philosophy. Devitt later argued (in Devitt 1989) that his causal framework could be used to give an account of the meaning (not just the mode of reference determination) of names, sharply disagreeing with the most popular post-Kripke way of understanding the meaning of names, namely, the Millian ‘direct reference’ approach of such philosophers as Nathan Salmon and and Scott Soames: (Unsurprisingly, not all Australasian philosophers of language have been convinced by Devitt’s causal approach to reference and meaning, even though they might have shared his naturalistic outlook and his disaffection for Millianism. Some have even argued for a return to a descriptivist account of names, albeit a broadly causal version; cf. Kroon 1987 and Jackson 1998b.)

Note that Devitt’s account is a theory of the reference and meaning of ordinary names, and the meaningfulness of fictional and other empty names may look problematic from this perspective. But Devitt thinks that these problems are tractable (Devitt 1989), unlike the problems faced by Millians. He thinks that even empty names have underlying causal networks, although these fail to be grounded in an actual individual (Devitt 1981a, Chap. 6). In the case of the metafictional statement ‘In the fiction, Holmes is a detective’, for example, the causal network underlying ‘Holmes’ points to the parts of the Holmes stories that help to make it true in the fiction that Holmes is a detective. And the negative existential statement ‘Holmes doesn’t exist’ is true to the extent that the causal network underlying tokens of the name ‘Holmes’ is not grounded in any individual.

Some of Devitt’s views show the influence of work by his erstwhile student Kim Sterelny (see, e.g., Sterelny 1983), with whom he also published a popular and influential textbook on the philosophy of language (Devitt and Sterelny 1987, 2nd edition 1999). This book clearly shows their disenchantment with much of contemporary philosophy of language, and later works by Devitt elaborated on his reasons for this disenchantment. In Devitt (1994), for example, he proposed answers to three questions: ‘What are the semantic tasks?’ ‘Why are they worthwhile?’ and ‘How should we accomplish them?’ urging a methodology to escape from the kind of intuition mongering he took to be rife in the philosophy of language. Devitt (1996) uses this methodology to criticise, inter alia, semantic holism and two-factor theories and to support a truth-referential semantics. More recently, Devitt has mounted an extensive critique of the Chomskian ‘psychological conception’ of linguistics, arguing instead for a ‘linguistic conception’ (see especially Devitt 2006).

Devitt left Australia for the USA in 1987, and work after this period should not be strictly construed as Australasian philosophy of language. There can be little doubt, however, that it is work done in the same naturalistic spirit as his earlier work, a naturalistic spirit encouraged by his early training in the Sydney philosophy department and fostered by teachers (Quine) and friends (in particular Hartry Field) at Harvard. There is also another sense in which we can see Devitt’s work as particularly fitting, giving this early training. As I mentioned earlier, the metaphysical orientation of Sydney philosophy in the 1950s and 1960s meant that it was bound to be suspicious of the linguistic turn in philosophy and that might further suggest that it would be less receptive to the burgeoning field of philosophy of language than, say, Melbourne philosophy. That has not been so. Instead, the kind of philosophy of language practised by Devitt and his students simply rejects the imperial ambitions of the linguistic turn in philosophy. Unlike Michael Dummett in particular, Devitt denies philosophy of language any claim to pre-eminence in philosophy. Nowhere is this more clear than in Devitt’s influential Realism and Truth (1984), where Devitt argues that realism is a metaphysical doctrine, not a semantic one, and that much damage is done by the thought that the task of philosophy is to ‘analyse our concept of X’ rather than to ‘explain the nature of X’. For Devitt, we should ‘Put Metaphysics First’. John Anderson would have been pleased.

Anti-realist Tendencies

Each of the three theoretical perspectives I have discussed so far—Davidsonianism, formal intensional semantics and Devitt-style causal semantics—has been the subject of criticism by other Australasian philosophers. Devitt in particular has been vociferous in his criticism of other approaches to the topics of reference and meaning. One significant critic of the presuppositions rather than the detail of these approaches has been Huw Price, who rejects truth-oriented accounts of meaning because they depend on what he regards as the unsustainable assumption that there is a viable distinction between assertoric and other discourse (Price 1988). Unlike expressivists, who have tended to prefer appropriateness conditions rather than truth-conditions for what they see as non-assertoric discourse, Price advocates a bifurcation between two notions of representation (an ‘external’, world-tracking notion and an ‘internal’, inferentialist notion) and a generous notion of assertion on which assertion functions as a coordination mechanism but where what gets coordinated depends on the practice or functional task of the utterances in question (and so on their style of representation). Price takes himself to remain a naturalist on this picture (Price 2008; see also Price 2004, which rejects the representationalist idiom). But the picture leaves the notion of truth-conditions as in a sense functionally ambiguous and certainly not well equipped to play a central role in a theory of meaning. Price, in fact, thinks that the best kind of theory of meaning does not yield a non-semantic reduction of statements of the form ‘x means F’, but—as is done by deflationists in the case of truth—explains the function of such a statement in terms which don’t require that it refers to substantial properties (Price 1997).

Price embraces a form of naturalism (perhaps nothing less could be expected of someone who was a Challis Professor at the University of Sydney) but resolutely rejects the kind of naturalistic referential semantics that someone like Devitt espouses. He thinks that such views result in an unsustainable kind of object naturalism (Price 2004). Still, he does not reject truth-conditions in favour of, say, Dummettian verification or justification conditions in a semantics of language. So he would not count as an anti-realist in the sense made famous by Michael Dummett. (Devitt, of course, thinks that the term ‘anti-realism’ is a misnomer when used in this way).

For the most part, Australasian philosophers have not been kind to such forms of anti-realism. Notable among works discussing Dummett’s philosophy is Green (2001a), which is a comprehensive account of how Dummett’s views about objectivity, normativity, systematicity, publicity and the dependence of thought on language developed from views found in Frege and Wittgenstein, and which sympathetically explains the anti-realist worry that ‘truth’ cannot be expected to be bivalent for a language in which meaning is use. But Green is at the same time careful to distinguish Dummett’s objectivist anti-realism about this or that discourse from subjective idealism, arguing that such objectivist forms of anti-realism are perfectly compatible with realism about common-sense physical objects or even realism about the entities of science (Green 2001a, p. 203; Green reinforces this point in Green 2005, which explains and criticises Dummett’s later view that a denial of bivalence always leads to a form of anti-realism). Green therefore shares Devitt’s scepticism that such semantic doctrines have much to do with the metaphysical issue of realism while being sympathetic to the underlying themes of Dummett’s program in the philosophy of language. (Green even supports the central Dummettian theme that language is prior in an account of thought, a view she defends in her critique of Davidson’s seemingly contextualist ‘A Nice Derangement of Epitaphs’ (Green 2001b), although in a form that insists on the priority of the existence of conventions of language rather than of the attribution to speakers of knowledge of such conventions).

Two other Australian philosophers who have discussed anti-realism in its various forms are Drew Khlentzos (2004) and Barry Taylor (2006). Khlentzos endorses naturalistic realism, with realism characterised in familiar metaphysical terms, but he accepts that there are apparently compelling arguments against such a realism which deserve a response. This includes Dummett’s argument against verification-transcendent truth-conditions, although Khlentzos is more impressed by Putnam’s model-theoretic argument. For Khlentzos, the most serious challenge to realism is one we derive from Putnam’s argument: the problem of explaining how the mind can form an adequate representation of a mind-independent reality.

While Khlentzos sees this as a challenge rather than a reason to reject realism, Barry Taylor draws the opposite conclusion (2006). Like Dummett, Taylor believes that the best way to formulate realism is as a claim about truth; specifically, that truth is objective, that is, public, bivalent and epistemically independent. After attempting to justify this characterisation of realism in the face of Devitt’s arguments for an explicitly metaphysical characterisation, Taylor argues at length that no known and defensible notion of truth preserves these realist theses, whether it be a notion of correspondence truth (here Putnam’s model-theoretic argument plays an important role) or a non-correspondence notion such as Wright’s superassertability, Tarskian truth or the notion of truth at play in McDowell’s ‘quietist realism’. But he resists opting directly for a Dummettian anti-realism, with its rejection of bivalence, and insists that less radical anti-realist options involving the rejection of the epistemic independence dimension of truth remain on the table. (Presumably Taylor came to think that the Davidsonian program he championed earlier in his career could only be defended if some such anti-realist revision of the notion of truth is adopted).

While anti-realism is one option for those who take Dummett’s version of a use theory of meaning to heart, another prominent view inspired by Wittgenstein’s work is sceptical rather than metaphysical. This is the scepticism about rule-following that Kripke’s 1982 work Wittgenstein on Rules and Private Language finds in Wittgenstein’s work. The so-called paradox of rule-following—that no course of action, such as using words with a certain meaning, can be determined by a rule because any course of action can be made out to accord with the rule—has generated a huge literature, one to which the (Irish-born) Australian philosopher Philip Pettit has made a seminal contribution. According to Pettit’s ‘straight’ solution to the paradox, cast far more widely than as a solution to a problem of linguistic meaning alone, a rule can be exemplified by cases if these give rise to an inclination in an agent to go on in a certain way, where the rule is the one that the inclination corresponds to in the actual world provided the inclination operates under favourable conditions (Pettit 1990). Importantly, Pettit thinks that if such rule-following is to be public, then the rule-followers must interact with one another as well as with their earlier and later selves, and that this has implications for our understanding of mental and social life in general. In particular, he argues in The Common Mind that rule-following marks off thinking subjects from other intentional systems, before going on to situate this picture of thinking subjects in a larger framework for social and political theory (Pettit 1993).

Language and Vagueness

Green, Khlentzos and Taylor engage directly with Dummett’s challenge to realism. Other Australian philosophers have engaged with certain Dummettian challenges to classical logic that don’t have such a direct bearing on the realism debate. The role of vagueness in the sorites paradox seems an obvious example, as even Dummett thought (Dummett 1975). Linda Burns (1991) offers an influential criticism of the case that Dummett and his follower Crispin Wright develop against standard models of language based on the alleged incapacity of such models to handle vagueness and the sorites, in particular their charge that observational predicates induce incoherence (see, e.g. Dummett 1975). Her solution is an example of a contextualist solution to the sorites: in context, a sorites-generating predicate is always interpreted in such a way as not to distinguish between a pair of items in a sorites series, and this produces contextual shifts along a sorites series since things do not remain the same as one moves along the series (technically, Burns advocates the view that the relevant predicates are governed by loose tolerance principles rather than Wright’s strict tolerance principles). Unlike the first contextualist, Hans Kamp, Burns uses this idea to provide a purely pragmatic analysis of the sorites paradox: classical logic and semantics are left intact, and the crucial induction premise is declared false.

Whether such techniques can be used to solve every instance of the sorites is in contention. In any case, there are well-known attempts by Australian philosophers to deal with vagueness by rejecting classical logic and semantics. Dominic Hyde, in particular, has written extensively on the topic (see especially Hyde 2008). Hyde rejects classical semantics, but he also rejects the widely accepted supervaluationist approach, which he sees as no more—and no less—compelling than the dual subvaluationist approach (the latter is a paraconsistentist approach that posits truth-value gluts where the former posits truth-value gaps). While Hyde professes a degree of uncertainty about the relative merits of gaps and gluts, he prefers the associated logic of vagueness to be a truth-functional logic: If, for example, A and ∼A are both gappy, then the same should be true of A ∨ ∼A, contrary to supervaluationism. Hyde combines this view with a tolerance for ontological vagueness and, indeed, thinks that the best defence of supervaluationism is based on the thought that vagueness is a merely semantic, representational phenomenon.

Burns and Hyde are just two of a number of Australian philosophers who have worked on vagueness. John A. Burgess and Lloyd Humberstone, for example, have argued for a variation on supervaluationist logic that abandons the latter’s logical conservatism by rejecting the law of excluded middle in the face of seeming counterexamples presented by vagueness (Burgess and Humberstone 1987). More recently, Burgess has presented a new objection, based on principles about the metaphysics of content, against the epistemicist view that there are sharp, but unknowable, boundaries to the extension of vague expressions (Burgess 2001). And not surprisingly, perhaps, given the Australian logical tradition, there is a growing body of work arguing for a paraconsistent approach to vagueness, beginning with Priest and Routley (1989). Beall and Colyvan (2001), for example, argue that a subvaluationist solution to the sorites paradox is superior to a supervaluationist approach, at least to the extent that it achieves uniformity with paraconsistent solutions to semantic paradoxes, while Hyde and Colyvan (2008) provide further reasons for pursuing the paraconsistent option.

The Return of Conceptual Analysis

The foregoing has highlighted the role played by formal semantic approaches as well as naturalistic approaches in the work of Australasian philosophers engaging in debates in the philosophy of language. The final major strand to be discussed combines elements of both an intensionalist approach to the theory of meaning and a naturalistic approach to the nature and function of language. It also returns us to the philosophical outlook that helped to spawn the philosophy of language in the 1950s and 1960s, an outlook—the linguistic turn—in which philosophical problems were regarded as (by and large) problems that could be solved through attending to the correct use of language, whether in its ideal form (ideal language philosophy) or in its natural form (ordinary language philosophy). For in the case of ordinary language philosophy, this outlook involved the methodology of conceptual analysis as broadly understood by Ryle, Grice, Strawson and Austin; and the preoccupation with conceptual analysis, seen through the lens of a naturalist, representationalist account of language and interpreted from within a two-dimensional modal semantic framework, has been an important part of the recent work of Frank Jackson.

Before discussing Jackson’s contribution to the philosophy of language, I will first turn to the topic of indicative conditionals, another area to which Jackson has made a signal contribution (Jackson 1987). Jackson’s contribution exploits a distinction for which Australasian philosophy is less well known: the Gricean distinction between what is said and what is implicated. (Indeed, as applied to such topics as relevant implication and entailment Australasian philosophy—in particular Australasian logic—is better known for a reluctance to understand these in pragmatic terms.) Jackson presents a theory that explains certain conflicting intuitions about indicative conditionals, among them the principle that one can pass from a statement involving the classic truth-functional connectives ‘and’, ‘or’ and ‘not’ to conditional statements (e.g. that ‘A or B’ implies ‘If not-A then B’) and the principle of the paradoxes of implication (that ‘Not-A; therefore, if A then B’ and ‘B; therefore, if A then B’ are both invalid). Jackson’s solution is to reject the latter principle, proposing instead that an indicative conditional and the corresponding material conditional have the same truth-conditions but not the same assertibility conditions. (Jackson thinks that subjunctive conditionals are very different and are to be understood in the manner made famous by Stalnaker and Lewis.) According to Jackson, the indicative conditional ‘If A then B’ is assertible iff (if and only if) the conditional probability Prob(B/A) is high, a principle due to Ernest Adams which Jackson supports by linking assertibility with a notion of robustness; that is, if the indicative conditional is assertible, it should remain assertible upon learning that A is true. Jackson thinks of such robustness as a matter of conventional rather than conversational implicature.

It should be said that while Jackson’s work on conditionals is particularly widely known, it is not the only influential work done by Australasian philosophers on the topic of conditionals. The Macquarie philosopher, Vic Dudman, initiated a wholly different research program in the 1980s on the impact of tense on the classification of ‘if’ statements. His work, which influenced Jonathan Bennett and a number of other prominent researchers, placed grammar before semantics and drew the line between various types of ‘if’ statements at a very different point from the dominant tradition followed by Jackson. For Dudman, the theoretically important dividing line among the statements (1) ‘If Oswald hadn’t killed Kennedy, someone else would/could have’, (2) ‘If Oswald doesn’t kill Kennedy, someone else will/might’, and (3) ‘If Oswald didn’t kill Kennedy, someone else did’ falls below (2) rather than—as the tradition has it—below (1), on the grounds, roughly, that (1) and (2) but not (3) display a forward time-shift. When they are interpreted this way, Dudman calls the former statements ‘conditionals’ and the latter ‘hypotheticals’ (Dudman 1989, 1994a). Using detailed grammatical analysis on such cases, Dudman argues that the subjunctive-indicative divide is vacuous, that the antecedent-consequent distinction is inapplicable to a vast range of ‘if’ statements (or, as he prefers, to the messages encoded in a vast number of English ‘if’ sentences), and that what he calls ‘conditionals’ are simple sentences that don’t have truth-values (Dudman 1989, 1994a, and 1994b). Another prominent critic of the standard distinction among indicative and subjunctive conditionals is La Trobe’s Brian Ellis, although his arguments rely far less on grammatical considerations and he certainly does not deny the antecedent-consequent distinction or accept Dudman’s conditional-hypothetical distinction (Ellis 1978, 1984: fn 9). For Ellis, both kinds of conditionals have truth-conditions but are non-truth-functional, with mood being a device for signalling what beliefs are being retained as a basis for reasoning from the antecedent supposition—a very different view from Jackson’s.

Jackson’s more recent work has been on the role and nature of conceptual analysis and is driven by his interest in metaphysics, the philosophy of mind, and metaethics. Of particular significance is Jackson (1998a), based on Jackson’s John Locke Lectures at Oxford University in 1995.Footnote 3 The focus of this work is ‘serious’ metaphysics and the ‘location problem’. ‘Serious’ metaphysics ‘is the attempt to give a comprehensive account of some subject-matter—the mind, the semantic, or, most ambitiously, everything—in terms of a limited number of more or less basic notions’ (p. 4), and the ‘location problem’ for an everyday property (whether it be artifactual, mental, semantic, social, economic, etc.) is to say how and why the property does or does not ‘get…a place in the scientific account of our world’ (p. 3). Jackson maintains what he calls ‘the entry by entailment thesis’: that ‘the one and only way of having a place in an account told in some set of preferred terms is by being entailed by that account’ (p. 5), where by ‘entailment’ he means ‘simply the necessary truth-preserving notion’ (p. 25). According to Jackson’s novel spin on this linguistic version of a supervenience claim, such a commitment to entailment theses requires serious metaphysicians to do conceptual analysis, since ‘conceptual analysis is in the very business of addressing when and whether a story told in one vocabulary is made true by one told in some allegedly more fundamental vocabulary’ (p. 28).

For Jackson, much as for early champions of conceptual analysis like Austin and Grice, to do conceptual analysis is to reflect on which possible cases fall under which descriptions; one’s intuitive judgements about hypothetical cases manifest one’s ‘theory’ of the relevant subject matter, and to the extent that one’s intuitions coincide with the folk, they reveal the folk theory. For example, conceptual analysis reveals that, according to our folk theory of water, water is ‘whatever actually is both watery [i.e. is found in rivers and lakes, falls from the skies, etc.] and is what we are, or certain of our linguistic forebears were, acquainted with’ (p. 39). Jackson thinks that such a conclusion amounts to a kind of description theory of the terms in question, although one that properly accommodates the Kripke-Putnam intuitions underlying causal theories of reference. (Jackson 1998b, a paper that clearly shows Jackson to be a descriptivist about proper names as well, provides more specific responses to Kripke’s arguments against descriptivism).

Conceptual analysis of this kind is what also underlies Jackson’s account of two kinds of intension that a sentence or term may have. What is ordinarily taken as the intension of a kind term like ‘water’ is a function from possible worlds to a substance that exists at those worlds; the intension of ‘water’, for example, takes a world to the substance H2O at that world. Jackson calls this the term’s C-intension because it is what we get when we consider ‘what the term applies to under various counterfactual hypotheses’ (p. 48) about this, the actual, world; and under any such counterfactual hypothesis, the actual watery stuff of our acquaintance remains H2O. But the term also has an A-intension, which is what we get when we consider ‘for each world w, what the term applies to in w, given or under the supposition that w is the actual world, our world’ (Jackson 1998b), that is, when we consider for each world w what instances there are of ‘the actual watery stuff of our acquaintance’ when those words are understood under the supposition that w is the actual world. (Using the terminology of Davies and Humberstone 1980, we might say that this shows the truth of ‘F(ixedly)A(ctual [water is the watery stuff of our acquaintance]’, and hence that ‘water is the watery stuff of our acquaintance’ is deeply necessary. Note, however, that Davies in particular has been critical of Jackson’s appeal to such a two-dimensional modal framework; see especially Davies 2004.)

Having argued that conceptual analysis provides us with a priori knowledge of A-intensions, Jackson is now able to argue that the entry by entailment thesis is best construed as an a priori deducibility thesis, and that physicalism, for example, is committed to the existence of conceptual entailments from the physical to the psychological. The case of ‘water’-talk again provides the model: ‘[W]e will be able to move a priori from…sentences about the distribution of H2O combined with the right context-giving statements, to the distribution of water.’ (p. 82). Thus, the sentence (i) ‘H2O covers most of the Earth’, together with (ii) ‘H2O is the watery stuff of our acquaintance’, conceptually entails (iii) ‘Water covers most of the Earth’, in view of the a priori status of ‘Water is the watery stuff of our acquaintance’ (p. 82). Jackson’s ambitious plan in Jackson (1998a) is to use the same methodology to solve the location problem for both colour and ethics—that is, to place both colour-talk and ethical-talk in the physical or descriptive picture of what the world is like.

Jackson’s work has been extremely influential, but also highly contentious, in large part because it goes against two widely accepted views in Anglo-American philosophy: the rejection of the analytic/synthetic distinction, made popular by Quine’s work (Jackson responds in 1998a, pp. 44–46), and the wholesale rejection of descriptivism about names and natural kind terms made popular by the work of Kripke and Putnam. Although Jackson is a naturalist, his attitude to conceptual analysis makes him a very different kind of naturalist from Michael Devitt, say, who accepts the anti-descriptivist causal theory of reference and is sympathetic to Quine’s attack on the analytic/synthetic distinction. He is also a very different naturalist from someone like Huw Price, who combines his naturalism with a bifurcated notion of representation and truth-conditions and a consequent denial that truth-conditions are central in a theory of meaning. Jackson, on the other hand, thinks that meaning settles how sentences represent things as being, in a single, standard sense of representation, and that how a sentence represents things as being determines truth-conditions. (This includes ethical sentences; in joint work with Philip Pettit (Jackson and Pettit 1998), he uses considerations from Lockean philosophy of language to argue that an alternative expressivist metaethics has independent problems.) Truth-conditions thus remain pivotal, even if we need two-dimensionalism to tell us what kind of truth-conditions are in play.

Jackson is not the only Australian philosopher to have made a contribution to the revival of conceptual analysis. Significant work has been done by David Braddon-Mitchell on the nature and role of folk-theories (see, e.g., Braddon-Mitchell 2004a and the argument in Braddon-Mitchell and Jackson 2006 for analytic functionalism), on the virtues of descriptivism as understood in two-dimensional terms (Braddon-Mitchell 2004b), and on the virtues of conditional analyses (Braddon-Mitchell 2003 claims that in the case of the term ‘quale’ such an analysis allows one to accept zombie intuitions while resisting a dualist conclusion). There is even a recent anthology, edited by Braddon-Mitchell and Robert Nola, on this style of conceptual analysis and its embedding in naturalism (Braddon-Mitchell and Nola 2009), a program that is often known as the Canberra Plan because many of its proponents have been associated with the Australian National University in Canberra and which counts among its most prominent and influential proponents the late David Lewis, Frank Jackson, Michael Smith and Philip Pettit. (The name ‘the Canberra Plan’ makes it sound as if we here have a home-grown philosophical program—something to rival Andersonianism in its Australianness. But of course this is far from being the case. Philosophy like so much else has now become truly globalised; and while some of the main players were Australians, David Lewis counts as one of the most important proponents of the Plan, among other things for the manner in which he combines physicalism with the idea that theoretical terms (including folk-psychological terms) can be defined via the Ramsification of theories. In general, Lewis’ influence on Australian philosophy—on philosophy in general, of course, but Australian philosophy in particular—has been immense.)Footnote 4

As the Canberra Plan is normally portrayed, it embeds conceptual analysis in a materialist form of naturalism. Jackson aside, however, the best-known Australian proponent of conceptual analysis, especially as seen through the lens of two-dimensionalism, is someone who is best known for his commitment to a non-materialist form of naturalism. This is the philosopher of mind David Chalmers, who used the machinery of two-dimensionalism in his influential The Conscious Mind (1996) to argue for mind-body property dualism. Chalmers’ work ranges widely over the philosophy of mind and metaphysics, but my focus here will be his contributions to the philosophy of language and in particular the distinctive way in which he understands two-dimensionalism. Like Jackson, Chalmers accepts two kinds of intensions: primary intensions, corresponding to Jackson’s A-intensions, and secondary intensions, corresponding to C-intensions. A primary intension is a function from scenarios (initially to be conceived of as centred worlds) to extensions. A secondary intension is a function from possible worlds to extensions. According to Chalmers, a sentence token S is metaphysically necessary iff the secondary intension of S is true at all worlds, while a sentence token S is a priori, or epistemically necessary, iff the primary intension of S is true at all scenarios. To justify the latter claim in particular, Chalmers argues for what he calls epistemic two-dimensionalism, according to which the scenarios that are in the domain of a primary intension represent highly specific epistemic possibilities and not, for example, contexts of utterance. The value of an expression’s primary intension at a scenario reflects a speaker’s rational judgments involving the expression, under the hypothesis that the epistemic possibility in question actually obtains. (For details, see Chalmers 2004, 2005.) For example, it is epistemically possible that our world is the XYZ-world (that is, that the liquid in the oceans and lakes is XYZ rather than H2O), and if it is, we should rationally endorse the claim ‘water is XYZ’, and we should rationally reject the claim ‘water is H2O’. So the primary intension of ‘water is H2O’ is false at the XYZ-world, and the primary intension of ‘water is XYZ’ is true there. Chalmers insists that this is not to say that names like ‘water’ are equivalent to (rigidified) descriptions, as Jackson suggests; rather, names have a normative inferential role that makes certain claims rational or irrational, given enough information. (Chalmers and Jackson 2001 set various doctrinaire disagreements aside to say why Chalmers and Jackson both think that conceptual analysis and a priori entailments are required if reductive explanation of the phenomenal in terms of the microphysical is to work).

Even more than Jackson, Chalmers thinks that two-dimensionalism is naturally combined with a semantic pluralism, according to which there are numerous entities which can play some of the explanatory roles that the notion of a proposition, for example, is supposed to play: not just primary propositions and secondary propositions (Jackson’s A-intensions and C-intensions for sentences) but also structured entities containing the primary and secondary intensions of component expressions of the sentences in question. He thinks that primary intensions, perhaps structured versions that combine with extensions, can be used to capture something like Frege’s notion of sense (Chalmers 2002a), while an appeal to a combination of structured primary and secondary intensions can help provide the semantics of propositional attitude ascriptions. Chalmers (2002b), for example, suggests that an utterance of ‘S believes that P’ is true just when the referent of S has a belief with the structured secondary intension of ‘P’ (in the mouth of the ascriber) and with an appropriate structured primary intension, where the range of what is ‘appropriate’ may depend on the context of utterance. As Chalmers acknowledges, such a view is closely related to ‘hidden-indexical’ analyses of belief ascriptions, with primary intensions playing the role of ‘modes of presentation’.

Conclusion

This review has inevitably been partial and to a degree biased—there has been significant work done by Australasian philosophers on particular issues in the philosophy of language that I have not tried to cover here. But even though the coverage has been partial, I think it has shown that Australasian philosophy of language has managed to develop a voice of its own or rather voices of its own. For it is abundantly clear that there is no Australasian or even Australian orthodoxy in the philosophy of language: no Australasian semantics, say, in the way there is—or was—a Californian semantics. Not only does the work of Australasian philosophers range widely, but even if we focus only on the theory of meaning we find many differences. While it is true that there is a widespread commitment to finding a naturalistic understanding of the phenomenon of meaning, it is also true that the concept of naturalism tends itself to be in contention, as is shown by the striking differences among the conceptions of meaning and of the representational nature of language found in the work of philosophers like Jackson, Devitt and Price. Another common theme is the way in which Australian philosophers in particular tend to have strong and distinctive views about the relation of philosophy of language to areas like metaphysics, although here again the views vary considerably. Jackson and Chalmers in particular think that armchair metaphysics is a possible and productive enterprise. Devitt thinks that this is a vain hope that ‘conceptual analysis’ at best yields folk-theories that may deserve rejection and that the philosophy of language cannot settle problems in metaphysics or even ethics. Barry Taylor demurs, insisting that sceptical worries about the semantic properties of language argue for a form of anti-realism. The voices, then, are many, and the noise they make discordant. Given the origins of philosophy of language in Australasia (and perhaps the nature of Australasian philosophy in general), we could hardly have expected anything less.