By Sari Pietikäinen, Alexandra Jaffe, Helen Kelly-Holmes, Nik Coupland
Sociolinguistics from the Periphery "presents a fascinating book about change: shifting political, economic and cultural conditions; ephemeral, sometimes even seasonal, multilingualism; and altered imaginaries for minority and indigenous languages and their users"
Date: Tue, 1 Jun 2004 11:22:25 +0200 From: Anne Reboul Subject: Reflections and Replies: Essays on the philosophy of Tyler Burge
EDITOR: Hahn, Martin; Ramberg, Bjørn TITLE: Reflections and Replies SUBTITLE: Essays on the philosophy of Tyler Burge PUBLISHER: MIT Press YEAR: 2003
Anne Reboul, Institute for Cognitive Sciences, CNRS, France
SUMMARY The book is neatly divided in two parts, the first one filled with articles commenting on various aspects of Tyler Burge's philosophy, the second part presenting Burge's answers to these comments. It opens with an introduction presenting Burge's contributions to philosophy as well as an outline of the following papers (both comments and answers). As the content of the individual papers will be presented below, I will only report on the presentation of Burge's philosophy. The editors (who authored the introduction) concentrate on Burge's contribution to anti- individualism (or externalism) about mental content. This began in a paper (''Individualism and the mental'') in which Burge introduced a variation on Putnam's Twin Earth thought experiment. In that experiment, Putnam (1975) introduced Twin Earth, which is physically identical to Earth but for one fact: what passes there for water does not have chemical composition H2O but XYZ. Thus, though a given Earthian and his/her Twin Earth counterpart (being themselves physically identical) are apparently in the same psychological state, the meanings for their respective words ''water'' are not identical. This led Putnam to the famous conclusion that ''meanings ain't in the head'', thought specifically to apply to natural kind terms through an intrinsic indexicality of these terms. In Burge's version, a patient goes to his/her doctor complaining of arthritis in his/her thigh, to be immediately informed by the doctor that arthritis being in the joint, the patient does not suffer from arthritis in his/her thigh. On Twin Earth, the term ''arthritis'' is not so restricted and the twin patient thus uses the word correctly but does not mean the same by it as the Earthian patient. Burge's conclusion differs from Putnam in that he concludes that the twins are not in identical psychological states, because they use different concepts with different intentional contents. Thus ''Burge's anti-individualism concerns thought, not linguistic meaning'' (xiv) and it is the social facts not the physical environment that varies. Burge also rejected in following papers the idea that it is indexicality which is the issue in Putnam's original experiment. Another important aspect of his views is the idea that we can entertain concepts that we incompletly understand. The debate has raged not only regarding the definition of ''concept'' but also on the possible consequences of Burge's specific brand of anti-individualism: for instance, doesn't it lead to a rejection of first-person authority? And what of the local supervenience of the mental on the physical?
The first paper, ''Burge, Descartes, and Us'', by Normore discusses a paper on Descartes by Burge and attempts to recast Descartes' contribution to philosophy of mind in a new perspective. Burge's claim in the original paper was that Descartes was the origin of individualism in philosophy of mind, where ''individualism'' is to be understood as the view that mental states can be individuated independently of the nature of the external world and its components. Normore proposes to go back to Descartes to check the validity of Burge's comment. This leads him to a central question in contemporary philosophy, i.e., how can the intuition that reference is direct, fixed by (causal) relations with external objects, be combined with another intuition to the effect that concepts and percepts can misrepresent their objects? This problem, according to Normore, is also at the heart of the Cartesian account of representation. Externalism was available to Descartes and it is Normore's contention that Cartesianism is, to a certain degree, externalist. The paper closes with a letter purporting to be from Descartes to Burge, the conclusion being that both Burge and Descartes are externalists, albeit of a different sort.
Barry Stroud signs the next paper, on ''Anti-individualism and Skepticism''. As he notes, anti-individualism about perception requires not immunity to error, but the ability to perceive things as they are. This raises the further question of whether this is enough to reject skepticism about perception, as was claimed by Davidson. This, according to Stroud, is not a claim that Burge would want to make. More modestly, Burge would want to prove ''for some of the particular things that philosophical scepticism would say we don't know, that we do or can know them after all'' (19). The reason for Davidson's stronger view is to be found in his notion of ''interpretation'' of one another's behavior and communication: this can only be done through common reference to external objects and states of affairs. This leads Davidson to the view that belief is, by its very nature, veridical. For Burge, by contrast, though anti- individualism is supported by our interpretive activities, this does not entail that beliefs are mostly true. Skepticism starts from the logical point that all our beliefs could be false. The idea that the more moderate anti-individualist view can defend is that, applied to any specific set of attributed beliefs, skepticism is not a relevant possibility: indeed, it is in this respect rather similar to Moore's paradox (i.e., ''it is raining, but I don't believe that it is raining''). Skepticism would not thereby be refuted, but it could be neglected.
The third paper, ''When Swampman get arthritis: ''Externalism'' in Burge and Davidson'', authored by Martin Hahn, compares the two varieties of externalism advocated by Burge and Davidson. Davidson distinguishes between physical and social versions of externalism, considering the first kind to be ''perceptual externalism''. This is where Burge and Davidson are mostly taken to agree though there is also a major point of agreement on social externalism, regarding the necessity of being in touch with other people with substantially similar innate perceptual and conceptual abilities. They also both claim that first- person authority is compatible with externalism. The main difference is in the dependence of content on social circumstances. Davidson's view on the matter is based on the triangulation argument, which derives from the thesis of radical interpretation, according to which meaning and content are assigned to others' thoughts and utterances on the basis of what we perceive of their environment and their perceptual relations with objects in that environment. This makes our thoughts conditional to living with creatures similar to us in the relevant ways. By contrast, Burge rejects radical interpretation and favors the moderate view that ''we can stand corrected in our own idiolectal usage'' (39). Regarding physical externalism, despite the similarities noted above, Burge's and Davidson's outlooks are also different in that, to go back to Twin Earth examples, Davidson would be inclined to consider the psychological states of the twins are identical or not depending on whether the chemistry on the two planets is sensitive to the difference between water and Twin water, while Burge would not take this last point into account. What's more, Davidson rejects the notion that having a concept does not entail having a complete mastery of it, a notion central to Burge's philosophy. The conclusion is that, despite common belief, Burge and Davidson turn out to agree more on social anti- individualism than they do on physical anti-externalism.
Donnellan turns to ''Burge's thought experiments'', outlining the difference between Burge's versions and Putnam's. Where Putnam relies on physical differences between Earth and Twin-Earth, Burge concentrate on differences in the linguistic communities. Both types of thought-experiments rely on experts, but Putnam's experts are experts about the physical world, while Burge's are those members of the speech community who can be expected to know the meaning of the words. Putnam's thought experiments highlight a very general semantic rule, which is not the case for Burge's. Finally, Putnam's invoke a deep indexical character, and again this is not the case in Burge. Donnellan takes up this last point, criticizing Putnam's original account for saying that natural kind terms have an indexical element, pointing out that this is incorrect on any reasonable account of indexicals. Rather the indexicality is to be found in the semantic rule for applying natural kind terms, through a reference to local paradigms. The relevant properties would be external but the semantic rule directing the speakers to them would be internal and common between Earthian and Twin-Earthians. As Donnellan points out, ''if all of this is correct, Putnam thought experiments (~E) cannot be used for an anti-individualist point'' (63). What's more the indexicality is important in that it is the basis on which Putnam's experiments can be generated. This is a major difference with Burge thought experiments where it is not clear that a semantic rule is necessary, because what is at issue is incomplete mastery. What is more, the principle of division of linguistic labor, advocated by Putnam, and which might be a common ground with Burge, is not dependent on Putnam's experiments. This leads Donnellan to suggest that in Burge's arthritis example, both patients have beliefs with the same content, the external circumstances determining not content but truth-values, which differ.
The next paper, by Joseph Owens, returns to indexicality (''Anti-individualism, indexicality and character''). The author examines the view that ''the individualistic character of the mental [can be] retained once we realize that the concepts [the twins] use in expressing their beliefs have an indexical element'' (77). His conclusion is that this line of defense is inappropriate. The idea behind the indexical view is that ''the problematic thought contents contain an indexical element'' (79). Its popularity has to do with Kaplan's theory of indexicals and, more specifically, with the notion of character, i.e., the constant linguistic meaning of a given indexical, later generalized to linguistic items in general. Thus, the twins agree on the characters and are in the same psychological state (identical cognitive content), even though their sentences do not express the same proposition. Though appealing, the solution does not work for an obvious reason: the stability of the character is irrelevant to the twins situation because they do not speak the same language. What is more, ''character itself does not supervene on nonintentionally individuated states'' (88; italics in the original). This is because ''the intuitions that gave rise to anti-individualistic model of the mental also support an anti-individualistic reading of character'' (94). Thus Burge is vindicated.
The next and sixth chapter, ''Competence with demonstratives'' is one of the two by a linguist, James Higginbotham, and one of the few in the book that relates to another aspect of Burge's philosophy, in this instance his (earlier) contributions to philosophy of language. The paper defends the view that it is impossible to preserve both reference and perspective in belief or speech reports. This is especially true of the first-person, though also true for unarticulated components, in, e.g., incomplete definite descriptions or quantificational noun phrases. Higginbotham turns to Burge's account of the place of demonstratives in theories of truth for natural languages. This had two features, truth-at-an-index, and the use of a bound variable in the statement of truth-conditions. The first one correspond to a rule of use, characteristic of both demonstratives and indexicals, and the second to the fact that this rule of use is not part of the truth- conditions. This has the consequence that though the perspective can be conveyed, it cannot be said, on pain of loosing referential faithfulness.
Christopher Peacocke, in the seventh paper, ''Implicit conceptions, understanding and rationality'', turns, as does Higginbotham, on Burge's early work in philosophy of language, specifically on Burge's reading of Frege on sense. The paper concentrates on primitive concepts, i.e., concepts which cannot be accessed through inference, but which are nonetheless rational. Peacocke's example is logical connectives, where a capacity to exercise simulation is present before any explicit rules are acquired, and, indeed, is the basis for the acceptance of such rules. This capacity is based on an implicit conception of the connectives. This is similar to the Fregean distinction between grasping a concept (possibly implicitly) and grasping a concept sharply (explicitly). The benefits involved in passing from an implicit to an explicit conception (which is not a simple endeavor) include generality, establishing clear limits and justification. The notion of implicit conception cannot be accommodated in conceptual role theories because they fail to explain how new principles can be arrived at. Implicit conceptions, which can be acquired as well as innate, may enter content-involving psychological explanations, for instance of categorical judgments. Finally, they are the basis of rational justifications for many inferences.
The next paper, by Fred Drestke, returns to Burge's anti- individualism and is entitled ''Burge on mentalistic explanations, or why I am still epiphobic''. It begins by mentioning the numerous points of agreement between the author and Burge, on realism about psychological content, semantic externalism and the non-epiphenomenalism of psychological content. The difference is that Drestke thinks that epiphenomenalism is a potential (though surmontable) problem for semantic externalism, while Burge does not even think that it is a problem. In Dretske's words, ''I can be puzzled about [how things can work that way], while conceding that things do work that way'' (154). The problem arises mainly for behavior that is not externally individuated, for example behavior that is described at the basic level of motion, regardless of the objects of the specific motion involved. This is a level at which externally individuated beliefs are dispensable in the explanation of the behavior. Here, Dretske introduces the example of the vending machine, which dispenses Cokes in exchange for quarters. Though it is designed to work with quarters, it would work just as well with any piece of metal relevantly similar to quarters. In other words, it is the intrinsic properties of quarters (those to which the machine positively answers), which enter in any lawlike explanation of the machine behavior, not their extrinsic properties (being quarters). Thus, these extrinsic properties are epiphenomenal to the delivery of Cokes. Mutatis mutandis, the same applies to mind-body interaction and epiphenomenalism, no matter how much one may reject it, has to be taken seriously and must be answered.
Ned Block takes the debate to the problem of consciousness in the next paper, ''Mental paint''. A recent question is whether qualia (the qualitative properties of conscious experience) are exhausted by their representational content or not. Block defends the negative answer, a position that he dubs ''phenomenism''. He shortly present and criticizes a version of functionalist internalist representationism before turning to externalist representationism, which he defends in part. He then points out that if qualia supervene on the brain, representationism is refuted. Representationism has often argued from the supposed diaphanousness of perception, which is correct for attention though not for awareness. There are however deeper questions: the very existence of ''mental paint'' (the irreducible qualia), whether this is or not accessible to introspection; the existence of qualia with no representational content. This leads Block to a triple distinction between the intentional content of an experience, the mental properties (mental paint) of such an experience, and the mental properties of non- representational experiences (e.g., orgasms). He concludes his paper by a discussion of various thought experiments, among them the Inverted Earth (identical to Earth but for the inversion of colors: e.g.,the sky is yellow there).
Bernard Kobes, in ''Mental content and hot self-knowledge'' returns to the problem of first-person authority, approaching it in an orginal way: he defends the idea that self-attribution of thoughts is not passive but active ~W it has a certain performative character ~W and that the thinker has knowledge of his performances. He takes on board quite a few of the theoretical notions defined in speech acts theory such as the distinction between mental content (propositional content) and mental relation (illocutionary force), as well as the distinction between two directions of fit, here mind-to-world vs. world-to-mind. The first (thetic) is typical of belief while the second is typical of intention and desire (telic). He reminds the reader of the supposed antagonism between first-person authority and an externalist semantics as well as of the solution proposed by Burge, to wit that the higher-order (self- attributive) thought shares its mental content with the lower order one. However, some criticisms have been leveled against this solution, first for thoughts relative to historical characters and events (''Socrates drank some hemlock''), where externalism is not entirely obvious, as well as memory of past thought episodes. Thus the extendibility of the solution is not obvious. Kobes, through the notion of the thinker as a cognitive agent, tries to answer these doubts. His solution is to treat the relation of the thinker to his/her self-attribution as telic, which gives him/her authoritative self-knowledge.
Brian Loar returns to the subject, partly treated by Higginbotham and Block, of ''Phenomenal intentionality as the basis of mental content''. He remarks that grasping other people's thoughts is not a mere matter of impersonal representation, but that it involves taking the same intentional perspective. This, if ''conceivings are in the head'' (229), and if intentionality is essential to them, seems to lead to the idea that intentionality is in the head, in opposition to externalist claims. This is not reflected in ''that'' clauses, no matter how ''oblique'' they are. Loar defends the notion that this is because mental content is phenomenal and is conveyed ''in the gaps between the words'' (230). On that view, ''the internal intentionality of perceptions and thoughts consists in their apparent directness, in their purporting subjectively to refer in various complex ways'' (231). The main thesis of the paper is that though externalists are right about reference and truth-conditions of thought, intentionality still is internal. The explanation for that apparent contradiction is that, contrary to externalist theses, ''intentionality does not presuppose reference and is not externally determined'' (231).
This leads us to the last chapter of the comments part of the book, ''Internalist explorations'', by Noam Chomsky. Chomsky begins by defining a Humean ''science of human nature'' as ''individualist and internalist'' (259), rejecting theories based on the commonsensical meaning of words like ''thought'' and ''belief'' in favor of theories based on the technical meaning of such words, that, according to him, can only be internalist. He illustrates his point by appeal to linguistics, and more precisely to the study of the language faculty, which is divorced from commonsensical and intuitive notions and based on the concept of I-language where I stands for ''internal and individual, and also intentional'' (263). He rejects Twin Earth and Swampman thought experiments on two grounds: judgements elicited in such a way are unclear and thus unreliable; folk semantics has nothing to bring to the scientific study of the language faculty.
The second part of the book, ''Tyler Burge replies'', opens with ''Descartes and anti-individualism: Reply to Normore''. Burge begins his paper by recognizing a mistake in his interpretation of Descartes as an individualist, though he insists that the interpretation of Descartes is complex regarding his individualist or anti-individualist position. Some of Descartes' positions are consistent with anti- individualism but do not entail it. The main problem is to do with Cartesian dualism (presupposing the metaphysical independence of the mental), which seems, on the face of it, contradictory with anti-individualism. If Descartes is not an individualist, then he ''must reconcile these views by holding that mental states and events that are about body are not essential to any given mind, or to being a mind'' (293), which, again, seems consistent with some Cartesian views. This leads to a new construal of Cartesian dualism, according to which ''some particular thoughts (~E) are necessarily dependent on physical properties'' (294). Then, the analysis of the cogito would be that ''the reflexive self-consciousness involved in the continuing I think is filled out by particular thoughts, which are themselves contingent modes of the mind'' (298). Burge then turns to his argument for perceptual anti-individualism in the paper that Normore comments, recognizing an error of presentation regarding the relation between the local failure of supervenience of intentional on physical states and anti-individualism. Such a failure is neither necessary for, nor identical with, anti-individualism. Finally Burge precises his anti-individualist view of perception and discusses Descartes' purported letter to him.
The next chapter, ''Some reflections on scepticism: Reply to Stroud'', begins by the reminder that ''some veridicality is implicit in the relevant conception of normality, but not in a way that guarantees that an individual has any veridical perceptions at all'' (335). He criticizes the coherence view of Davidson, noting that most perceptual beliefs are not inferentially or derivately warranted. This leads him to scepticism and anti-individualism: ''Scepticism is about knowledge or justification. Neither anti- individualism nor the slogan that error presupposes veridicality says a word about knowledge or justification'' (338). Thus anti-individualism is not an answer to scepticism, though it can be a component to such an answer.
Chapter 15, ''Davidson and forms of anti-individualism: Reply to Hahn'', opens with remarks about the relation between Burge's position on de re thoughts (they are necessary for having propositional attitudes, including de dicto thoughts and they are incompletly conceptualized) and his anti-individualist position (which concerns both de re and de dicto attitudes) in the paper that Hahn comments. Burge then remarks that anti-individualism is not limited to the individuation of concepts but rather to that of mental states and thus is compatible with various views about concepts. What is more anti-individualism is committed both to the notion ''that the individuation presupposes a background of referential success [and] that attitudes are in part individuated in terms of nonintentional relations (~E) that the individual bears to objects, properties, or relations in the environment'' (351). This does not reject internal (e.g., inferential) relations about attitudes. Burge then turns to Hahn's comparison between him and Davidson and admits that he rejects interpretationism. He briefly discusses Swampman and turns to the issue of anti-individualism and inateness, noting that his position is metaphysically compatible with innateness. He then turns to his own brand of social externalism, insisting that on his view, the patients in the arthritis thought experiment have different concepts (because of their respective linguistic communities) and hence different beliefs.
''The thought experiments: Reply to Donnellan'' begins by a comparison between Burge's arthritis and Putnam's Twin Earth thought experiments, noting that the main similarity is that both types of experiments rest ''on the fact that we are not omniscient and the fact that there is possible slack between what we know descriptively about the referents, or correct applications, of our concepts and what their referents are'' (363). Burge then discusses the four differences highlighted by Donnellan between his and Putnam's experiments, concentrating on the last one, the place of indexicality, which he thinks is misleading. He questions the indexical semantic rule proposed by Donnellan, insisting that, as the psychological states of the twins are not identical, it is just not necessary for it to be indexical and the rule itself is neither necessary nor plausible. More generally, Burge rejects the use of such metalinguistic rules to account for incomplete mastery of concepts.
''The indexical strategy: Reply to Owens'' is short given a general agreement between Burge and Owens on the subject. Thus Burge limits himself to a few qualifications to Owens' paper and to a discussion of a thought experiment proposed by Owens.
The eighteenth chapter ''Tracking perspective: Reply to Higginbotham'' concerns one of the two contributions by linguists to the book. Although Burge expresses his sympathy with most of Higginbotham's paper, he disagrees on the idea that preserving reference implies losing the perspective of others in indirect discourse. His argument rests on the difference between the rule for contextual reference for a given indexical and the meaning of that indexical, thus rejecting Higginbotham's division of the lexicon in the two classes of items with meanings and items with rules of use. Burge concludes that ''a satisfying investigation of language must sometimes go beyond the bounds of linguistic meaning, or linguistic understanding, conservatively construed'' (382).
The next chapter, ''Concepts, conceptions, reflective understanding: Reply to Peacocke'', begins by an acknowledgment that reflective acceptance of conceptual truths and the application of incompletely understood concepts may be accounted for in mentalistic terms through the notion of implicit conceptions. Burge nevertheless tries to precise the notion of implicitness used by Peacocke, proposing that ''implicit conceptions are unconscious psychological conceptual structures that explain our ability to apply concepts to cases, or to realize that principles involving concepts are true'' (383), proposing that such implicit conceptions might be neurally represented in a syntactic way, though this still leaves it open whether such conceptions are subpersonal or personal. His doubts concern whether implicit conceptions can be as widely used as Peacocke seems prepared to allow. In addition, implicit conceptions could themselves be incomplete, which leads to a futher application of anti- individualism. The paper ends with a (slightly critical) discussion of Peacocke's interpretation of Leibniz and Frege and a defense of rationalism about truths.
Chapter 20, ''Epiphenomenalism: Reply to Dretske'', opens with a statement to the effect that there is indeed a disagreement about epiphenomenalism, which concerns both its being a threat and the method for its rejection. Though Burge claims that he is interested in the mind-body problem, he owns to a negative attitude regarding epiphenomenalism as a serious notion and about materialist metaphysics. Returning to epiphenomenalism and anti- individualism, he notes that though some relational properties are causally irrelevant, not all relational properties are and that, similarly, though some higher- level properties are causally irrelevant, not all higher- order properties are. Burge localizes Dretske's worry about epiphenomenalism in a confusion between mental events as individuated through their contents and the relational facts that underlie those contents. If the confusion is led to rest, it takes care of the worry about epiphenomenalism. The paper ends with a rejection of materialism regarding mental events.
In the next chapter, ''Qualia and intentional content:Reply to Block'', Burge argues, mainly in accordance with Block's view, that having color concepts partly depends on ''bearing relations to the colors in a broader environment'' (405) and agrees ''that there are non representational qualitative mental properties'' (405). This is, again in agreement with Block, against the representationalist views of sensations. Burge then turns to Block's discussion of Inverted Earth, outlining the importance of evolution in color perception. Thus, on Inverted Earth, ''representational content changes without any corresponding change in phenomenal character'' (411). However, ''the perceptual intentional content (~E) will commonly be in some way, at some level, different, if phenomenal character is different'' (412). Burge then turns to Block's remark that phenomenal character may vary among normal humans. Burge defends the notion that this variation is consigned not to global identification of color, but rather ''to a narrowly discriminated and quickly forgotten shade'' (413), and is thus irrelevant for intersubjective color typing.
Burge then returns to authoritative self-knowledge in ''Mental agency in authoritative self-knowledge: Reply to Kobes''. He is in global agreement with Kobes, notably on the point that ''some performative knowledge of one's propositional attitudes is partly constitutive of being a critically rational agent'' (417). Burge notes that knowledge about one's own actions does not have the same roots depending on whether it concerns physical or mental actions. Knowledge of physical action is subject to brute contingencies and hence to brute error, which is not the case for knowledge of one's own mental states. Burge then introduces a distinction between performative self- knowledge which is logically self-verifying (e.g., I am hereby entertaining the thought that writing requires concentration) and which corresponds to pure cogito cases and performative self-knowledge that is not (e.g. I am hereby thinking (in the sense of comitting myself to the view) that writing requires concentration), or impure cogito cases, of which, however, Burge remarks that they are abnormal if not pathological. He then notes that performative self-knowledge involves a reflexive element: ''there is (~E) a reflexive second-order element in the logical form of the first-order thoughts'' (419). However, Burge considers that ''the formation of a great number of beliefs, particularly perceptual beliefs, is not strictly an activity'' (419) and that the performative model thus cannot fully explain first-person authority. ''Performative and reflexive cases are such that the intentional content that they attribute is thought and thought about at the same time. So the content of the attributed bottom-level attitude and the content attributed in the self- attributional thought are locked together'' (426). This means that in slow-switching Twin Earth cases both contents will come apart at the intentional level. In non-reflexive cases, memory will ''preserve the content between different attitudinal states over time'' (431).
''Phenomenality and reference: Reply to Loar'' opens with the acknowledgement of a fundamental difference, given that Loar defends an internalist view, but expresses a doubt as to the reasons why Loar thinks an internalist view necessary. According to Burge, an internalist view is not necessary to fight materialism: indeed the view that a conceiving's intentional properties are essential to it, even combined with anti-individualism defeats, at most, type and token identity materialism. Anyway, ''spatial location is not the central issue'' (435). Burge then turns to opaque contexts and insists that ''differences in oblique occurrences in true propositional attitude attributions prima facie signal differences in mental content'' (438). What is more, intentional states may lack referents: this fact does not contradict anti-individualism, and neither do brain-in-the-vat thought experiments or any other cases of non-reference.
The final chapter, ''Psychology and the environment: Reply to Chomsky'', opens with a commendation of Chomsky's work and a defense of it against frequent philosophical objections. It begins with the points of agreement, i.e., the rejection of eliminativism about mental kinds, the distinction between scientific and commonsensical psychological notions, the inaccessibility to consciousness of many linguistic and psychological structures, the frequent innateness of linguistic and psychological abilities, the notion that the holism of meaning is no barrier to its scientific study. Burge then turns to the differences, notably in terms of internalism or individualism, though he notes that the notion of internalism defended by Chomsky is not well-defined enough to allow precise discussion. He notes that the notion of I- language is not incompatible with anti-individualism (which ''does not presuppose the existence of public languages'' (453)). What is more, ''internalism or individualism (~E) is not simply a claim that psychology studies the internal states of individuals'' (453). Burge also considers Chomsky's skepticism about intuitions in thought experiments an overreaction. He then turns to the part of Chomsky's article where the author tries to justify his internalist position (where ''internalism'' is interpreted as meaning that no cognitive mechanism can be investigated as related to anything external) by appeal to various scientific theories, notably Marr's theory of vision and notes that Chomsky does not give an adequate discussion of Marr's work, in which the ''main objective is explicitly stated to be that of explaining how we visually determine the properties of actual objects in physical space that we in fact visually represent as they are'' (464). Thus, semantic theories can be both internal in Chomsky's sense and anti-individualist in Burge's.
CRITICAL EVALUATION In ''The voyage out'', Virginia Woolf has one of her characters reminisce about his Cambridge youth and the philosophical discussions he then had with his fellow students, making him say ''It's the arguing that counts''. So it was then, at the end of the 19th and beginning of the 20th century at Cambridge when Georges Moore and Bertrand Russell were the philosophers in residence and so it is now in analytic philosophy. Yet, it would be very short-sighted of linguists to dismiss analytical philosophy as mere hair- cutting, and this is especially true of the questions raised by Burge's work. The issue of anti-individualism is relevant at most levels in linguistics (though maybe not in phonology), the more obvious being semantics and pragmatics. Parts of the debate in the present book are deeply relevant to issues in contemporary pragmatics, such as for instance the appropriate conclusion to draw from Burge's arthritis thought experiment, as well as the chapters outlining the difference between Burge's and Davidson's views of social anti-individualism. One such issue may be the Gricean notion of non-natural meaning and the way one could or not link it to those two versions of social anti-individualism. Another issue, on the semantic side, would be the degree of agreement and the points of disagreement between different semantic approaches (e.g., model-theoretical versus situation semantics) and the specific brand of anti-individualism that has been defended by Burge. There is no space here to go into details and I will not attempt to do so, but all of these questions should be taken into account and thought about seriously by linguists. Thus, it is a book well worth reading for any audience interested in semantics and pragmatics, as well as cognitive science more broadly. Though densely packed and complex in some places, this book is in general very clear and highly rewards reading efforts.
REFERENCES Burge, T. (1979), ''Individualism and the mental'', in Midwest Studies in Philosophy 4, 73-121.
Putnam, H. (1975) ''The meaning of 'meaning''', in Mind, Language and Reality. Philosophical Papers, vol. 2, Cambridge, Cambridge University Press, 215-271.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Anne Reboul is a First Class Research Fellow at the French Center for Scientific Research (CNRS). 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 an Encyclopaedic Dictionary of Pragmatics and quite a few papers in French and English, on pragmatics and/or philosophic subjects.