Three Kinds of Nonconceptual Seeing-as (Review of Philosophy and Psychology, special issue edited by Robyn Carson and Kepa Korta): It is commonly supposed that perceptual representations in some way embed concepts and that this embedding accounts for the phenomenon of seeing-as.  But there are good reasons, which will be briefly reviewed here, to doubt that perceptions embed concepts.  The alternative is to suppose that perceptions are marks in a perceptual similarity space that map into locations in an objective quality space.  From this point of view, there are at least three sorts of seeing-as.  First, in cases of ambiguity resolution (such as the duck-rabbit), the schematicity of the figure leaves us with a choice as to where in perceptual similarity space to place a mark (for example, closer to the marks that represent rabbits or closer to the marks that represent ducks).  Second, in cases where expertise affects perception (as when, for example, we learn to distinguish various kinds of tree leaves), the accumulation of perceptual landmarks permits a more precise placement of a mark in perceptual similarity space.  Third, extensive experience with an object (e.g., the family dog) allows similarity to that object to serve as an acquired dimension in perceptual similarity space, which in turn affects the relative similarities of other objects.

Open Texture and Schematicity as Arguments for Non-referential Semantics
(forthcoming in Sarah-Jane Conrad and Klaus Petrus, eds., Meaning, Context, and Methodology, De Gruyter Mouton): Many of the terms of our language, such as “jar”, are open-textured in the sense that their applicability to novel objects is not entirely determined by their past usage.  Many others, such as the verbs “use” and “have”, are schematic in the sense that they have only a very general meaning although on any particular occasion of use they denote some more particular relation.  The phenomena of open texture and schematicity constitute a sharp challenge to referential semantics, which assumes that every non-logical term has a definite extension.  A different, non-referential approach to formal semantics defines truth as relative to a context and defines contexts as built up from exclusively linguistic entities.  For any given utterance of a sentence, there will be one of these contexts that pertains to it.  In this framework, open texture and schematicity can be understood as consequences of the complex nature of the pertaining relation between contexts and utterances.

Do Perceptions Justify Beliefs? The Argument from "Looks" (forthcoming in J. Gersel, R. T. Jensen, M. S. Thaning and S. Overgaard, eds., In Light of Experience: Essays on Reason and Perception, Oxford University Press): Why should we believe that perceptions justify beliefs? One argument starts with the premise that sentences of the form “a looks F” may be used to justify conclusions of the form “a is F”. That will show that perceptions justify beliefs only if the term “looks” in “a looks F” expresses a peculiar modality that indicates that the proposition in which it is incorporated is the content of a visual perception. But in that case, the premise, “a looks F”, does not provide any justification for the conclusion that “a is F”; so the argument does not work anyway.

Presuppositions as Anaphoric Duality Enablers (forthcoming in a special issue of Topoi, edited by Filippo Domaneschi, now available on-line on journal website): The key to an adequate account of presupposition projection is to accommodate the fact that the presuppositions of a sentence cannot always be read off the sentence but can often be identified only on the basis of prior utterances in the conversation in which the sentence is uttered. In addition, an account of presupposition requires a three-valued semantics of assertibility and deniability in a context. Presuppositions can be explicated as sentences that belong to the conversation and the assertibility of which ensures that the remaining assertibility and deniability conditions of the presupposition-bearing sentence are dual to one another. The prevailing approach to presuppositions, grounded in Heim’s context-change semantics, can be criticized both on philosophical grounds and for failing to accommodate the phenomena.

How Many Bare Demonstratives are There in English? (Linguistics and Philosophy 37, 2015: 291-314): In order to capture our intuitions about the logical consistency of sentences and the logical validity of arguments, a semantics for natural language has to allow for the fact that different occurrences of a single bare demonstrative, such as "this", may refer to different objects. But it is not obvious how to formulate semantic theory in order to achieve this result. This paper first criticizes several proposals: that we should formulate our semantics as a semantics for tokens, not expressions, Kaplan's idea that syntax associates a demonstrative with each occurrence of a demonstrative, Braun's idea that a context may specify shifts in context across the evaluation of the expressions in a sentence; and Predelli's idea that we should countenance different classes of contexts. Finally, a solution is proposed that allows that a natural language persists across the addition of basic lexical items but defines logical properties in terms of language stages. A surprising result is that we do not need to think of demonstratives as taking different referents in different situations.

The Illusion of Semantic Reference (in Andrea Bianchi, ed., On Reference, Oxford University Press, 2015, pp. 21-39): A lot of us have given up on the idea that there will be a naturalistic account of the relation of semantic reference and so have resolved to formulate our theories of semantics and communication without appeal to semantic reference. Still, there is a resilient intuition to the effect that I know the extensions of the terms of my language. This paper explicates that intuition without yielding to it. The key idea is to give a “skeptical” account of what it is to “know the meaning” of a word, by which I mean an account of the status that is granted to a person in saying that he or she “knows the meaning” of a word..

Incomplete Thoughts (in Laurence Goldstein, ed., Brevity, Oxford University Press, 2013, pp. 74-90): It is often assumed that, though we may speak in sentences that express propositions only inexplicitly, our thoughts must express their propositional contents explicitly. This paper argues that, on the contrary, thoughts too may be inexplicit. Inexplicit thoughts may effectively drive behavior inasmuch as they rest on a foundation of imagistic cognition. The paper also sketches an approach to semantic theory that accommodates inexplicitness in mental representations as well as in spoken sentences.

Logical Nihilism in Contemporary French Philosophy (Teorema 32 (2013): 65-79): Recanati takes for granted the conveyance conception of linguistic communication, although it is not very clear exactly where he lies on the spectrum of possible variations. Even if we disavow all such conceptions of linguistic communication, there will be a place for semantic theory in articulating normative concepts such as logical consistency and logical validity. An approach to semantics focused on such normative concepts is illustrated using the example of “It’s raining”. It is argued that Recanati’s conception of semantics as involving the pragmatics of saturation and modulation cannot account for the logical properties of “It’s raining”.

Perception without Propositions (in John Hawthorne and Jason Turner, eds., Philosophical Perspectives 26: Philosophy of Mind, Blackwell, 2012: 19-50): In recent years, many philosophers have supposed that perceptual representations have propositional content. A prominent rationale for this supposition is the assumption that perceptions may justify beliefs, but this rationale can be doubted. This rationale may be doubted on the grounds that there do not seem to be any viable characterizations of the belief-justifying propositional contents of perceptions. An alternative is to model perceptual representations as marks in a perceptual similarity space. A mapping can be defined between points in perceptual similarity space and points in an objective quality space. The correctness of perceptual representation can then be defined as a kind of accuracy of mapping rather than as the truth of a proposition. The phenomenon of seeing-as can be accounted for as a matter of the location of marks in perceptual similarity space relative to other marks in perceptual similarity space. Perceptual representations, on this account, will not justify beliefs, but they may nonetheless guide judgment.

Semantics and Pragmatics (in Delia Graff Fara and Gillian Russell, eds., Routledge Companion to the Philosophy of Language): Semantics deals with the literal meaning of sentences. Pragmatics deals with what speakers mean by their utterances of sentences over and above what those sentences literally mean. However, it is not always clear where to draw the line. Natural languages contain many expressions that may be thought of both as contributing to literal meaning and as devices by which speakers signal what they mean. After characterizing the aims of semantics and pragmatics, this chapter will set out the issues concerning such devices and will propose a way of dividing the labor between semantics and pragmatics. To semantics belongs the job of defining the conditions under which a sentence is true relative to a context. To pragmatics belongs the job of explicating the conditions under which a given context pertains to a given conversation.

Indirect Discourse, Relativism, and Contexts that Point to Other Context (in François Recanati, Isidora Stojanovic, Neftali Villanueva, eds., Context-dependence, Perspective and Relativity in Language and Thought, Mouton, 2010): Some expressions, such as “all” and “might”, must be interpreted differently, relative to a single context, when embedded under “says that” than when unembedded. Egan, Hawthorne and Weatherson have appealed to that fact to argue that utterance-truth is relative to point of evaluation. This paper shows that the phenomena do not warrant this relativistic response. Instead, contexts may be defined as entities that assign other contexts to contextually relevant people, and context-relative truth conditions for indirect discourse sentences can be satisfactorily formulated in terms of such contexts.

What Tipper is Ready for: A Semantics for Incomplete Predicates (Noûs 46, 2012: 61-85): This paper presents a precise semantics for incomplete predicates such as “ready”. Incomplete predicates have distinctive logical properties that a semantics needs to explain. For instance, “Tipper is ready” logically implies “Tipper is ready for something”, but “Tipper is ready for something” does not imply “Tipper is ready”. It is shown that several approaches to the semantics of incomplete predicates fail to accommodate these logical properties. The account offered here defines contexts as structures containing an element called a proposition set, which contains atomic propositions and negations of atomic propositions. The condition under which “Tipper is ready” is true in a context is defined in terms of the contents of the proposition set for the context. On this account, the content of the context pertinent to a conversation must be determined not by what speakers have in mind but by relations of objective relevance.

Contexts in Formal Semantics (Blackwell Philosophy Compass 5 (2010): 568-578): Recent philosophical literature has debated the question of how much context-relativity needs to be countenanced in precise semantic theories for natural languages and has displayed different conceptions of the way in which it might be accommodated. This article presents reasons to think that context-relativity is a phenomenon that semantic theory must accommodate and identifies some of the issues concerning how it ought to be accommodated.

(with Matthew Van Cleave) Linguistic Practice and False-belief Tasks (Mind and Language 25 (2010): 298-328.): Jill de Villiers has argued that children's mastery of sentential complements plays a crucial role in enabling them to succeed at false-belief tasks. Josef Perner has disputed that and has argued that mastery of false-belief tasks requires an understanding of the multiplicity of perspectives. This paper attempts to resolve the debate by explicating attributions of desires and beliefs as extensions of the practices of making commands and assertions, respectively. In this light one can explain why desire-talk will precede belief-talk and why even older children will have difficulty attributing desires that conflict with their own.

Global Domains versus Hidden Indexicals (Journal of Semantics 27 (2010): 243-270): Jason Stanley has argued that in order to obtain the desired readings of certain sentences, such as “In most of John's classes, he fails exactly three Frenchmen”, we must suppose that each common noun is associated with a hidden indexical that may be either bound by a higher quantifier phrase or interpreted by the context. This paper shows that the desired readings can be obtained as well by interpreting nouns as expressing relations and without supposing that nouns are associated with hidden indexicals. Stanley's theory and the present alternative are not equivalent, however. They differ over the status of sentences such as “Every student is happy and some student is not happy”. On Stanley's theory, this sentence will be true in some contexts, while on the present alternative it will be true in no context. Considerations in favor of the present theory's verdict on such sentences are presented. The broader question at issue is the correct way to incorporate context-relativity into formal semantics.

Against Accommodation: Heim, van der Sandt, and the Presupposition Projection Problem (in John Hawthorne, ed., Philosophical Perspectives, 22, Logic and Language, Blackwell Publishing, 2008: 171-205): This paper criticizes the dominant approaches to presupposition projection and proposes an alternative. Both the update semantics of Heim and the discourse representation theory of van der Sandt have problems in explicating the presuppositions of disjunctions. Moreover, Heim's approach is committed to a conception of accommodation that founders on the problem of informative presuppositions, and van der Sandt's approach is committed to a conception of accommodation that generates over-interpretations of utterances. The present approach borrows Karttunen's idea that instead of associating presuppositions with sentences, we should define the conditions that contexts must meet in order to satisfy-the-presuppositions-of a sentence. However, in place of Karttunen's conception of contexts in terms of common ground, the present theory substitutes a conception of contexts as objective entities that are independent of the attitudes of the interlocutors. Contexts, so conceived, may be defined as containing sets of relevant possibilities. This allows us to define the conditions under which a context satisfies-the-presuppositions-of a disjunction.

Zero Tolerance for Pragmatics (Synthese 165 (2008): 359-371, in a special issue edited by Isidora Stojanovic): The proposition expressed by a sentence is relative to a context. But what determines the content of the context? Many theorists would include among these determinants aspects of the speaker's intention in speaking. My thesis is that, on the contrary, the determinants of the context never include the speaker's intention. My argument for this thesis turns on a consideration of the role that the concept of proposition expressed in context is supposed to play in a theory of linguistic communication. To illustrate an alternative approach, I present an original theory of the reference of demonstratives according to which the referent of a demonstrative is the object that minimally and best satisfies certain accessibility criteria. Although I call my thesis zero tolerance for pragmatics, it is not an expression of intolerance for everything that might be called “pragmatics.”

A Critique of the Similarity Space Theory of Concepts (Mind and Language 22 (2007): 317-345.): A similarity space is a hyperspace in which the dimensions represent various dimensions on which objects may differ. The similarity space theory of concepts is here defined as the thesis that concepts are regions of similarity spaces that are somehow realized in the brain. Proponents of such a theory of concepts include Paul Churchland and Peter Gärdenfors. This paper argues that the similarity space theory of concepts is mistaken because regions of similarity spaces cannot do the work of concepts. In particular, regions of similarity space cannot serve as the components of judgments. In the course of this critique, it will emerge that even if similarity spaces cannot model concepts, they may model a kind of nonconceptual representation.

On the Alleged Priority of Thought over Language, in Savas L. Tsohatzidis, ed., John Searle's Philosophy of Language: Force, Meaning, and Thought, Cambridge University Press, 2007, pp. 125-142. It is obvious that there are kinds of cognition -- mental problem solving -- that do not require spoken language. But it should not be obvious that peculiarly conceptual thought is independent of spoken language. This paper is a critical survey of arguments concluding that conceptual thought must be independent of language. The special emphasis is on arguments that John Searle has put forward, but others are considered as well. These include the claim that only the intentionality of thought is "intrinsic", arguments from the nature of speech acts, appeals to the fact that animals and babies think, and the computational theory of mind (this last not being one of Searle's arguments). Finally, there is an argument from a certain conception of linguistic communication.

The Circle of Deference Proves the Normativity of Semantics (Rivista di Estetica (special issue: essays in honor of Diego Marconi) 34 (2007): 181-198.) According to normativism about meaning, as I define it, a statement to the effect that a word has a certain meaning is in effect a proposal. It is a proposal to use a word in a certain way. If the proposal is accepted, then it carries normative force. This paper is a defense of normativism, so defined. The key premise of my argument is that for every group of users of a word, the members of that group regard themselves as responsible to the usage of the other members of the group.

Against Stepping Back: A Critique of Contextualist Approaches to the Semantic Paradoxes (The Journal of Philosophical Logic 35 (2006): 393-422): A number of philosophers have argued that the key to understanding the semantic paradoxes is to recognize that truth is essentially relative to context. All of these philosophers have been motivated by the idea that once a liar sentence has been uttered we can "step back'' and, from the point of view of a different context, judge that the liar sentence is true. This paper argues that this "stepping back'' idea is a mistake that results from failing to relativize truth to context in the first place. Moreover, context-relative liar sentences, such as "This sentence is not true in any context'' present a paradox even after truth has been relativized to context. Nonetheless, the relativization of truth to context may offer us the means to avoid paradox, if we can justifiably deny that a sentence about a context can be true in the very context it is about.

Semantics for Deflationists (in Deflationism and Paradox, Oxford U.P, 2005: 148-176. This paper spells out the positive theory sketched at the end of "Against Stepping Back".): According to deflationists, [p] is true is in some sense equivalent to p. The problem that the semantic paradoxes pose for the deflationist is to explicate this equivalence without relying on a semantics grounded in the sort of real reference relations that a deflationist thinks do not exist. More generally, the deflationist is challenged to give an account of logical validity that does not force us to countenance such relations. (The usual model-theoretic definition seems to presuppose that there is some special interpretation, the intended interpretation, such that truth simpliciter is truth on that intended interpretation. So if the deflationist adopts this sort of definition, the deflationist will be challenged to identify the intended interpretation without positing real reference relations.) Fortunately, a precise semantics compatible with the deflationist philosophy can be had as follows: First, we define a context as a certain sort of set constructed from a basis of literals (atomic sentences and negations of atomic sentences). This formal account of contexts has to be supplemented with an account of the conditions under which a structure satisfying the formal definition is the structure of that kind pertinent ot a given conversation. For each syntactic type of sentence, we define the conditions under which a sentence of that type is assertible relative to a context. In particular, we define the conditions under which sentences of the form " [p] is true" are assertible in a context, and we define the conditions under which sentences of the form "[p] is assertible in context G" are assertible in a context. Finally, logical validity is defined as preservation of assertibility in a context. It is demonstrated that this approach to semantics resists the semantic paradoxes.

The Belief-Desire Law (Facta Philosophica 7, 2005: 121-144): Many philosophers hold that for various reasons there must be psychological laws governing beliefs and desires. One of the few serious examples that they offer is the belief-desire law, which states, roughly, that ceteris paribus people do what they believe will satisfy their desires. This paper argues that, in fact, there is no such law. In particular, decision theory does not support the contention that there is such a law.

On the Evidence for Prelinguistic Concepts (Theoria (Spanish) 54 (2005): 287-297): Language acquisition is often said to be a process of mapping words into pre-existing concepts. Some researchers regard this theory as an immediate corollary of the assumption that all problem-solving involves the application of concepts. But in light of basic philosophical objections to the theory of language acquisition, that kind of rationale cannot be very persuasive. To have a reason to accept the theory of language acquisition despite the philosophical objections, we ought to have experimental evidence for the existence of concepts in prelinguistic children. One of the few lines of research that attempts to provide such evidence is the work of Paul Quinn, who claims that looking-time results show that four-month old infants form "category representations". This paper argues that Quinn's results have an alternative explanation. A distinction is drawn between conceptual thought and the perception of comparative similarity relations, and it is argued that Quinn's results can be explained in terms of the latter rather than the former.

Attitudes without Psychology (Facta Philosophica 5, 2003: 239-256): Many philosophers hold that beliefs and desires are theoretical entities postulated for the sake of predicting and explaining people's behaviors. This paper offers a very different perspective on the nature of beliefs and desires. According to this, the first step is to understand the nature of assertion and command. Then, to understand the nature of belief and desire, what one must do is extend one's understanding of assertion and commandto assertions and commands on behalf of others; for to attribute a belief is to make an assertion on someone's behalf, and to attribute a desire is to make a command on someone's behalf. From this perspective we can recognize that explanation and prediction are not the primary rationale for attributions of beliefs and desires. A number of foundational issues will also be addressed, having to do with cognition, semantics and ontology.

Truth, Propositions and Context (in Philosophical Dimensions of Logic and Science. Selected Contributed Papers from the 11th International Congress of Logic, Methodology, and Philosophy and Science, Artur Rojszczak, Jacek Cachro and Gabriel Kurczewski, eds., Kluwer, 2003): Whether a sentence is true or not is relative to a context. It is commonly supposed that a sentence is true if and only if the proposition that it expresses in that context is true. Four motives for this analysis are considered and each is shown to be unpersuasive.

Social Externalism and Linguistic Communication (in Meaning, Basic Self-Knowledge and Mind: Essays on Tyler Burge, María-José Frápolli and Esther Romero, eds., CSLI Publications, 2003: 1-33): According to the expressive theory of communication, the primary function of language is to enable speakers to convey the content of their thoughts to hearers. According to Tyler Burge's social externalism, the content of a person's thought is relative to the way words are used in his or her surrounding linguistic community. This paper argues that Burge's social externalism refutes the expressive theory of communication.

Situated Inference versus Conversational Implicature (Noûs 35, June 2001: 163-189): As Grice defined it, a speaker conversationally implicates that p only if the speaker expects the hearer to recognize that the speaker thinks that p. This paper argues that in the sorts of cases that Grice took as paradigmatic examples of conversational implicature there is in fact no need for the hearer to consider what the speaker might thus have in mind. Instead, the hearer might simply make an inference from what the speaker literally says and the situation in which the utterance takes place. In addition, a number of sources of the illusion of conversational implicatures in Grice's sense are identified and diagnosed.

T-Schema Deflationism versus Gödel's First Incompleteness Theorem (Analysis 61, April 2001: 129-136): I define T-schema deflationism as the thesis that a theory of truth for our language can simply take the form of certain instances of Tarski's schema (T). I show that any effective enumeration of these instances will yield as a dividend an effective enumeration of all truths of our language. But that contradicts Gödel's First Incompleteness Theorem. So the instances of (T) constituting the T-Schema deflationist's theory of truth are not effectively enumerable, which casts doubt on the idea that the T-schema deflationist in any sense has a theory of truth. (The argument in section 2 of "Semantics for Deflationists" supercedes this paper.)

Deflationism and Logic (Facta Philosophica 1, 1, October 1999: 167-195): Inference rule deflationism is the thesis that the nature of truth can be explained in terms of the inference rules governing the word "true". This paper argues, first, that, in light of the semantic paradoxes, the inference rule deflationist must reject some of the classical rules of inference. It is argued, secondly, that inference rule deflationism is incompatible with model theoretic approaches to the definition of logical validity. Here the argument focuses on the question whether the number of primitive referring expressions in a natural language is denumerably infinite. Finally, it is argued that these conclusions pertain to T-schema deflationism and Horwich's minimal theory as well. If you would like a copy of this paper, ask me for one. It will be hard to get otherwise.

What is a Context of Utterance? (Philosophical Studies 91, August 1998: 149-172): For many purposes in pragmatics one needs to appeal to a context of utterance conceived as a set of sentences or propositions. The context of utterance in this sense is often defined as the set of assumptions that the speaker supposes he or she shares with the hearer. I argue by stages that this is a mistake. First, if contexts must be defined in terms of shared assumptions, then it would be preferable to define the context as the set of assumptionsthat the interlocutors really do share. Second, not all shared assumptions belong to the context, because not all are relevant. Third, hearers need not accept every member of the context, because some presuppositions are informative. Finally, presupposition coordination problems show that contexts may have contents that even the speaker does not accept. Contexts, we may conclude, are mind-transcendent. In one sense of the term a "presupposition" is an interlocutor's take on this mind-transcendent context.

Universal Instantiation: A Study of the Role of Context in Logic (Erkenntnis 46,March 1997: 185-214): The rule of universal instantiation appears to be subject to counterexamples, although the ruleof existential generalization is not subject to the same doubts. This paper is a survey of ways of responding to this problem, both conservative and revisionist. The conclusion drawn is that logical validity should be defined in terms of assertibility in a context rather than in terms of truth on an interpretation. Contexts are here defined, not in terms of the attitudes of the interlocutors, but in terms of the goals of conversation, and assertibility is explained in terms of cooperation.

Domain of Discourse (Mind 106, January 1997: 1-32): The proposition expressed by an utterance of a quantified sentence depends on a domain of discourse somehow determined by the context. How does the context of utterance determine the conten tof the domain of discourse? Many philosophers would approach this question from the point of view of an expressive theory of linguistic communication, according to which the primary function of language is to enable speakers to convey the propositional contents of their thoughts to hearers. This paper argues that from this point of view there is no persuasive treatment of the determinants of the domain of discourse. The argument focuses onan abnormal case in which the domain the speaker has in mind is not evident to the hearer. In this way the question concerning the determinants of the domain of discourse is used to challenge the expressive theory of communication. See also my "Intelligibility in Semantics: Reply to van Deemter", Mind, 107 (1998): 447-50.

A New Skeptical Solution (Acta Analytica 14, 1995: 113-129): Kripke's puzzle about rule-following is a form of the traditional problem of the nature of linguistic meaning. A skeptical solution explains not what meaning is but the role that talk of meaning plays in the linguistic community. Contrary to what some have claimed, the skeptical approach is not self-refuting. However, Kripke's own skeptical solution is inadequate. He has not adequately explained the conditions under which we are justified in attributing meanings or the utility of the practice of attributing meanings. An alternative skeptical solution may be founded on a nonepistemic conception of assertibility. Roughly, a sentence is assertible if it facilitates cooperation. The function of meaning-talk is to resolve certain sorts of conflicts in assertion. Attributions of meaning to persons outside the community may be a proper expression of a practice whose reason for being lies entirely within the community.

An Extraterrestrial Perspective on Conceptual Development (Mind and Language 8, Spring 1993: 105-130): The network theory of conceptual development is the theory that conceptual developmentmay be represented as a process of constructing a network of linked nodes. The nodes of such a network represent concepts and the links between nodes represent relations between concepts. The structure of such a network is not determined by experience alone but must evolve in accordance with abstraction heuristics, which constrain the varieties of network between which experience must decide. This paper criticizes the network theory on the grounds that current proposals regarding these abstraction heuristics all fail, and further, that, given certain plausible assumptions, no viable account of these abstraction heuristics will be possible. Abstraction heuristics cannot be universal principles of rational thought because virtually no concept is intrinsically unsuitable for use in a true and useful representation of reality. Nor can they be species-specific natural conventions because in that case, it is argued, we would not be able even in principle to learn to understand the language of creatures who used different ones.

The Lockean Theory of Communication (Noûs 26, Fall 1992: 303-324.): The Lockean theory of communication is here defined as the theory that communication takes place when a hearer grasps some sort of mental object, distinct from the speaker's words, that the speaker's words express. This theory contrasts with the view that spoken languages are the very medium of a kind of thought of which overt speech is the most basic form. This article is a critique of some of the most common motives for adopting a Lockean theory of communication. It is not enough that words in some sense express thoughts. It is not enough that animals and prelinguistic infants in some sense think. It is not enough that speakers mean something by what they say or that hearers must understand a speaker's presuppositions. On the contrary, any explanation of how children can learn to communicate in the way the Lockean imagines will presuppose that words can instill beliefs in some way more fundamental than the Lockean theory itself can explain.

Mental Content and the Division of Epistemic Labor (Australasian Journal of Philosophy 69, September 1991: 302-318): Tyler Burge's critique of individualistic conceptions of mental content is well known.This paper employs a novel strategy to defend a strong form of Burge's conclusion. The division of epistemic labor rests on the possibility of language-mediated transactions, such as asking for something in a store and getting it. The paper shows that any individualistic conception of content will render such transactions unintelligible.

Semantics without Reference (Notre Dame Journal of Formal Logic 31, Summer 1990: 437-461): A theory of reference may be either an analysis of reference or merely an account of the correct use of the verb "refer". If we define the validity of arguments in the standard way, in terms of assignments of individuals and sets to the nonlogical vocabulary of the language, then we will be committed to seeking an analysis of reference. Those who prefer a metalinguistic account, therefore, will desire an alternative to standard semantics. One alternative is the Quinean conception of logical validity as essentially a matter of logical form. Another alternative is Leblanc's truth-value semantics. But these prove to be either inadequate for purposes of metatheory or philosophically unsatisfactory. This paper shows how validity (i.e., semantic consequence) may be defined in a way that avoid the problems facing these other alternatives to standard semantics and also permits a metalinguistic account of reference. The validity of arguments is treated as a matter of logical form, but validity for forms is defined on analogy with the definition of semantic consequence in truth-value semantics. (A more radical kind of semantics without reference is the context logical approach represented in several of my other publications.)

How to Learn a Language Like a Chimpanzee (Philosophical Psychology 3, March 1990: 31-53): This paper develops the hypothesis that languages may be learned by means of a kind of cause-effect analysis. This hypothesis is developed through an examination of E. Sue Savage-Rumbaugh's research on the abilities of chimpanzees to learn to use symbols. Savage-Rumbaugh herself tends to conceive of her work as aiming to demonstrate that chimpanzees are able to learn the "referential function" of symbols. Thus the paper begins with a critique of this way of viewing the chimpanzee's achievements. The hypothesis that Savage-Rumbaugh's chimpanzees learn to use symbols by means of cause-effect analysis is then supported through a detailed examination of the tasks they have learned to perform. Next, it is explained how language-learning in humans might be conceptualized along similar lines. The final section attempts to explain how the pertinent cause-effect analysis ought to be conceived. (This paper was published with a reply by Savage-Rumbaugh. See the same issue, pp. 55-76.)