I explore one apparent source of conflict between our naïve view of grammatical properties and the best available scientific view of grammatical properties. That source is the modal dependence of the range of naïve, or manifest, grammatical properties that is available to a speaker upon the configurations and operations of their internal systems -- that is, upon scientific grammatical properties. Modal dependence underwrites the possibility of conflicting grammatical appearances. In response to that possibility, I outline a compatibilist strategy, according to which the range of grammatical properties accessible to a speaker is dependent upon their cognitive apparatus, but the properties so accessible are also mind-independent.

In this paper my concern is to evaluate a particular answer to the question of how we acquire mastery of the syntax of our first language. According to this answer children learn syntax by means of scientific investigation. Alison Gopnik has recently championed this idea as an extension of what she calls the ‘theory theory’, a well established approach to cognitive development in developntental psychology. I will argue against this extension of the theory theory. The general thrust of my objection is that at the point at which children are acquiring knowledge of syntax they are not in a position to engage in far-reaching scientific investigation. Or, if they are, there are no mechanisms in place to ensure that their scientific investigations will generate a common body of knowledge so making linguistic convergence a mystery. That this is so is a product of two salient features of scientific confirmation. I will conclude that my objections to the theory theory put pressure on learning theories in general.

The article takes up a range of issues concerning knowledge of language in response to recent work of Rey, Smith, Matthews and Devitt. I am broadly sympathetic with the direction of Rey, Smith, and Matthews. While all three are happy with the locution ‘knowledge of language’, in their different ways they all reject the apparent role for a substantive linguistic epistemology in linguistic explanation. I concur but raise some friendly concerns over even a deflationary notion of knowledge of language. Against Devitt I have more serious worries. The latter half of the paper seeks further clarification of Devitt’s realism and raises concerns over its ability to reflect the shape and content of linguistics.

Boghossian claims that we can acquire a priori knowledge by means of a certain form of argument, our grasp of whose premises relies on the existence of implicit definitions. I discuss an objection to his ‘analytic theory of the a priori’. The worry is that in order to employ this kind of argument we must already know its conclusion. Boghossian has responded to this type of objection in recent work, but I argue that his responses are unconvincing. Along the way, I resist Ebert’s reasons for thinking that Boghossian’s argument fails to transmit warrant from its premises to its conclusion.

Chomsky and others have denied the relevance of external linguistic entities, such as E-languages, to linguistic explanation, and have questioned their coherence altogether. I discuss a new approach to understanding the nature of linguistic entities, focusing in particular on making sense of the varieties of kinds of “words” that are employed in linguistic theorizing. This treatment of linguistic entities in general is applied to constructing an understanding of external linguistic entities.

The new Chomskian orthodoxy denies that our linguistic competence gives us knowledge of a language, and that the representations in the language faculty are representations af anything. In reply, I have argued that through their intuitions speaker / hearers, (but not their language faculties) have knowledge of language, though not of any externally existing language. In order to count as knowledge, these intuitions must track linguistic facts represented in the language faculty. I defend this idea against the objections Collins has raised to such an account.

Elaborating on views I have expressed elsewhere, I argue that the common-sense notion of linguistic competence as a kind of knowledge is both required by common-sense explanatory and justificatory practice and furthermore fully compatible with the non-intentional characterization of linguistic competence provided by current linguistic theory, which is itself non-intentional.

It can be assumed that if any part of our mental life is innate it could in principle be developed in private, i.e. is not of necessity a social product. According to the argument in de Sousa (1980), emotions can be subjected to rationality assessments, making them a part, albeit special (borderline), of our ‘rational life’. Contribution of emotions to the conduct of ‘rational life’ is important, as the characteristics of belief and action most commonly associated with rationality do not provide sufficient grounds to guide an organism towards any particular course of action. By asking whether emotions (such that are still subjectable to minimal rationality assessments) can be developed in private (in a sense of Wittgenstein’s Philosophical Investigations) we are asking whether there can be any part of our ‘rational life’ that could be innate, and thus not a result of social conditioning. A brief survey of the related issue of interpretation of Wittgenstein’s arguments against private language and rule-following reveals that the issue is not whether socially non-conditioned emotions could be experienced (exist), but whether we would ever, given the absence of ‘investigation independent standards of correctness’, be able to know that they are or are not.

This paper takes up the two main points in John Collins “Note” (2008b), which responds to my paper, “Explanation and Reality in Linguistics” (2008). (1) Appealing to what grammars actually say, the paper argues that they primarily explain the nature of linguistic expressions. (2) The paper responds to Collins’ criticisms of my view that these expressions have many of their properties by convention.

This note briefly responds to Devitt’s (2008) riposte to Collins’s (2008a) argument that linguistic realism prima facie fails to accommodate unvoiced elements within syntax. It is argued that such elements remain problematic. For it remains unclear how conventions might target the distribution of PRO and how they might explain hierarchical structure that is presupposed by such distribution and which is not witnessed in concrete strings.

Against criticism from Georges Rey I defend both my earlier account of sentence realization and my objection to his own ‘folie-a-deux’ account. The latter has two components, one sceptical (sentences and other standard linguistic entities are rarely if ever realized [‘produced’, ‘tokened’, ‘uttered’]) and the other optimistic (this is a benign outcome since communication is unaffected by our being mistaken in assuming that they are realized). Both components are flawed, notwithstanding Rey’s defence. My non-sceptical account of sentence realization avoids the difficulties his faces, as well as those he raises for it.

According to the “Folieism” I have been recently defending, communication is a kind of folie à deux in which speakers and hearers enjoy a stable and innocuous illusion of producing and hearing standard linguistic entities (“SLE”s) that are seldom if ever actually produced. In the present paper, after summarizing the main points of the view, I defend it against efforts of Barber, Devitt and Miščević to rescue SLEs in terms of social, response-dependent proposals. I argue that their underlying error is a failure to appreciate the important shift of the explanatory locus in modern linguistics, from external objects to internal conceptions. I go on to show how (i) pace Devitt, this shift is entirely compatible with there being conventional aspects to language, and also serves to distinguish the ease of natural language from the waggle dance of the bees; and (ii) pace Barber and Smith, it is compatible with an appearance / reality distinction, and with reliance on testimony in epistemology. I conclude with further arguments about why, pace Collins and Matthews, intentionality is a crucial feature of linguistic explanation, even if it is ultimately spelt out largely in terms of computational role.

This paper defends Some anti-Chomskian themes in Ignorance of Language (Devitt 2006a) from, the criticisms of John Collins (2007, 2008a) and Georges Rey (2008). It argues that there is a linguistic reality external to the mind and that it is theoretically interesting to study it. If there is this reality, we have good reason to think that grammars are more or less true of it. So, the truth of the grammar of a language entails that its rules govern linguistic reality, giving a rich picture of this reality. In contrast, the truth of the grammar does not entail that its rules govern the psychological reality of speakers competent in the language and it alone gives a relatively impoverished picture of that reality. For, all we learn about that reality from the grammar is that it “respects” the rules of the grammar.

The lynch-pin of the structuralist account of logic endorsed by Koslow is the definition of logical and modal operators with respect to implication relations, i.e. relative to implication structures. Logical operators are depicted independently of any possible semantic of syntactic limitations. It turns out that it is possible to define conjunction as well as other logical operators much more generally than it has usually been, and items on which the logical operators may be applied need not be syntactic objects and need not have truth values.In this paper I analyse Koslow’s structuralist theory and point out certain objectionable aspects to as well as reasons why such a theory does not fulfil the (possibly unjustified) expectation of getting defined a universal logical structure.

Many contemporary philosophers contend that a positive answer to Molyneux’s Question -- the question of whether a “man born blind and made to see” would be able to identify spatial figures, without touching them, on first viewing -- requires that there be a *rational connection* between the representations of those figures afforded by vision and by touch. This paper explores the question of what this could mean if the representations are non-discursive, or “pure recognitional” concepts, and argues that the most plausible answer to this question can be invoked to resolve analogous questions about the individuation of phenomenal concepts.

The sense in which having the available belief that P gives one a reason for believing that one believes that P is just that if one has that available belief one is thereby justified, or warranted, in believing that one has it. In explaining why it is so it helps to bring in the notion of rationality. We noted earlier that it is a requirement of full human rationality that one regularly revise one’s belief system in the direction of greater consistency and coherence, and, as a condition of one’s being able to do this, that one have access to its contents and their relations to one another. Judging that one believes something when one does, and judging that one doesn’t believe something when one doesn’t, are manifestations of the satisfaction of this requirement of rationality. That seems a sufficient reason to say that one is warranted in doing so.

Intentionalism holds that two experiences differ in their representational content if and only if they differ in phenomenal character. It is generally held that Intentionalism cannot allow for the possibility of spectrum inversion without systematic error, unless it abandons the idea that, for example, the qualitative character of color experience is inherited from the qualitative character of colors. The paper argues that the conjunction of all three -- the possibility of spectrum inversion, Intentionalism, and the inheritance thesis -- can be consistently, and plausibly, accepted.

In “Reference Borrowing and the Role of Descriptions,” Dunja Jutronić criticizes my view of the borrowing of names and natural kind terms. These terms should be treated, she argues, in the same way as I have tentatively suggested kind terms like ‘sloop’ should be: borrowers need to associate a categorial description that is true of the referent. I am not persuaded. Still, perhaps the suggestion should be extended to these terms anyway. I propose a way to test whether it should that does not rest on intuitions about reference.