The Continuum Companion to the Philosophy of Language

Manuel Garcia-Carpintero, Max Kölbel, Manuel Garcia-Carpintero, Max Kolbel, Max Kölbel, Manuel Garcia-Carpintero, Max Kolbel

• Publisher: Continuum
• Year: 2012
• Print ISBN: 9780826444066
• eBook ISBN: 9781441196385
• Edition: 1
• Topic: Sprachen & Linguistik
• Subtopic: Sprachwissenschaft

The Continuum Companion to Philosophy of Language offers the definitive guide to contemporary philosophy of language. The book covers all the fundamental questions asked by the philosophy of language - areas that have continued to attract interest historically as well as topics that have emerged more recently as active areas of research. Ten specially commissioned essays from an international team of experts reveal where important work continues to be done in the area and, most valuably, the exciting new directions the field is taking. The Companion explores issues pertaining to the nature of language, form semantics, theories of meaning, reference, intensional contexts, context-dependence, pragmatics, the normativity of language, analyticity, a priority and modality. Featuring a series of indispensable research tools, including an A to Z of key terms and concepts, a detailed list of resources and a fully annotated bibliography, this is the essential reference tool for anyone working in the philosophy of language.

1 Editorial Introduction: History of the Philosophy of Language¹

Manuel García-Carpintero²

The Philosophy of Language has a history almost as long as the history of Philosophy itself. Plato’s Cratylus and Sophist, and Aristotle’s De Interpretatione and Prior Analytics, contain important reflections on topics such as the conventionality of language, the subject–predicate structure, valid inference and its relations with the structure of language and thought, truth, or the ontological implications of linguistic categories. Medieval philosophers carried out studies of reference (“suppositio”) and generalization as sophisticated as any. The Port-Royal logicians, Hobbes and Locke took those discussions forward, and, in the latter case, anticipated current concerns about the way natural kind terms work. In the following few pages, however, I will limit myself to drawing a very rough (and rather idiosyncratic) map of the terrain of the contemporary scene, as it was set out in the work of Frege, Russell and the early Wittgenstein – the presupposed common background, taught to beginners in the discipline, for the themes to be further explored from a present-day perspective in the ensuing chapters. In the first part of the chapter, I will outline some core issues as they are presented in what in my view is the insightful systematic articulation of Frege’s and Russell’s themes in Wittgenstein’s Tractatus Logico-Philosophicus. In the second part, I will sum up the main issues, describe some contributions to them in Frege, Russell, Wittgenstein and other historical landmarks, and indicate how they are approached today, as presented in the ensuing chapters. The introduction concludes with a brief discussion of research methods and problems in the field.

Meaning and Modality in the Tractatus

The core issues in the philosophy of language are first put forth with compelling self-conscious depth in Wittgenstein’s Tractatus Logico-Philosophicus, his appraisal of the presuppositions of Frege’s and Russell’s Logicist Program – even if the book would not have been possible without Frege’s and Russell’s ground-breaking research. It is true that, in contrast with Frege’s and Russell’s works, the Tractatus is an opaque piece, whose claims (and even more, the reasons, arguments or at least motivations for them) are difficult to make out, in this respect a reflection of the rather dogmatic methodological attitude of its author. It is also true that such dogmatism appears to have precluded Wittgenstein from seeing the, in some cases glaringly manifest, difficulties for the views he had put forward, and the extent to which the alternative views of his two predecessors, which he had haughtily dismissed, were much more sensible. However, in my view it was in the Tractatus that the proper dimensions and interconnections of the main problems confronted afterwards in the discipline are clearly envisaged for the first time. Neither Frege nor Russell appears to have paid much thought to what has become, since the Tractatus, a core issue in the philosophy of language – the link between grasping the representational contents of thoughts and sentences, and knowledge of modality; or so I will try to suggest in the next few paragraphs. For the most part they aim not mainly to establish this perhaps idiosyncratic historical point, but to sketch out these core problems, so that later we can trace the relations with how they are approached today, as presented in the chapters to follow.

Those core problems in the philosophy of language only perspicuously adumbrated in the early history of analytic philosophy in the Tractatus concern the relations between meaning, modality and our knowledge of them. Frege’s project, which he pursued relentlessly for most of his intellectual life and whose (from his perspective) tragic failure Russell spotted, was the Logicist Program, aimed at proving that arithmetic reduces to pure logic. Frege’s work was hardly a fully-fledged failure: he had come very close to at least reducing arithmetic to logic and set theory, along the lines used later in Russell’s and Whitehead’s Principia or in the independently pursued Cantorian program. In the process, he came up with outstandingly significant conceptual innovations, from modern logic and semantics to an original and influential view in the philosophy of mathematics that many still think fundamentally correct. However, a full appraisal of the epistemological and ontological yields of the project required an examination of the epistemological and ontological status of logic and logical validity themselves; and that in its turn leads to a thorough examination of the nature of the representational devices through which we carry out logically valid inferences: natural language and the thoughts it conveys (what we may call a theory of intentionality). Frege and Russell somehow saw this, and in fact made suggestions about the matter (outlined below) at times more sensible than those in the Tractatus, at times simply incorporated into it. But it is only in that work, I think, that the nature of the problems and their interconnections is systematically realized, through the realization that representation in natural languages and in thought is inextricably tied up with discrimination between possibilities.

Notoriously, the Tractatus contains a flawed theory of intentionality, the so-called “picture theory”; but, more than its failures, what is interesting for our present purposes is to appreciate what it set out to achieve – especially how Wittgenstein hoped that it would deliver what in his view Frege and Russell had failed to provide: a philosophically adequate account of logical validity and hence of the foundations of their logicist project.³ To put it in the metaphor he later used in the Investigations, criticising his earlier views, Wittgenstein’s objection in the Tractatus to the view on the nature of logical validity that Frege and Russell had defended is that it does not account for the “sublimity” of logic: they did not account for the characteristic modal properties of logical truths and validities, and our knowledge thereof, as resulting from essential properties of the representational means in which they are cashed out. It is such an account, according to him, that the picture theory provides.

According to Frege and Russell, logically valid propositions, and inferential transitions among them, are distinguished by their maximal generality; for instance: given that a equals b, and b equals c, we can infer that a equals c, no matter what a, b and c are. According to the Tractatus, however, this is wrong (Tractatus 6.1231). On the one hand, some logical truths are not literally speaking general (if Hesperus is Phosphorus, and Phosphorus is Venus, then Hesperus is Venus is itself a logical truth); on the other, a general truth may well be only accidentally true (we can express in purely general terms the claim that there are infinitely many things, which according to Wittgenstein is not a logical truth). Logical validities are necessary; and they are a priori.⁴ Frege’s and Russell’s proposals do not account for this crucial fact: why should maximal generality entail necessity and apriority? It was the fact that, in his view, the picture theory accounted for it that mainly recommended it in his eyes. The picture theory is relevant to solve the problem because for Wittgenstein logical validities are expressed in natural languages (Tractatus 5.5563) – or the thoughts they convey – whose essential representational properties the picture theory characterizes. Artificial languages, far from being “ideal languages” worth studying in their own right as more adequate to carrying out valid inferences – as Frege and Russell thought – are mere “frictionless planes”; they are useful fictions whose study is a convenient means to exhibit in a simpler way the logical properties of our ordinary assertions and thoughts.

Aside from its motivation as a way of accounting for the modal properties of logical truth and validity, Wittgenstein supported his picture theory of intentionality arguing that only such a theory accounts for two fundamental facts about representation in language and thought. First, we understand linguistic representations and grasp thoughts (at least in paradigm cases, let us say, so as not to prejudge any relevant issue) without knowing whether or not they are correct, whether or not the represented reality is in fact as represented; I will summarize this henceforth with the slogan “representations may fail”. Second (“representations may be new”), we can understand or grasp immediately, without further explanation, representations that we have not encountered before.⁵ How is the picture theory supposed to deal effectively with these explanatory issues? (There will be no point in considering the further issue of whether it really is the only theory that accounts for them.) The picture theory, as I understand it, ascribes to any intentional system, i.e., any system exhibiting the two properties to be explained, two crucial semantic features, which we may describe as an external and an internal one. The external ingredient comprises a lexicon and the correlations of the items in it with independent objects, correlations which Wittgenstein thought of as consisting of implicit ostensive definitions. The internal ingredient is an abstract syntax applying to the items in the lexicon which signifies, by way of what Goodman (1976, 52) calls exemplification,⁶ identical relations between the items correlated with them by the external ingredient. It is the latter feature that makes sentences and thoughts into pictures: the distinguishing feature of pictures is that they represent properties that they themselves exemplify; they represent thanks to the fact that there is a range of properties they literally share with the represented situations.

Let us see how this is supposed to solve the first problem, that representations may fail. The syntax determines a class of well-formed elementary sentences; not just any concatenation of items in the lexicon is acceptable, only some are permitted. Each of them is in that respect a possibility: it is possible to say it, as opposed to abstaining from saying it, independently of the others. Saying is here the lowest common factor of different speech acts – asserting, ordering, conjecturing, requesting, and so on – whose distinguishing differences Wittgenstein thought irrelevant for his concerns. The syntax thus determines a class of maximal “discourses” – allowed combinations of the two designated possibilities for each elementary sentence. Correspondingly, given that the syntax is shared by the lexicon and correlated items, it determines the possibility that the combination of items corresponding to the names in any given elementary sentence (a state of affairs) obtains, and the possibility that it does not obtain. It determines thereby a corresponding logical space of maximal combinations of these two possibilities for each state of affairs; only one of them can be actualized, constituting the actual world. What is required to understand a sentence is to know the interpreted lexicon from which it is built, and its logical syntax; what is thereby known is a possible state of affairs, the class of maximal combinations constituting the logical space compatible with its obtaining, what Wittgenstein calls (following Frege) the sentence’s truth-condition; it is not required to know whether or not this class includes the actual world.⁷

According to this, all (and only) truth-conditions are (contents of) possible sayings, not only those expressed by elementary sentences. Some appropriate set of expressions (the “logical constants”, on the Tractarian account) is needed, to gain the additional expressive potential needed to express all truth-conditions. But the claim made about the explanatory virtue of the picture theory for the case of elementary sentences is intended to apply also to complex sentences including these expressions. Understanding them requires, according to the picture theory, knowing the interpreted lexicon, their logical syntax and the identical “syntax” in the world signified by exemplification, plus the set of logical constants needed in order to express every possible truth-condition thereby determined. This assigns to any non-defective (neither tautologous nor contradictory) sentence a truth-condition, without thereby establishing whether or not it actually obtains. Wittgenstein (Tractatus, 2.1511; cf. Investigations, §§ 95, 194) particularly liked the fact that this little theory accounts for the first problem of intentionality, that representations may fail, while preserving an essential connection between linguistic representations and the world – and thus representations are of real items, not some intermediate ghosts, as in representationalist accounts of perceptual experience. This is achieved in that the represented possible states of affairs are made of real objects, constituting the actual world (all possible worlds, given that all lexical items are on the Tractarian view Kripkean “rigid designators”, designating the same entity with respect to all possible worlds) and of equally real, possibility-determining, “syntactical” relations between them.

Accounting for the second explanatory issue (that representations may be new), assuming the picture theory as presented, is straightforward. Knowing the lexicon, the logical syntax that as we have seen signifies by exemplification, and the relevant set of logical constants suffices for understanding sentences beyond those that one has in fact encountered; in contrast, the meaning of any new lexical item must be explained to us.

Finally, this is how the picture theory is supposed to account for the “sublimity” of logic, the fact that we know a priori necessary truths and relations of necessary truth preservation, to conclude this sketchy outline: “It is the peculiar mark of logical propositions that one can recognize that they are true from the symbol alone, and this fact contains in itself the whole philosophy of logic” (Tractatus, 6.113). If the relations that determine which states of affairs are possible are reflected by identical relations determining which combinations of lexical items are logico-syntactically well formed, we have at the very least the impressionistic beginnings of an explanation. Knowing the facts that determine which possibilities there are, which ones correspond to a given saying, and which ones, expressed by a given saying, are included in the ones expressed by others is already a presupposition of understanding those (or any) sayings. Logical truth is just truth with respect to all possibilities, and logical validity the containment of all the possibilities for the premises in the possibilities for the conclusion. All these matters are determined by the logico-syntactical relations determining well-formedness, signified by exemplification (what I called the “internal” semantic relations). No particular set of “external” semantic relations (no specific lexicon, set of correlations with external objects) must be known for that, although some must; in that respect, the knowledge might be considered a priori.

I have summarily sketched the picture theory of representation that appears to be propounded in the Tractatus, the evidence alleg...

Table of contents

• Cover
• 1 Editorial Introduction: History of the Philosophy of Language
• 2 On the Nature of Language: A Basic Exposition
• 3 Formal Semantics
• 4 Theories of Meaning and Truth Conditions
• 5 Reference
• 6 Intensional Contexts
• 7 Context Dependence
• 8 Pragmatics
• 9 Semantic Normativity and Naturalism
• 10 Analyticity, Apriority, Modality
• 11 New Directions in the Philosophy of Language
• A-Z of Key Terms
• Select Bibliography
• Index