Language and Philosophical Problems

Sören Stenlund

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781134952243
• Edition: 1
• Topic: Filosofia
• Subtopic: Storia e teoria della filosofia

Language and Philosophical Problems investigates problems about mind, meaning and mathematics rooted in preconceptions of language. It deals in particular with problems which are connected with our tendency to be misled by certain prevailing views and preconceptions about language. Philosophical claims made by theorists of meaning are scrutinized and shown to be connected with common views about the nature of certain mathematical notions and methods. Drawing in particular on Wittgenstein's ideas, Sren Stenlund demonstrates a strategy for tracing out and resolving conceptual and philosophical problems. By a critical examination of examples from different areas of philosophy, he shows that many problems arise through the transgression of the limits of the use of technical concepts and formal methods. Many prima facie different kinds of problems are shown to have common roots, and should thus be dealt and resolved together. Such an approach is usually prevented by the influence of traditional philosophical terminology and classification. The results of this investigation make it clear that the received ways of subdividing the subject matter of philosophy often conceal the roots of the problem.

1

LANGUAGE, MIND, AND MACHINES

1 INTRODUCTION

There is a prevailing tendency in current philosophy and linguistic theory to forget the difference between a theoretical representation and what it represents, to conflate the rules that govern technical notions and methods in current theorizing about language with the rules of our language as they are manifest in our use of language in ordinary situations. The more established certain technical notions and methods tend to become, the more instinctive this tendency becomes and, as a consequence, questions of the correctness and of the limitations of the technical terms and methods are not raised. The limits of the applicability of the technical terms and methods are taken to be the limits of language. Notions of language connected with technical and formal methods are treated as fundamental.

This tendency is most problematic and seductive when the technical notion consists of a technical use of a familiar word or phrase (such as ‘language’, ‘sentence’, ‘name’, ‘is true’, ‘refers to’, ‘interpretation’, ‘meaning’, etc.) which already has an established non-technical use. Formal similarities between the technical and the non-technical uses—which may have inspired the new notion—may then make it appear as though the technical notion were not a construction, not a new concept, but something inherent in the old notion, something which has now ‘been made explicit’. The rules defining the technical notion are erroneously treated as principles of the (hidden true) nature of the old notion. And this confusion is reinforced by the employment of the same familiar word in these two conceptually different forms of use.

This is how several problematic concepts and methods of contemporary linguistic theory and philosophy have originated in mathematical logic and meta-mathematics. Crucial for this influence of formal logic on linguistic theory was the adoption of the linguistic perspective in meta-mathematics, i.e. the idea of conceiving certain mathematical calculi as formal languages. This point of view stimulated the introduction of linguistic terminology into metamathematics, for instance the use of the word ‘language’ in terms like ‘meta-language’ and ‘object-language’, and terms like ‘expression’, ‘form’, ‘meaning’, ‘interpretation’, ‘translation’, ‘syntax’, ‘semantics’, ‘denote’, ‘use’, ‘mention’, ‘proposition’, ‘assertion’, etc. These terms were, however, given a new, technical use in meta-mathematics, a use which was governed for the most part by the efforts of mathematical logic, and these efforts consisted essentially in the construction and study of various mathematical calculi, such as the predicate calculus, by means of mathematical methods, and employed the ‘idealizations’ characteristic of mathematical work.

As a result of having uncritically brought this technical use of traditional linguistic terminology back into the study of real languages, modern linguistic theory and philosophy of language is dominated by a view of language that one might call the calculus conception, i.e. the view of language in general as a calculus or formal system similar to the systems of formal logic. It is in the context of this view that the new technical uses of linguistic terms appear to be unproblematic and even justified when applied to ordinary languages.¹ The important and difficult thing to understand about this calculus conception is how it manifests itself as a fundamental conception in the use of familiar words from traditional linguistic prose and ordinary discourse, and not so much in the explicit use, which also occurs, of mathematical notation and techniques.

It is therefore also important to distinguish between the calculus conception as a paradigm for a technical, scientific methodology (e.g. for so-called model theory and formal semantics), and the same conception as a fundamental conception of language. As a scientific methodology, as a paradigm for constructing models of various linguistic phenomena, it has its limitations (as have all scientific methods), and its applications have to be justified by their success in solving scientific and technical problems. As such it has indeed been successful, not least within the development of computer science and the construction of techniques for ‘natural language processing’. But this success in the application of thesescientific methods and principles does not justify taking them as general normative principles, and that is what is involved in the attitude towards the calculus conception as a fundamental conception. The rules determining the technical use of words like ‘language’, ‘expression’, ‘sentence’, ‘proposition’, ‘reference’, ‘interpretation’, etc., which were originally adapted to the study and the description of formal systems, are taken without investigation or justification to be appropriate for the description and study of language in general. In this way the scientific methods, notions, and techniques acquire the role of a ‘philosophical methodology’ with the claim to being the appropriate tool for answering philosophical questions on the nature and function of language through the construction of theories of meaning, theories of speech-acts, theories of language learning, and so on. But in this way the conceptual confusions are only increased. For instance, on the basis of the notion of a linguistic expression originating in meta-mathematics it becomes a major problem ‘how purely physical sounds issuing from the mouth of a speaker can mean something, can express thoughts and intentions’, and on the basis of the calculus picture of a language, it becomes something of a mystery, in need of theoretical explanation, how it is possible that people can understand sentences which they have never actually seen or heard before.

The difficulty with philosophical problems of this kind is to resist the temptation to treat them as scientific problems to be dealt with by theory construction and theoretical explanation, because even if such an approach may lead to good and useful technical results (as by-products), it will (for that very reason) conceal the conceptual confusion which is the cause of the philosophical problems. What will solve them is conceptual clarification of the basic notions and not additional constructions and new technical notions developed on the basis of the ones that are not sufficiently understood.

In this book it will be shown how various manifestations of this calculus conception are at the root of several conceptual and philosophical problems of current interest. It will be seen that many of these conceptual problems have roots crossing the current classification boundaries of a variety of philosophical problems. As a consequence, this study will be devoted to related problems in the philosophies of language, mathematics, and mind, and in the philosophical discussion of Artificial Intelligence.

2 THE NOTION OF NATURAL LANGUAGE

What then is involved, more specifically, in the calculus conception as a fundamental conception² of language?

1 It manifests itself in a characteristic use of the term ‘natural language’. Natural languages are conceived as being ‘in principle’ formal languages. (This is explicit, for instance, in Davidson’s and Montague’s writings.)

2 Separation of form and content, expression and meaning. It is supposed to be possible to give (at least in principle) a specification of all the external features of the expressions of a language that are relevant to their meaning without referring to or presupposing the meaning or the use of the expressions in this specification. Considerations of meaning and use may be necessary to finding the specification, to isolating the relevant features of an expression, but once they are found the specification can be stated and understood ‘in abstraction from content and use’. (I call this the external or mechanical notion of the form of an expression and its use, and I shall contrast it below with what I call the logical form of an expression and its use. This notion of logical form is not to be confused with the form of an expression as represented by means of the methods of formal logic, by means of formalization, which is an instance of external form.)

3 The relation between a language and its use in real-life situations is taken to be of the same kind as the relation between a calculus or theory (such as the probability calculus) and its application. Language is exhaustively defined as language through its syntactical and semantical rules. The pragmatic rules, the rules for the use of a language in real situations, are determined on the basis of its syntax and semantics, which are therefore supposed to be conceptually prior to and independent of the pragmatic rules. This could be stated more generally as follows: the logical grammar of the expressions of a language are supposed to be formally specifiable (with the notion of ‘form’ as mechanical form), i.e. the conditions for the use of an expression are supposed to be specifiable uniformly within some systematic framework external to the situations in which the conditions obtain.

4 Molecularity. There is something which constitutes the ‘basic semantical units’ or the ‘molecules’ of language (usually it is the sentences of the language). Meaning is determined uniquely by the notion of a certain form of these molecules. (If e.g. the molecules of language are sentences there is supposed to exist a formally specifiable notion of sentencehood, or of the syntactical form of sentences, on the basis of which meaning is determined.)

5 Features depending on the more specific notion of a calculus as a calculus of functions (like the predicate calculus).

(i)	Compositionality, or more generally: the notions of ‘constituent’, ‘part/whole’, ‘simple/composite’ are used with the functional notation of mathematics as a paradigm. The grammar of these notions is the grammar of the mathematical use of these expressions.
(ii)	‘Determination’ and ‘dependence’ as functional, i.e. the notion ‘…is determined by…’ used in this theorizing on language is conceptually of the kind that is expressed by a mathematical function (even when it is not explicitly expressed in mathematical notation).
(iii)	Logical form is in general represented by means of functions. (This itself is a cause of much conceptual confusion. There is, for instance, no room for internal relationships and logical dependences in contrast to functional relationships. The notions of dependence on situations, context dependence, and so on are construed as functional dependences.)
(iv)	The notion of a rule that is used is the mathematical or algebraic notion (as when it is said that by means of the syntactical rules infinitely many sentences can be generated).
(v)	The notions of ‘finite’, ‘infinite’, ‘sequence’, ‘string’, etc. that are used in this theorizing on language are the ones that we normally employ in connection with mathematical calculi.

More generally, the kind of ‘idealization’ or ‘abstraction from physical and practical circumstances’ that is typical of mathematical work, is taken for granted as appropriate also in the study and the description of the nature of language, which means of course that such physical and practical circumstances are considered inessential to language as language. So to the calculus conception as a fundamental conception there belong definite attitudes to and presuppositions about the essence of language in general.

The term ‘natural language’ is used in opposition to the terms ‘formal language’ and ‘artificial language’, but the important difference is that natural languages are not actually constructed as artificial languages and they do not actually appear as formal languages. But they are considered and studied as though they were formallanguages ‘in principle’. Behind the complex and the seemingly chaotic surface of natural languages there are—according to this way of thinking—rules and principles that determine their constitution and functions; and it is assumed that this hidden structure can be presented as a theory similar to the syntactic and semantic theories of formal systems—with the difference that such a theory will be enormously more complicated, so complicated in fact that we can never hope actually to construct one for more than fragments of a natural language.

What this attitude amounts to in practice, in the practice of constructing theories of language (where it does mean something specific), is the following: there are no conceptual limitations to the applicability of the methods and concepts adopted in the description of formal systems to the study of languages in general. The limitations that exist concern only technical and empirical matters and differences in degree of complexity.

To use the term ‘natural language’ as it is employed in current philosophy of language is essentially to commit oneself to this dogma. And to say in this use that ‘a natural language is not a calculus or formal system’ tends to mean that natural languages do not present themselves as formal systems and they did not originate as artificial languages. To that everybody will of course agree. But the statement will not be understood as questioning the calculus conception as a fundamental conception, precisely because it is taken as fundamental.

As regards the current use of the term ‘natural language’ one could also say that it presupposes a general notion of language under which both natural and artificial languages fall. What then is the common idea of language here? It seems to me that this common idea, in its most general form, is the following: a language is determined (or given) by its vocabulary (its lexicon) and grammar, or—through the influence of meta-mathematics and logical semantics—by its syntax and semantics.

This, however, is an answer to the question: ‘What is the general idea of language in traditional linguistic theory?’, or ‘What, according to traditional linguistic theory, would the common idea of giving a (complete) description of a language amount to?’ The essential and common feature of languages in general is identified with the paradigm of traditional theorizing about language. In much current linguistic theory this takes the form of conceiving a language as ‘being ultimately’ a formal system.

3 CONCEPTUAL INVESTIGATION

If we disregard the use (or family of uses) of the word ‘language’ in traditional and current linguistic theory, and look at languages, not as empirical phenomena to be scientifically explained, but as they present themselves in life, in their use in human affairs in general (i.e. not only in the human activity of theorizing about language), we find features that are much more fundamental. We find, for instance, that language appears as various forms of linguistic communication between human beings. Language is connected to communication and this connection is not just an empirical or historical fact, it is a conceptual connection. The ideas of language and communication are inseparable in a sense which makes it misleading to say that the main purpose or the main function of language is communication between human beings—misleading because it invites the view of language in general as an invention, as something that was constructed (like a nomenclature, a technical terminology or a formal system) for a particular purpose; and it suggests that it would have been possible to have the idea of communication (of thoughts, ideas, information, messages, requests, moods, etc.) without having the idea of language. But our concept of communication extends as far as language.

In recent so-called pragmatics and speech-act theory the importance of the ‘communicative aspect of language’ is in a certain sense recognized, but only as an additional aspect, neglected in traditional linguistic theory. Its conceptual importance is not recognized. The theories of pragmatics attempt to supplement the view of language of traditional grammar and formal logic by accounting for linguistic phenomena where the traditional view deviates from actual linguistic practice. The various features of linguistic communication and of our use of language are considered as empirical phenomena to be explained theoretically and accounted for on the basis of a picture of a language as being fundamentally a complicated (formal) system. The use of linguistic expressions in communication is conceived as being the (‘tacit’) application of a system of specifiable rules and prin...

Table of contents

• Front Cover
• Half Title
• Title Page
• Copyright
• CONTENTS
• Preface
• 1 LANGUAGE, MIND, AND MACHINES
• 2 NOTIONS OF LANGUAGE AND THEORIES OF MEANING
• 3 FORM AND CONTENT IN MATHEMATICS
• Notes
• Bibliography
• Index