Online research in philosophy

This paper continues a strain of intellectual complaint against the presumptions of certain kinds of formal semantics (the qualification is important) and their bad effects on those areas of artificial intelligence concerned with machine understanding of human language. After some discussion of the use of the term epistemology in artificial intelligence, the paper takes as a case study the various positions held by McDermott on these issues and concludes, reluctantly, that, although he has reversed himself on the issue, there was no time at which he was right.

Formal semantics is an approach to SEMANTICS1, the study of meaning, with roots in logic, the philosophy of language, and linguistics, and since the 1980’s a core area of linguistic theory. Characteristics of formal semantics to be treated in this article include the following: Formal semanticists treat meaning as mind-independent (though abstract), contrasting with the view of meanings as concepts “in the head” (see I-LANGUAGE AND E-LANGUAGE and MEANING EXTERNALISM AND INTERNALISM); formal semanticists distinguish semantics from knowledge of semantics (Lewis 1975, Cresswell 1978), which has consequences for the notion of semantic COMPETENCE. A central part of the meaning of a sentence on this approach is its TRUTH CONDITIONS, and most although not all formal semantics is model-theoretic, relating linguistic expressions to model-theoretically constructed semantic values cast in terms of truth, REFERENCE, and possible worlds. This sets formal semantics apart from approaches which view semantics as relating a sentence just to a representation on another linguistic “level” (LOGICAL FORM) or a representation in an innate LANGUAGE OF THOUGHT. The formal semanticist could accept such representations as an aspect of semantics but would insist on asking what the model-theoretic semantic interpretation of the given representationlanguage is (Lewis 1970). Formal semantics is centrally concerned with COMPOSITIONALITY at the SYNTAX-SEMANTICS INTERFACE, how the meanings of larger constituents are built up from the meanings of their parts on the basis of their syntactic structure, and with the relation between compositional SENTENCE MEANING and meaning in discourse.

This essay considers what it means to understand natural language and whether a computer running an artificial-intelligence program designed to understand natural language does in fact do so. It is argued that a certain kind of semantics is needed to understand natural language, that this kind of semantics is mere symbol manipulation (i.e., syntax), and that, hence, it is available to AI systems. Recent arguments by Searle and Dretske to the effect that computers cannot understand natural language are discussed, and a prototype natural-language-understanding system is presented as an illustration.

Some twenty years ago, semanticists of natural language came to be overwhelmed by the problem of semantic analysis of belief sentences (and sentences reporting other kinds of propositional attitudes): the trouble was that sentences of the shapes X believes that A and X believes that B appeared to be able to have different truth values even in cases when A and B shared the same intension, i.e. were, from the viewpoint of intensional semantics, synonymous 1 . Thus, taking intensional semantics for granted, belief sentences appeared to violate the principle of intersubstitutivity of synonyms. The verdict of the gurus of intensional semantics was that hence intensional semantics is inadequate, or at least insufficient for the purposes of analysis of propositional attitudes; and that we need a kind of a ‘hyperintensional semantics’.

1. Formal semantics in linguistics -- 2. Generalized quantifier theory -- 3. The interface between syntax and semantics -- 4. Anaphora, discourse, and modality -- 5. Focus, presupposition, and negation -- 6. Tense -- 7. Questions -- 8. Plurals -- 9. Computational semantics -- 10. Lexical semantics -- 11. Semantics and related domains.

How can computers distinguish the coherent from the unintelligible, recognize new information in a sentence, or draw inferences from a natural language passage? Computational semantics is an exciting new field that seeks answers to these questions, and this volume is the first textbook wholly devoted to this growing subdiscipline. The book explains the underlying theoretical issues and fundamental techniques for computing semantic representations for fragments of natural language. This volume will be an essential text for computer scientists, linguists, and anyone interested in the development of computational semantics.

A primary problem in the area of natural language processing has been semantic analysis. This book looks at the semantics of natural languages in context. It presents an approach to the computational processing of English text that combines current theories of knowledge representation and reasoning in Artificial Intelligence with the latest linguistic views of lexical semantics. The book will interest postgraduates and researchers in computational linguistics as well as industrial research groups specializing in natural language processing.