Abstract
This chapter relates our approach to semantics and pragmatics described in the previous chapter to traditional theories of philosophical logic and model theory. This discussion is relevant for two reasons. Logical semantics in general, and predicate calculus in particular, have become the main paradigm for analyzing meaning in natural language. At the same time, predicate calculus plays an important role in Artificial Intelligence, including automatic theorem proving, rule based deductions, and planning. Therefore, this chapter may be read as an investigation of the ontological assumptions of model-theoretic semantics, and as a study of the “epistemological problems” of Artificial Intelligence (in the sense of McCarthy (1977)).
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
For an alternative solution see Hausser (1979b), pp. 39ff and Section 7.
The attempt to define meanings solely in terms of system internal oppositions originated with Saussure. This structuralist goal was carried on by Carnap in Der Logische Aufbau der Welt. Goodman’s (1951) critique of this work showed conclusively that a meaning analysis based solely on oppositions is not capabable of unique identification.
Hausser (1981).
See Hausser (1979a,b, 1981) for further discussion.
See Putnam (1975), chapter 12, who points out that the notion of an intension as a concept, on the one hand, and the notion of an intension as an extension determining function, on the other, are two different notions which are incompatible. The reason, according to Putnam, is that concepts are something mental and thus in the head of the speaker-hearer. Since Putnam chooses to define meaning as an extension-determining function (cf. op.cit, p. 270), he is led to the counterintuitive conclusion that “meanings just ain’t in the head” (op.cit., p. 227).
Another aspect supporting the classification of Carnap’s intensions as -sense (or extensional) is the fact that one may define a “strictly intensional logic” where expressions always denote intensions and are therefore not ambiguous between a “sense” and a “referent.” Such a system, called IL1, was presented in Hausser (1979b, 1984a) and is described by Peregrin and Sgall (1987) as follows: “…the recursion of the logical rules operates uniformly on the level of intensions; consequently there is no operator ‘’ in IL1 and no basic emphasis on the concept of extension. It is worth noting that if we accept the definition of the concept of intensional logic as given, for example, by T.M.V. Janssen, then IL1 is not intensional, since the rule of intersubstitutivity of identicals holds without restriction.” A presuppositional variant of IL1 is defined in 15.5.1.
Barwise and Perry (1983). See Pollard and Sag (1987, pp. 1–6) for a description. An “idealistic,” <+constr,+sense>, reinterpretation of situation semantics is proposed in Ishimoto (1987). Such a reinterpretation, however, runs counter to the most fundamental motivation of situation semantics, which is an attempt to maintain (a version of) mathematical realism in the analysis natural language semantics: situation semantics adopts an extremely wide notion of external reality in order to avoid description of the speaker-hearer internal information processing. Instead, the external platonic structures postulated by situation semantics are supposed to “classify” internal states (see Barwise and Perry (1983), p. 226).
Kamp (1981a).
See the description of “translating an assertion into a mental model,” Johnson-Laird (1983), pp. 249f.
This point is emphasized by McDermott (1976), who notes that the possible referents of an expression cannot be fixed in advance. “The uses of reference in discourse are not the same as those of naming in internal representation…In discourse, a speaker will introduce a hand and easily refer to ‘the finger’. Frame theorists and other notation-developers find it marvelous that their system practically gives them ‘the finger’ automatically as a piece of the data structure ‘at hand’. As far as I can see, doing this automatically is the worst way of doing it. First, of course, there are four or five fingers, each with its own name, so ‘the finger’ will be ambiguous. Second, a phrase like ‘the finger’ can be used in so many ways that an automatic evaluation to Finger 109 will be wasteful at best.…It seems much smarter to put knowledge about translation from natural language to internal representation in the natural-language processor, not in the internal representation.”(McDermott, (1976), pp. 149f.)
Hausser (1979b, c, 1981, 1983a,b, 1984a,b).
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1989 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Hausser, R. (1989). Meaning, Truth and Ontology. In: Computation of Language. Symbolic Computation. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-74564-5_13
Download citation
DOI: https://doi.org/10.1007/978-3-642-74564-5_13
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-74566-9
Online ISBN: 978-3-642-74564-5
eBook Packages: Springer Book Archive