Lexical Semantics and Textual Interpretation

1. The fundamental references here are Grice 1975 and Eco 1979.

2. Grice's Co-operative Principle, Greimas' fiduciary contract, and Eco's co-opera tive activity of the reader are similar concepts that all refer to an extratextual framework of restrictions on inferential processes that the text simultaneously suggests and requires.

3. In saying "semantics of the proper noun" I do not intend to claim that proper nouns have a meaning rather that a direct reference, but only that proper nouns possess some semantic features by virtue of their form, typically, but not only, the feature of male/female. Think, for example, of the connotations associated with certain proper nouns which allow us to make numerous inferences about the regional, social, or cultural origins of the people to which they refer. For a similar position, cf. Lehrer 1992.

4. On the principle of relevance, cf. Grice 1975 and Sperber and Wilson 1986.

5. Such a principle of economy exists at the same level as Grice's maxims or Sperber and Wilson's principle ol relevance; it is independent of the specific content of interpreta­tion and constitutes a general and extratextual cognitive principle.

(1. For a discussion ol the presuppositions ol factive verbs, cf. Kiparskyand Kiparsky 1971-

Notes to Pages 208-1x5

7. In this case, the principle of relevance would operate in the following way: if a text says something there must be a good reason for it; two actions are being talked about, and therefore there must be some relation.

8. They coincide with ptagmatic inferences based on conversational implicature.

9. A particularly illuminating example of this mechanism and how it can be ex­ploited is to be found in the story by Alphonse Allais examined in Eco 1979.

10. Cf. Schank and Abelson 1977.

u. Cf. Jackendoff 1983; Geurts 1985; Heyer 1985; Descles 1987; Kleiber 1988.

12. With natural kinds, typicality relates to a regularity in morphological configura­tion (color, shape, dimensions, etc.); in tetms referring to social situations or institutions like the restaurant, typicality relates to the cultural model and in some cases would be close to what we would call connotations more traditionally.

13. I consider here only the properties that I have defined as essential; clearly bachelor is associated with a richer and more detailed frame that also contains many typical proper­ties that are similar in this case to cultural connotations. Typical properties, however, can by definition be erased; in that they are default values they can always be suspended with­out negating the term.

14. Fillmore (1985) uses the term context free negation for cases like this. Even though this does not seem to me the most apptopriate name because no sentence is ever produced free from a context (with the exception of linguists' examples), one can, however, agree with the sense in which he talks about context free negation (also called "within frame"). For Fillmore, a context free negation is one that is capable of activating its own appropriate context of interpretation, evoked on the basis of terms present within it, in this case bache­lor, without needing a prior context. This is very similar to what I called the "indexation to a standard context" of reference.

15. The whole of Fillmore's argument (1985: 241) is very similar to my own, ex­cept that Fillmore does not develop the argument of the different internal structure of properties.

16. The general approach of these works is that there exist semantic presuppositions, dependent on the semantics of terms, which impose their content in the discourse, accord­ing to a mechanism similar to the one I described in section 8.2.2. For a general review of the problem of presuppositions, cf. Violi 1986; Levinson 1983, chapter IV; Ducrot 1972.

17. In fact, it is not difficult to imagine a context in which new information is con­veyed by presupposed content, as in the sentence We regret to inform you that your article has not been accepted, in which the new information is the factivity of the subordinate clause presupposed by the predicate regret.

18. For a more extensive analysis of this point, cf. Violi 1988; Eco and Violi 1987.

19. Cf. Violi 1986.

20. The background-foreground mechanism is an extremely general mechanism of human cognition, applying also, but not only, to visual perception. Cf. Jackendoff 1983, 1992; Talmy 1978.

10. The Many Dimensions of Meaning

1. If we use the term chair to refer to a tree trunk while we are having a picnic, we activate an abductive process of interpretation that makes possible a semantic overextension of the term; however, if I say Bring me the red coat when there are only black coats, such overextension is impossible.

2. This is the localist hypothesis that I spoke about briefly in chapter 1.

3. In particular, the distinction between an identification enigma and a reference enigma seems problematic and unclear.

,|. CI'. Creimas and Couiu'-s (1979: jji). In Crcimasian semiotics (unlike in Pottier,

Notes to Pages 226-229

for example) semes are metalinguistic elements, exclusively differential, a minimal inven­tory of primitive and universal terms from which to derive all the sememes of a natural language. "Theoretically it is easy to imagine that the combining of twenty-odd semic categoties . . . could produce such a quantity of sememes as to entirely satisfy the needs of a natural language. Semic categories inventoried in this way will without doubt contain the set of universals of language. It is in this sense that one can talk about semes as the minimal units of signification" (Greimas and Courtes 1979: 333; my translation). In these words can be rediscovered the classical project of structural semantics, and an attempt to inventory the ultimate primitives from which, by a process of combination, all elements of content can be derived. We have already discussed the failure of this project. However, the search for homogenous classes of semes, or semantic features, underlying lexical manifestation, can be liberated from these presuppositions and can prove extremely interesting even if the semic inventory cannot constitute the limited inventory that gives rise to the lexicon of a language in its entirety. This path makes it possible to reflect on the existence of basic semantic nuclei, or primitives, and on the deep categories upon which surface lexical con­figurations are organized.

5. Cf. Greimas and Courtes 1979. The main differences in the development of Greimas' theory are terminological: while in Semantique structuralehe is still talking about semiological and semantic levels, this terminology is subsequently abandoned because it is a source of potential equivocation, and replaced with the opposite pairs, figurative/abstract and exteroceptive/interoceptive.

6. Cf. in particular Petitot (1992) who emphasizes the depth of interoceptive or abstract semes, stressing their nature as substantial universals, connected to the roots of the imaginaty. Care needs to be taken, according to Petitot, not to confuse the deep, semio-narrative "semantic" of Greimas' theory with the "semantic" of linguistic theories of the sign. This would lead to a confusion between "an anthropology of the imaginary or a metapsychology with a linguistics 01 semiotics of the sign, which would make the theory as a whole incomprehensible" (Petitot 1992: 37j; my translation).

7. The uncertainties are heightened by the scatcity of precise indications: in the Dictionary of Greimas and Courtes (1979), there is very little concrete exemplification of the two semic levels and for the abstract plane they boil down to concepts of relation/term, object/process.

8. On this point, cf. Bozzi (1990: 97; my translation): "colors, sounds, flavors, etc., as much as movement, shape, weight, the position of objects in space depend on many conditions that it is not at first easy to imagine. ... In our heads we have simulacra of objects, schemata of events, shadows of the qualities of things, stereotyped models of facts and events, which we require in order to talk appropriately and to plan movements and actions. Very efficacious summaries in the formation of which'a minimal principle is ruth­lessly at work (the minimum number of articulations possible, the minimum amount of functional dependency, minimum variability, etc.)."

9. Think, for example, of the meaning of carrot: here color seems to be so important to semantic competence that it can be extended to othet contexts such as hair color in carrot top.

10. Cf. Schreuder, Flores d'Arcais, and Glazenborg 1984; Flores d'Arcais, Schreuder, and Glazenborg 1985; Flores d'Arcais and Schreuder 1987.

11. Pairs of terms may be united by solely perceptual relations, like apple and bull which have their spherical shape in common but not their functional properties or category membership; by solely conceptual relations, like apple and banana, which are perceptually dissimilar bin belong to the same conceptual category; or by both perceptual and concep­tual relations {apple anil orange).

11. Л great quantity of data proves (he existence of differences in the process of acquisition ol the semantic properties o/ referential terms: according (0 .1 hypothesis by

(Ink (1973)1 ''"' ^msl propertlei learned lor this clan of cermi derive from perceptual at