< previous page page_18 next page >

Page 18
happen for the same semiotic reasons for which lexical meanings are graspable and for which a sentence can be both generated and interpreted. The only problem is to insert into semantic compositional spectrums also contextual and circumstantial selections (Theory, 2.11), to add over-coded rules (Theory, 2.14), and to consider within the sememic representation also textual operators (see 0.6.2). The essay on Peircean semantics (Chapter 7) should encourage such a perspective.
In 0.6.2 we shall see that even a textual category such as 'frame' is based upon the model of a sememic representation in terms of case grammar. We shall also see, in 0.7.4, that there is a strong structural similarity between this type of sememic representation and the more abstract structures. Therefore, we can assume that a sememe is in itself an inchoative text whereas a text is an expanded sememe. 9
That is why it is not so realistic to consider the boxes of Figure 0.3 as 'real' steps in text interpretation: they are virtual poles of an interpretative movement which is far and away more continuous and whose timing is rather unpredictable.
Having said this, we can proceed with examining the various codes and subcodes of box 1.
0.6.1.1. Basic Dictionary.
At this sublevel the reader resorts to a lexicon with the format of a basic dictionary and immediately detects the most basic semantic properties of the sememes involved, so as to make a first tentative amalgamation. If the text says that /once upon a time there was a young princess called Snow White. She was very pretty/, the reader detects by a first semantic analysis of «princess» that Snow White is surely a «woman». The sememe «princess» is virtually much more complex (for instance, «woman» entails «human female», and a human female should be represented by many properties such as having certain body organs, and so on). At this point the reader does not know as yet which of these virtual properties must be actualized. This decision will be helped only by further amalgamation and by textual operators. At this sublevel the reader also actualizes the syntactic properties of the lexemes (singular, feminine, noun, and so on) and can begin to establish co-references.
0.6.1.2. Rules of Co-Reference.
On the basis of the first semantic analysis and of the detection of syntactical properties, the reader disambiguates anaphorical and deictic expressions (various shifters). Thus he is able to decide that the /she/ of the text quoted above must be referred to the princess. We shall see in 0.6.2 that one cannot disambiguate most of these co-references without resorting to textual operators. The reader in any case outlines here the first tentative co-textual relations. At this level the reader operates every transformation from surface to deep

 
< previous page page_18 next page >