The Stoic model of sign assumes, therefore, the form of the inference (p q), where the variables are neither physical realities nor events, but the propositions that express the events. A column of smoke is not a sign unless the interpreter sees the event as the true antecedent of a hypothetical reasoning (if there is smoke . . .) which is related by infer-ence (more or less necessary) to its consequent (. . . then there is fire). This is why the Stoics can say, and they do, that the sign is a lekton and, therefore, an incorporeal. The sign is not concerned with that smoke and that fire, but with the possibility of a relationship between antecedent and consequent regulating of any occurrence of the smoke (and of the fire). The sign is type, not occurrence.
By now it is clear how, in the Stoics’ semiotics, the theory of language becomes rightfully associated with the theory of signs. In order to have signs, propositions must be formulated, and the propositions must be organized according to a logical syntax which is reflected and made pos-sible by the linguistic syntax (see Frede 1978). Signs emerge only insofar as they are rationally expressible through the elements of language. Language is articulated inasmuch as it expresses meaningful events. It must be stressed that the Stoics do not still say that words are signs (at most they say that words serve as vehicles for types of signs).
The lexical difference between the couple semainon/semainomenon and the simeion remains. But the common and obvious etymological root is an indication of their relatedness. We could have the Stoics say, as Lotman does, that language is a primary modeling system, through which the other systems are expressed.
Referring once again to contemporary theories (see also Todorov 1977), we could then say that the linguistic term and the natural sign are constituted by a double relation of signification, a double elevation trans-latable into the Hjelmslevian model of connotation (in the diagram form popularized by Barthes; see Figure 1.1).
The word /smoke/ refers to a portion of content segmentation which we will conventionally designate as «smoke». At this point, we have three alternatives, whether intensional or extensional: (a) «smoke» con-notes «fire» on the basis of an encyclopedia-like representation which takes into account metonymic relationships of effect-to-cause (a case grammar accounting for ‘actants’ like Cause or Agent can represent rather well this sort of meaning postulate); (b) the sentence /there is smoke/ expresses as its content the proposition «there is smoke» which, always by virtue of an underlying encyclopedic representation including frames or scripts (see 3.2 of this book), suggests as a reasonable infer-ence «there is fire» (notice that we are still at an intensional level, since the possibility of the inference is coded among the properties of smoke, independently of any actual world experience); (c) in a process of refer-ence to states of the actual world the proposition «there is smoke», on the grounds of the aforementioned meaning postulates, leads to the in-dexical proposition «therefore here there is fire», to be evaluated in terms of truth values.
When I perceive a cloud or a column of smoke as mere physical events, they do not differ from any sound which I can perceive without attributing a semantic relevance to it (as the barbarian does). But if, on the basis of a preexisting rule, I know that smoke in general refers to fire, then I make the event pertinent as a single expression of a more general content, and the smoke I perceived becomes the perceptive con-tent «smoke». This first movement, from the sensation to the percep-tion invested with meaning, is so immediate that we tend to consider it as semiotically irrelevant. Gnoseology has always questioned precisely this presumed immediacy of sensation and perception.
Ε C
ΕC
FIGURE1.1
Even according to the medieval perspective, simplex apprehensio, that is the first operation of the intellect, allows one to grasp the thing in its essence through the phantasm, but it is only by the act of judgment, that is, the second operation of the intellect, that the thing is recognized and viewed as relevant for the purpose of further predications. It is not by chance that gnoseology talks of the perceptive meaning where the term ‘meaning’ seems at once a semantic category and a category of the phenomenology of perception. Actually, in order to grasp, from a series of sensory data, the form ‘smoke’, I must already be directed by the belief that smoke is relevant to the making of further inferences. Otherwise, the smoke provided for me by the sensation remains a potential perception which I have not yet make pertinent as smoke, but as mist, miasma, or as any exhalation which is not caused by combus-tion. Only if I already know the general rule which makes for ‘if smoke, then fire’ am I able to render the sensory datum meaningful, by seeing it as that smoke which can reveal fire.
1.8. Unification of the theories and predominance of linguistics
Some centuries later, in De Magjistro, Augustine will definitely bring to-gether the theory of signs and the theory of language. Fifteen centuries before Saussure, he will be the one to recognize the genus of signs, of which linguistic signs are a species, such as insignias, gestures, ostensive signs. But in so doing Augustine delivers to the tradition that follows him a problem that not even the Stoics had clearly solved. Augustine had actually provided a solution, but he had failed to stress it sufficiently to make it indisputable. The Stoics had left unresolved the problem of the difference between the relation of linguistics expression to content on the one hand (what Hjelmslev will call denotation) and the relation of sign-proposition to consequent meaning on the other. One suspects that the first level may still be based on equivalence, while the second is doubtlessly based on inference (Figure 1.2).
However, we must ask whether or not this difference is based on a curious ‘optical illusion’. From the moment in which Augustine intro-duces verbal language among signs, language starts to appear in an awk-ward position. Being too strong, too finely articulated and therefore scientifically
Ε C
Ε C
FIGURE1.2
analyzable (and the work of the Hellenistic grammarians must be kept in mind in this respect), language could hardly be the object of a theory of signs born in order to describe the relationship between natural events, so elusive and generic (and we will see how much the Stoics’ inference was epistemologically open to a continuum of relationships of necessity and weakness). Since language was increasingly believed to be the semiotic system which could be analyzed with the most profit (a careful study of this aspect of the history of semiotics would be very worthwhile) and the system which could serve as a model for all other systems (translating every other semiotic onto the plane of its content), the model of the linguistic sign gradually came to be seen as the semiotic model par excellence.
By the time this conclusion was reached (the definitive sanction took place with Saussure), the linguistic model was crystallized into its ‘flat-test’ form, the one encouraged by the dictionaries and, unfortunately, by a lot of formal logic which had to fill its empty symbols only for the sake of exemplification as well. As a consequence, the notion of meaning as synonymy and as essential definition began to develop.
/Man/ is equivalent to «rational animal» in certain contexts, but cer-tainly not in the expression /mom, there is a man with a package to deliver/, where the content «man» can be analyzed according to many properties (male, unknown, human being, person of a low social extrac-tion, even foreign presence or threat), but certainly not as a rational animal. Aristotle delivered to us the principle of (biconditional) equiva-lence between term and definition by genus and species because he worked only on categorematic terms to be inserted within assertive propo-sitions. The Stoics, on the other hand (see Frede 1978; Graeser 1978), thought that every syntactic category had its semantic counterpart, including syncategorematic terms. If the complete lekta derived from the combina-tion of incomplete ones, they had also to include conjunctions, articles, and pronouns. Augustine later shows that even prepositions have mean-ing.
1.9. The ‘instructional’ model
In De Magistro 2, Augustine analyzes with Adeodatus the verse by Virgil «Si nihil ex tanta superis placet urbi relinqui,» and defines the eight words as «octo signa.» He then proceeds to analyze the meaning of /si/ and to point out that this term conveys a meaning of «doubt». And since he knows that поп esse signum nisi aliquid significet, he is forced to define the meaning (certainly not the referent!) of /nihil/. Granting that it is im-possible to produce signs which do not say anything, and since the meaning of /nothing/ does not seem to be either an object or a state of the world, Augustine concludes that this term expresses an affection of the sou/, that is, the state of mind which, although not recognizing something, recognizes at least its absence.
Augustine then goes on to ask what /ex/ means. He refuses to accept the synonymical answer, according to which /ex/ would mean «de». This synonym is an interpretation that must in turn be interpreted. The con-clusion is