23 In this regard, see Picardi in the introduction to the Italian version of Davidson 1984 and Picardi 1992. Picardi (1992: 253) wonders about the relation between the theories an interpreter must have at his disposal in order to understand a language and the theories the interpreter must construct every time for each individual interlocutor at each stage of the conversation. I don’t think Davidson does anything to solve this problem, and this is precisely because, perhaps for linguistic reasons, he makes no distinction between langue and parole, in other words between the meaning of the terms of a language and the sense of utterances and texts.
24 For this, see the observations in Alac 1997.
25 For this, the reader is referred to the analysis made in Zijno 1996 on the positions of Davidson and Sperber-Wilson. It is clear that none of these authors maintains that there are no linguistic conventions, and that all of us follow determined rules, both for presupposing the interlocutor’s beliefs and for negotiating pertinency and elaborating inferences regarding the communicative situation. Nevertheless, the emphasis is laid on the work for «minimizing disagreement,» letting it be understood that, having a good theory of the speaker, one might do without a theory of language. Yet when it is said that «communicating means trying to modify the cognitive environment of another individual» (Zijno 1996, 2.1.2.) and that a cognitive environment for an individual is the set of facts manifest to him (Sperber and Wilson 1986: 65), this cognitive environment comes to look very like what I call a CT, and in order to presuppose it on the speaker’s part, I too must have a representation in the form of NC. The inference and the contract regard the effort required to make our cognitive environments publicly compatible. This is the case with my Ayers Rock example. It is clear that if someone says to me that Ayers Rock is an animal, I infer from this that his cognitive environment is not only rather unlike mine but also unlike the publicly agreed one. Minimizing disagreement means leading the other to accept at least in part an NC that the Community finds passably acceptable. At most I can extend the principle of charity beyond normal limits, if I am talking with a primitive who really does see Ayers Rock as an animal. But I agree to adapt my cognitive environment to his only for the purposes of communicative interaction, which I feel ought to be safeguarded at all costs. Afterward, I still think that that mountain is not an animal. To put it brusquely, the popular principle whereby one always agrees with the insane does not mean that the Community accepts their viewpoint. It is another matter if the Community turns out to be wrong, and the one we thought a madman was right: history teaches us that this has often happened in the past, and the Community has taken a bit of time to modify what, by social decree, everyone thought was right. In short, therefore, negotiation does not institute a cognitive environment, it takes account of previous cognitive environments, corrects them, and attempts to homogenize them.
26 To say that we negotiate from time to time does not mean to say that stronger and stabler conventions do not gradually sediment. See Dummett 1986: 447–58. A fine contractual view of meaning is to be found in Bruner, one of whose merits is having put the problem of Meaning at the center of the Cognitive Sciences. Not only does he state that culture makes meanings public and shareable (and he knows of the Peircean idea of the public nature of interprétants), but he also maintains that, although all our discourses are ambiguous and polysémie, we are always able to make their meaning public through negotiation (1990: 13).
27 Marconi 1997, 5, also contains reflections on the Schtroumpfs, quoting my article «Schtroumpf und Drang» in Alfabeta, 5 September 1979 (now in Sette anni di desiderio, Milan: Bompiani, 1983: 265–71).
28 What is the cognitive universe of the Schtroumpfs like? Given that they indiscriminately give the name schtroumpf to houses, cats, mice, and bachelors, does this perhaps mean that they do not possess these concepts and cannot distinguish between a cat and a bachelor? Or have they a system of expression (a lexicon, in particular) that is rather poor, but a system of content that is at least as vast and articulated as the experiences allowed by their environment? Or again, since the Schtroumpf language makes it as possible to say Beethoven’s Fifth Schtroumpf as Beethoven’s Schtroumpf Symphony or the Fifth Symphony of Schtroumpf (but never Schtroumpfs Schtroumpf Schtroumpf!), perhaps they have a lexicon as rich as ours and use the all-purpose homonym for reasons of laziness, aphasia, affectation, or secrecy. But does using only one word for many things not lead them to see all things united by some strange relation? If eggs, spades, and mushrooms are all schtroumpfs, do the Schtroumpfs not live in a world where the links between eggs, spades, and mushrooms are far fuzzier than they are in our or in Gargamel’s world? And if this is the case, does this confer upon the Schtroumpfs a deeper and richer contact with the totality of things, or does it make them unable to make a correct analysis of reality, immuring them inside the imprecise universe of their pidgin? These are all questions I feel I cannot answer here, but I have listed them in order to say that Peyo’s stories, even though conceived for children, pose some serious semiotic problems for adults.
Chapter Five / Notes on Referring as Contract
1 For example, the one used in Santambrogio (1992), which deals with reference to «general objects.» Santambrogio intended to study how we can deal in terms of quantification with sentences about general objects. In a truth-functional semantics the problem is of some interest, but I hold that in such a case referring to something becomes a synonym for talking about something. Every time we talk, we talk about something, but, then, I don’t see what specific phenomenon is signified by the term referring.
2 For the difference between semiotic and factual judgments, see A Theory of Semiotics, 3.2. If I said that all platypuses lay eggs, and even if I quantified it, as I did for the properties of mammals, I would certainly not be referring to all the platypuses that exist or have ever existed, because the existence of barren platypuses cannot be excluded. Simply, I would once more be saying that, whatever the animal one wishes to apply the term platypus to, it would have to be an animal that has the property of laying eggs. One might argue that the state of being a mammal is not the same as nursing one’s young: at first sight this would seem to be the case, because the fact that platypuses nurse their young has been proved by various observation sentences, while their status as mammals depends on taxonomic convenience. But since taxonomy registers as mammals those animals attributed with the property of nursing their young, and numerous observation sentences tell us that platypuses nurse their young, we can consider the two sentences as equivalent from our standpoint. Those who make such sentences refer to nothing but contribute to the reconfirmation of the social agreement regarding the MC to be assigned to the corresponding term, in other words, regarding the format of the categorial system assumed within a given conceptual schema.
3 The referential function is not necessarily expressed by the grammatical form. Let’s take a sentence such as Napoleon died on the fifth of May. The sentence would be understood as referential if it was uttered in the same month by a courier arriving in London from Sant’Elena. If a scholar were to say, on the basis of newly discovered documents, that Napoleon did not die on the fifth of May, he would certainly still be referring to Napoleon as an individual, and if he said, All the history books I have studied give erroneous information about Napoleon, he would certainly be referring to all the single history books he had consulted. But if, in answering a question in a history exam, a student were to say Napoleon died on the fifth of May, I would doubt that this was still a sentence with a referential function. The student, completely uninterested in Napoleon, is merely citing an encyclopedic datum to please the teacher. In other words, the student is only trying to show that he knows the cultural convention whereby the notion of Napoleon is associated with the property of having died on the fifth of May 1821, exactly in the same way as if he had replied to the chemistry teacher that water is H2O (where it is very clear that the reference is not to water but to what the current textbooks have to say about the matter). If the student were to say that Napoleon died on the eighteenth of June 1815, the teacher would tell him that he had a poor recollection of what the textbooks say, given that this is recorded as the date of
