4 Why this was never said or admitted is explained very well by Ducrot. For Saussure and the structuralist school, signifieds were purely differential and were not defined by their content. The meaning of a sign registered only the features that distinguished it from the other signs in a language and not a description of its possible referents: To return to the Aristotelian example, the meaning of homme does not involve the feature «featherless,» because as a matter of fact the natural classification incorporated in French does not oppose homme and oiseau within a category bipède, but homme and animal within a category être animé (Ducrot and Schefer 1995: 303).
5 I must admit that, in previous works, I gave rise to the misunderstanding that semiotics should not take any interest in processes of referring, and that it was possible to deal with both the problem of identifying the referent and the problem of acts of reference in a unitary fashion. But my polemic was due to the fact that in those works I wanted to emphasize how culture constituted a system of content, and how discourses produced an effect of truth, so it seemed less important to establish what individual or state of affairs was being referred to by saying that Dion runs. Naturally no one thought that language was not used for referring to something. The problem hinged on seeing reference as a function of meaning and not vice versa. The second half of A Theory of Semiotics treats of what happens when we express indexical judgments and of how, in referring to objects, perceptual data are compared with cultural data; while the chapters on sign production hinge largely on the tasks involved in the interpretation of symptoms, prints, clues, and toposensistive vectors in order to learn something about that which is the case, and about how we construct or take as signs examples, samples, or projections to refer, indicate, designate, or portray objects of the world. In conclusion, my preoccupation with abduction, and with regard not only to general laws but also to facts—as happens in the investigations of Sherlock Holmes (Eco and Sebeok 1983)—means that I was interested in the mental mechanisms by means of which we come to say something true or at least likely with reference to specific individuals and events. I am grateful to Augusto Ponzio (1993: 89) for having observed that in A Theory of Semiotics I moved from an apparently «antireferential» semiotics to a «non immediately referential semiotics.» In other words, if at first it seemed as if I was stating that semiotics has nothing to do with our relation with reality, in the second phase I was saying that it is not possible to explain how we refer to reality if we do not first establish how we give meaning to the terms we use.
6 I am condensing an argument put forward by Bonomi (1994, 4). In 1934, Carlo Emilio Gadda wrote an article, «Morning in the Abattoir.» This was a description of the town abattoir in Milan. In Bonomi’s mental experiment, the article remained unpublished, it never mentioned the city of Milan, and a researcher had found a manuscript copy among Gadda’s papers, taking it for a piece of narrative fiction. If the researcher later discovered that the text was a newspaper piece, to be judged in terms of true/false, even though he changed his opinion about the nature of that text, he would have no need to reread it. The world described, the individuals who inhabit it, and their properties would all still be the same, and the researcher would then simply «project » that representation onto reality. Therefore «for the content of an account that describes a certain state of things to be understood, it is not necessary for the categories of true and false to be applied to that content.»
7 For my objections with regard to rigid designation, see Eco 1984, 2.6.
8 In an a technical way, Campanini proceeds much like the pseudointelligent computer involved in project Eliza. In this experiment the computer, which obviously does not understand what its human interlocutors say, is instructed to take the subject of an interlocutor’s sentence and use it for the construction of a question that seems intelligent. If the interlocutor confesses to having parental problems, then the computer responds by saying, «Tell me a bit about your parents.» In the sarkiapone sketch, Campanini restricts himself to grasping the name of the property supposed by Chiari and in essence replies, «But the sarkiapone does not possess the property you have mentioned.»
9 With regard to this type of existential presupposition, see Eco and Violi 1987 and Eco 1990, 4.4.
10 On this point we could dust off the old medieval argument as to whether existence is an accident of the essence (attributed to Avicenna) or an act of the essence (Aquinas). There is no doubt that it is necessary to distinguish between the predicative and existential use of the essence (for a crystal-clear synthesis, see Piattelli Palmarini 1995, 11). This makes it possible to clarify a point made in 2.8.3 in reference to the interpretation offered by Fumagalli (1995), who says that while in Peirce’s early work the three categories of Firstness, Secondness, and Thirdness were elements of the proposition, in his later writings they were moments in experience. As a moment of experience, existence is not a predicate, it is the clash with something that stands against me, before me, anterior to all conceptual elaboration—and it is this immediate sensation of being that I was talking about in 1.3. It is a prepredicative existence. But when, on the other hand, I state that Paris has the property of existing in this world while Calvino’s Invisible Cities do not have it, I have moved on to existence as predicate.
11 For the paradoxes of rigid designation in scientific contexts, see Dalla Chiara and Toraldo di Francia 1985.
12 One can object that the domain label (uk., or fr., or it.) can be a description, insofar as it defines at least the baptismal area. If it were true, the description would be too vague and inscrutable. But a British citizen is free to subscribe to a French site: therefore the domain loses any descriptive value it may have had.
13 One might say that no one is really interested in the solution. But if the Jekyll-Mary relationship were to result in the birth of a son, Charles, who at the age of twenty discovers there are two Jekylls, then he might be seriously interested in knowing who his biological father is. But owing to the daily sexual relations between Jekyll and Mary, it would be impossible to determine the day on which Charles was conceived, so we would have someone who knows that his father was certainly one of the two Hyde brothers, but no amount of trying to find out would tell him which of the two it was.
14 Like Putnam (1992), Holmes thinks that the Thing-in-Itself is not so much an unknowable by definition as an ideal limit of knowledge. Therefore I agree also with F0llesdal (1997: 453): the rigidity of the designation is a regulative idea; in the Kantian sense of the term, a normative notion.
15 Among the infinite contributions to the debate on the murder of Smith, I will mention only the three texts I had in mind as I was writing this paragraph: Bonomi (1975: 4), Santambrogio (1992), and, above all, Berselli (1995 1.3).
16 In the original version, Nancy wanted to marry a Norwegian. To the best of my knowledge the example appears in McCawley 1971, but perhaps it was already in circulation. Interesting suggestions on Nancy’s case were sent to me (in manuscript form) by Franz Guenthner in the course of a debate at the Centro di Semiotica in Urbino in the seventies.
17 I am always irritated when I receive a postcard, let’s say, from Bali with «Greetings, John.» Which John? Is it possible that this John does not know there are many other people in the world with the same name as his, and that I know at least a score of them? Can he possibly think he is the only John I know? I am citing a very common case. This means that people think of their names in terms of rigid designation. But the fact that people make this mistake (or fall victim to this weakness) is no reason for philosophers to make it too.
18 As was pointed out in note 16, I took the example of Nancy from McCawley. Are we dealing with the same Nancy? It is true that his Nancy wanted to marry a Norwegian and mine wanted an analytic philosopher, but the same Nancy might be very willful. Or she might harbor the strange notion that all analytic philosophers are Norwegian. When we talk about my Nancy, are we talking about McCawley’s Nancy? As can be seen, negotiating a reference is a very complex operation.
19 Inconceivable worlds (in narrative and the figurative arts) are an example of impossibilia, i.e., worlds that the reader is led to conceive only as far as it
