(iii) Citizen Kane, who dreams of building the perfect residence, finds it in Europe in the abbey of Cognac, which has remained intact since the time of its construction. He buys it, has it dismantled and the stones numbered before having it shipped to Xanadu and reconstructed. Is this the same abbey? He certainly thinks so, but certain supercilious European critics and historians think otherwise. Their preference is not for identity of materials or form but for homolocality. Would they therefore be obliged to say that Port Royal (which no longer exists) is more original than Cognac (which fundamentally exists, even though it is in the wrong place)?
(iv) The buildings in the Valley of Kings in Egypt risked being flooded when it was planned to build a new dam. UNESCO had those buildings dismantled stone by stone and then reconstructed in another valley. Are these the same buildings? UNESCO’s assumption is that they are, what counts is the form and the identity of the materials, but those who contested the authenticity of Kane’s reconstruction would have to disagree. Why should cases (iii) and (iv) be different? Why do we think that UNESCO has the moral and scientific right to do what Kane did arbitrarily and out of personal interest?
(v) The Parthenon in Nashville (Tennessee) was designed to resemble the formal structure of the original Parthenon in all respects, so much so that there was a rumor (how veracious, I do not know) to the effect that, after the last war, in order to restore parts of the Parthenon in Athens, experts went to gather evidence on the basis of the Nashville Parthenon. Furthermore, the Nashville Parthenon is painted, as the original edifice is supposed to have been. Yet no one would dare to consider it original, even though the form is the same, simply because the stones are not the same ones, because it does not stand in the same place (among other things, it stands on a plain and not on an acropolis), and, above all, because the other one is still there.
(vi) Poland (as a political entity) has been one of the most tormented nations in history: all one has to do is glance at a historical atlas to see how its frontiers have expanded and contracted depending on the period; indeed, at a certain point it all but disappeared from the map. What does the name Poland refer to? It depends on the historical context in which it is used. Is the sentence Bialystok belongs to Poland true or false? It depends on when it was said.22
On Ahab’s other Leg
In the light of a contractual theory of reference, I think it may also be possible to solve the knotty problem of referring to fictitious characters, such as Sherlock Holmes or Pinocchio. If a strong ontological version of reference (from the standpoint of the eyes of a Divine Mind) is asserted, then all the arguments that have filled tens and hundreds of books can apply to fictitious characters.23 If a weak ontological version (internal realism, reference in the eyes of a Community) is accepted, the discourse seems less dramatic, because we would refer to Hamlet every time we assumed we were dealing with the character described in Shakespeare’s possible world and regarding whom all encyclopedias recognize some properties (though not others), just as the encyclopedias agree in saying that water is H2O.
The interesting problem is not whether fictional characters exist in the same way as real people: in that case, the answer is «no,» not even if one were to accept the realism of Lewis (1973: 85), for whom possible worlds are just as real as the one in which we live from day to day. The interesting problem is why we can refer to them in the same way in which we refer to real people, and we understand one another perfectly well both when we say that Napoleon was the husband of Josephine and when we say that Ulysses was the husband of Penelope. This happens because the encyclopedias agree in assigning to Josephine the property of having married Napoleon as her second husband and to Penelope that of having married Ulysses.
It has been said that narrative worlds are always little worlds, because they do not constitute a maximal and complete state of things (see Pavel 1986; Dolezel 1989: 233 ff.; Eco 1990, 4). In this sense narrative worlds are parasitical, because, if the alternative properties are not specified, we take for granted the properties that hold good in the real world. In Moby-Dick it is not expressly stated that all the sailors aboard the Pequod have two legs, but the reader ought to take it as implicit, given that the sailors are human beings. On the other hand, the account takes care to inform us that Ahab had only one leg, but, as far as I remember, it does not say which, leaving us free to use our imagination, because such a specification has no bearing on the story.
Once we have accepted the commitment to read a story, we are not only authorized but also invited—if we so wish—to make inferences both on the basis of events narrated and on those presupposed. In principle we could do the same thing with a sentence that refers to events that really happened and with a sentence from fiction. Given Julius Caesar was assassinated in the Senate, in Rome, on the Ides of March, 44 B.C., we can infer in what year ab urbe condita the event took place (but we have to decide if it refers to the dating of Cato the Elder or to that of Varro). Given D’Artagnan arrived in the city of Meung, on a sorrel nag at least fourteen years old, on the first Monday of the month of April 1625, by consulting a universal calendar one might conclude that the first Monday of that April was the seventh.
But while it is of some interest to know in what year ab urbe condita Caesar died, it is not narratively of interest to know that d’Artagnan arrived in Meung on the seventh of April. It is of interest to establish that Hamlet was a bachelor, because the observation has some bearing on an understanding of his psychology and of the business with Ophelia. But when, at the end of chapter 35 of The Red and the Black, Stendhal, in recounting how Julien Sorel tries to kill Madame de Rénal, concludes, «Il tira sur elle un coup de pisto-let et la manqua; il tira un second coup, elle tomba.» Is there any sense in wondering where the first bullet ended up?
As has already been stated in Eco (1979), fictional characters have different types of properties.
(i) We have, first and foremost, those properties that are not made explicit by the text but must be presupposed in the sense that they cannot be denied: a character’s hair may not be described, but this is no reason for the reader to presume he is bald. The extent to which such properties may not be denied is seen in the processes of intersemiotic translation: if in a film version of the tale, Julien Sorel went off try to commit murder without his shoes (not mentioned in the story), the matter might appear curious.
(ii) Then there are those properties that in Eco 79 are called S-necessary (or structurally necessary), such as the property of maintaining, within the possible narrative world, reciprocally defining relationships with other characters. In the narrative world of Madame Bovary, there is no other way to identify Emma if not as the wife of Charles, who in his turn has been identified as the boy seen by the narrator at the beginning of the novel; any other narrative world in which Madame Bovary were to be the wife of Monsieur Homais would be another world, with its complement of different individuals (in other words, we would no longer be talking of Flaubert’s novel but of a parody or remake of it).
(iii) The properties explicitly attributed to the characters in the course of the story are seen as particularly evident, such as having done this or that thing, being male or female, young or old. They do not all have the same narrative value: some have an important bearing on the story (e.g., the fact that Julien shot Madame de Rénal), others less (the fact that he fired as the lady was at prayer with her head lowered, and that he fired two shots instead of one only). We can make a distinction between essential and accidental properties.
(iv) Finally there are the properties that the reader infers from the story, which are sometimes crucial for its interpretation. In order to make inferences, accidental properties are sometimes transformed into essential ones: for example, the fact that Julien’s first shot missed can allow us to infer that he was particularly nervous at that moment (in point of fact, a few lines before, it says that his arm was trembling), and this changes the nature of his deed, no longer due to cold determination but to a disordered passionate impulse. Just to stay with Stendhal, with regard to Armance there is critical debate as to whether Octave de Malivert was really impotent, since the text does
