que luego destruía inmediatamente temiendo que fueran desiguales homenajes: ¡Huyendo vas Lilia de mí, / oh tú, cuyo nombre ahora / y siempre es hermosa flor / fragrantísimo esplendor / del cabello de la Aurora! . . . Pero no le hablaba, sino con la mirada, lleno de litigioso amor, pues que más se ama y más se es propenso al rencor, experimentando calofríos de fuego frío excitado por flaca salud, con el ánimo jovial como pluma de plomo, arrollado por aquellos queridos efectos de amor sin afecto; y seguía escribiendo cartas que enviaba sin firma a la Señora, y versos para Lilia, que guardaba celosamente para sí y releía cada día.

Escribiendo (y no enviando) Lilia, Lilia, vida mía / ¿adónde estás? ¿A dó ascondes / de mi vista tu belleza? / ¿O por qué no, di, respondes / a la voz de mi tristeza?, multiplicaba sus presencias. Siguiéndola de noche, mientras volvía a casa con su doncella (Voy siguiendo la fuerza de mi hado / por este campo estéril y ascondido . . .), había descubierto dónde vivía.
I think that this rewriting represents an act of fidelity and that the Spanish text produces exactly the effect aimed at by the Italian original. It is true that a sophisticated reader would realise that all references are to Spanish and not Italian poetry, but the story takes place in a historical period where Northern Italy was largely under Spanish influence, and Lozano made clear that ‘the condition for using that material was that it was poorly known’. Moreover, Lozano made a collage of different texts, so that it was difficult even for Spanish readers to identify the sources. They were rather invited to ‘smell’ a cultural climate. Which was exactly what I wanted them to do when I wrote the Italian.

There is only one objection. The Spanish version seems to mention, apropos Lilia, things, acts and features that the original Italian text does not. To what extent should a translation be allowed to say what the original did not? We shall discuss this embarrassing question in my next chapter.

NOTES

1. Helena Lozano Miralles, ‘Quando el traductor empieza a inventar: la creación léxica en la versión españ ola de Baudolino de Umberto Eco’, in P. Capanaga and I. Fernández García, eds., La Neología (Zaragoza: Pórtico, forthcoming).
2. Taylor, Christopher J., ‘The Two Roses. The Original and Translated Versions of The Name of the Rose as Vehicles of Comparative Language Study for Translators’, in Avirovič Ljljana e Dodds John, eds., Umberto Eco, Claudio Magris. Autori e tradduttori a confronto (Trieste, 27–8 November 1989) (Udine: Campanotto, 1993), pp. 71–9.
3. Schifano (and this is the only flaw in a perfect translation), pulled by his linguistic automatism, translated baumes du Perou. This anachronism can be pardoned because I say in the opening pages of my novel that the manuscript that allegedly inspired me was a French nineteenth-century translation of a lost medieval text, and thus that Perou can be attributed to my pseudo-source. In fact Schifano had chosen, as a stylistic solution, not so much the imitation of a medieval chronicler but rather the style of a nineteenth-century novelist. In any case, better Mecca than Peru.
4. Torino: Einaudi, 1973.
5. El nom de la rosa, tr. Josep Daurell (Barcelona: Destino, 1985).
6. Obviously the meaning of the quotation is irrelevant. For the curiosity of my readers I can say that it was a magic spell suggested for medical purposes.
7. Sylvie, tr. Ludovic Halévy (London: Routledge, 1887); Sylvie, tr. Richard Aldington (London: Chatto and Windus, 1931); Sylvie, tr. Richard Sieburth (London: Syrens, 1995).
8. Milano: Frassinelli, 2001.
9. See my ‘Rhetoric and ideology in Sue’s Les Mystères de Paris’, in The Role of the Reader (Bloomington: Indiana U.P., 1979).
10. See my ‘Casablanca: Cult movies and intertextual collage’, in Faith in Fakes (London: Secker, 1986). Further edition as Travels in Hyperreality (London: Picador, 1987).
11. On the equivalent effect see Eugene Nida, Towards a Science of Translation (Leiden: Brill, 1964). For a larger discussion on these topics see Susan Bassnett-McGuire, Translation Studies (London/New York: Methuen, 1980). On functional equivalence see Ian Mason, ‘Communicative/functional approaches’, and on skopos theory Christina Schäffner, ‘Skopos theory’, as well as Hans Vermeer, ‘Didactis of Translation’, all in Mona Baker, ed., Routledge Encyclopedia of Translation Studies (London: Routledge, 1998). Also in this encyclopaedia see Dorothy Kenny, ‘Equivalence’, where several kinds of equivalence (referential, denotative, connotative, text-normative, pragmatic, dynamic, formal, functional and so on) are listed.
12. Helena Lozano Miralles, ‘Comme le traducteur prit possession de l’Ile et commença à traduire’, in Jean Petitot and Paolo Fabbri, eds., Au nom di sens. Autour de l’oeuvre d’Umberto Eco. Colloque de Cerisy-la-Salle 1996 (Paris: Grasset, 2000).

III. Translation and reference

In my first essay I stated that one of the kinds of equivalence that could be judged as essential for a translation was referential equivalence. In very simple terms, a translation should convey the same things and events as the original.

I mean reference in its strictest sense,1 that is, an act by which, once one knows the meaning of the uttered words, one determines states of a possible world (which can be either the world we are living in or the one described by a novel), and asserts that in a given spatiotemporal situation certain things or certain events happen. As Strawson said, ‘mentioning or referring is not something an expression does; it is something that someone can use an expression to do’.2 Cats are mammals is not, from my point of view, an act of referring, because it simply establishes which properties we should assign to cats in general in order to use, in the course of a verbal interaction, the word cat. The same would happen with the phrase unicorns are white.

In Strawson’s terms, ‘to give the meaning of an expression . . . is to give general directions for its use to refer to or to mention particular objects and persons; to give the meaning of a sentence is to give general directions for its use in making true or false assertions’. If somebody says that cats are amphibians or unicorns have a striped mantle, one should not say that these two assertions are ‘false’, but more properly that they are ‘wrong’, at least if we follow zoology manuals and all the traditional descriptions of unicorns found in ancient bestiaries. In order to decide that cats are amphibians is a correct assertion I would have to ask the whole of society to restructure its entire system of natural classifications, as happened when the learned community decided to consider the assertion dolphins are fishes as wrong.

On the other hand expressions like there is a cat on the mat, my cat Felix is sick or Marco Polo said to have seen unicorns are referring to situations of the actual world (including the fact that Marco Polo reported having seen unicorns). These assertions can be empirically tested and judged as true or false. In ordinary situations we react to cats are mammals and there is a cat on the mat in two different ways. In the first case we open an encyclopaedia in order to see if the statement is correct; in the second, if we do not trust the speaker, we check de visu, to ascertain whether the statement is true.

Now, most of the texts people translate are reports on facts, narrations, poems and so on, and all of them mention something that should be taken as if it were the case. A newspaper article saying that so and so died yesterday presupposes that we take for granted that so-and-so is really dead. A novel saying that Prince Andrei died commits the readers to take for granted that (in the possible world of that novel) Prince Andrei really died – to such an extent that, if the narrator shows him alive and well in the course of the novel, the reader will feel astonished. In the same vein they will consider the assertion of another character, who says that Andrei is still living, a lie.

A novel describes a world (a possible one, even though not necessarily a fictitious one, as in historical novels). Translators are not allowed to change the true references to that world and no translator could say, in his version, that David Copperfield lived in Madrid or Don Quixote in Devonshire.

Such restrictions admit many exceptions. For instance, when a given expression has a connotative force it must keep the same force in translation, even at the cost of accepting changes in denotation. To make a very elementary example, if an English text says that it is raining cats and dogs, no Italian translator is obliged to respect the reference and to translate as piove cani e gatti because the expression does not mean anything in Italian. In Italian one would say piove come dio la manda.

Thus a provocative question can be: to what extent, in order to preserve its proper effect, can a text be altered without violating the equivalence in reference?

Disregarding reference

In my translation of Sylvie, an instance in which I felt obliged to make a modification in terms of reference was that of the visit to Châalis in chapter 7. Nerval says that it takes place le soir de la Saint-Barthélemy and a little further on he speaks of le jour de la Saint-Barthélemy. There is no contradiction because the evening of St Bartholomew belongs to St Bartholomew’s day. Translators usually faithfully render the evening in the first occurrence, and the day in the second.

Now, for every French reader it is enough to mention la Saint-Barthélemy to evoke the night of one of the most cruel massacres of European history, but in Italian the same connotation is provided by the expression ‘la