ABSTRACT
Textual relations How does a self -translation relate as a text to 'normal' translations? Can it be said to possess its own distinctive character? In an essay on James Joyce's own Italianizing of two pas sages from his Work in Progress (the future Finnegans Wake) , Jacqueline Risset answers in the affirmative. Unlike translations ' in the usual sense of the word' ( 1984: 3) , she argues , Joyce's texts are 'no pursuit of hypothetical equivalents of the original text (as given, definitive) but as a later elaboration represent ing . . . a kind of extension, a new stage, a more daring variation on the text in process' ( 1984: 6). This allows her to oppose Joyce's auto-translation to the ' fidelity and uninven tiveness ' ( 1 984: 8) that characterized the French translation of the same passages, by a team that included no less than Philippe Soupault, Yvan Goll, Adrienne Monnier and Samuel Beckett. What is at stake here is the old notion of authority , of which original authors traditionally have lots and translators none. Since Joyce himself wrote these second versions in idiomatic and creative Italian, they are invested with an authority that not even an 'approved' translation by diverse hands can match. The public's preference for an author's translation is less based on an extensive study of its intrinsic qualities - though Risset does conduct such an examination - than on an appreciation of the process that gave birth to it. The reason for this state of affairs is quite obvious, as Brian Fitch points out: ' the writer translator is no doubt felt to have been in a better position to recapture the intentions of the author of the original than any ordinary translator' ( 1988: 1 25). In terms of its produc tion, an auto-translation also differs from a normal one, if only because it is more of a double writing process than a two-stage reading-writing activity. As a result, the orig inal's precedence is no longer a matter of 'status and standing' , of authority , but becomes 'purely temporal in character' (Fitch 1988: 1 3 1 ). The distinction between original and (self-)translation therefore collapses, giving place to a more flexible terminology in
which both texts are referred to as 'variants ' or 'versions' of equal status (Fitch 1988: 1 32-3).
