A few years ago I signed up to do a translation of Nouvelles Impressions d’Afrique by Raymond Roussel (1877-1933) for Princeton University Press. I’d never done a translation before, but, well, I’d written a biography of Roussel so knew his oeuvre pretty thoroughly, and had read Nouvelles Impressions a number of times … How hard could it be?
To the translator the poem presents certain obvious difficulties. The least of these is in fact the most obvious: the poem is written in a very peculiar way, for it makes extensive, very extensive, use of brackets and footnotes. The main text of each of its four cantos is conceived as one gigantic sentence, with a single main verb. The opening lines of these cantos look innocent enough: in each case they set about describing some site in Egypt, but before long Roussel has interrupted this description with a parenthetical thought, and this parenthetical thought is itself then interrupted by another, and then comes a footnote, which will have its own sets of parentheses… Therefore, to figure out its opening syntactic unit you have to turn forward to the lines that succeed the canto’s final closing bracket, that is to the canto’s closing lines, then do the same for the lines introduced by the first bracket and that are picked up between its second last and last bracket, and so on, flipping backwards and forwards, gradually working your way into its middle.
It’s a weird way of constructing a poem, and means it certainly doesn’t look like any other poem ever written on the page. By line 21 of Canto I, for instance, one is up to 5 brackets and the first footnote:
(((((parfois une étincelle,
L’entourant de pompiers qui grimpent à l’échelle,
Fait d’un paisible immeuble un cratère qui bout1))))),
(((((sometimes a spark,
Causing firemen, climbing ladders, to surround it,
Makes of a peaceful block of flats a boiling crater1))))),
The footnote then tells us that what we really need when a fire breaks out is a giant like Gulliver, who puts out a fire in the Royal Apartments in Lilliput by peeing on them.
But these brackets and footnotes don’t in themselves present a particular problem to the translator. You’ll have noticed from the extract from the French above that the poem rhymes étincelle and échelle, and, being the methodical man he was, the bout after which we get the footnote Roussel then rhymes with the first line of that footnote. Like most of his earlier long poems, Nouvelles Impressions is written in rhyming alexandrines that alternate masculine and feminine rhymes (ie. rhymes ending in a mute e). But I very quickly decided I wouldn’t be emulating these rhymes. Two earlier distinguished translators, Kenneth Koch (whose version of Canto III was published in 1964) and Ian Monk (whose version in pentameter couplets of the whole thing was published in 2004) had amply demonstrated what really extreme feats of ingenuity and invention use of corresponding rhymes could result in, but my primary aim was to make the French comprehensible.
I started off trying to do it all as literally as possible. I’ve just been reading the great translations of Rilke made by Ruth Speirs in Cairo in the 1940s; in a Note she writes: ‘What I have aimed to do is give as exact a translation of Rilke as is humanly possible.’ I suppose that’s what every translator wants, though no doubt their own personality creeps in insidiously.
As a neophyte to the translating business, what most amazed me was how much it resembled, emotionally, what I imagined it felt like to be chained to a corpse! Not that I didn’t enjoy it, in a peculiar sort of way, but I found the lack of freedom translating entails quite mind-bending. You have to turn into English what he or she has written in French, German, Serbo-Croat, Japanese, whatever. How different this was from fooling around creating ‘versions’, as Pound had done with Fenollosa’s notes for his Cathay poems, as I’d done myself with bits of Apuleius and Ronsard and Boethius and Tacitus and Lucretius and Pliny. Here you had to get it right.
It’s made me feel enormous but unenvying admiration for translators, and led me to abandon earlier plans to translate more Roussel, namely his Textes-genèse (one of which, ‘Parmi les noirs’, has been brilliantly Englished by Ron Padgett), and his fragmentary L’Allée aux lucioles. As I neared the end of Canto IV, sleep deserted me night after night. I found it hard to think of anything else; to breathe; translator’s claustrophobia – is there such a condition?
Now I hold the finished book in my hand, it all of course seems ‘worth it’, though that ‘it’ still lurks in a secret strata of my consciousness, warning me not, at least in the foreseeable future, to try ‘it’ again.