I've been translating a bit from Camões' Lusíadas and thought I'd jot down some thoughts, interspersed with polemical asides.
First on the matter of form. English is quite capable of handling the rhyme requirements of Camoes' ottava rima. Byron's Don Juan is proof enough of that, opening as it does with the following stunt-rhymed stanza:
Bob Southey! You 're a poet—Poet-laureate,
And representative of all the race;
Although 't is true that you turned out a Tory at
Last,—yours has lately been a common case;
And now, my Epic Renegade! what are ye at?
With all the Lakers, in and out of place?
A nest of tuneful persons, to my eye
Like four and twenty Blackbirds in a pye;
But unlike Byron, and unlike Dante, Camoes didn't mind banal rhymes. His verse is full of rhymes that are based purely on the same morpheme. Dante only occasionally allows himself a rhyme like abandonnai/intrai/trovai (which are rhymes in roughly the same way that harrowing/wandering/opening are.) Camoes doesn't seem to mind them at all. His rhymes are semantically interesting, but often satisfy no more than the bare minimum of requirements for what counts as a rhyme. He is also not shy about wrenching things about for a rhyme, as when he gives the form amareilo (for amá-lo-eis), or forces the reader to read Taprobána at the end of a line, instead of the normal Tapróbana.
To translate Camoes into classical ottava rima, as many have done, is possible in English. The first translator of the Lusíadas did exactly that. To do the same kinds of things in octave rhyme in English as can be done in Portuguese (or for that matter Italian) is possible only to a point. The recent octave rhyme translations of Ariosto and Tasso by David Slavitt and Max Wickert demonstrate nicely that the same form actually lends itself to different — but overlapping — possibilities in English and Romance. It will not do to exaggerate nature of the difference, but it is real. When you need syntactic scrambling, say, to tack a string of adjectives after the noun they modify, there are certain effects you can't quite duplicate. There are other things you can do more easily.
In English poetry, end-stopped lines are not as important to maintaining rhythmic feel as they are in Portuguese, which means that enjambment can be used far more robustly. And with no requirement for a caesura anywhere in particular, you can use a syntactic break in the same spot in a series of lines to set them apart from what comes before and after them. You can also switch up the rhythm using different numbers of unstressed syllables between verse beats. This is because traditional English prosody is based on word-accent, and not just syllable count.
It's worth noting — if only because people so commonly believe otherwise — that this difference is not a strictly linguistic difference. There is no intrinsic reason why English-speakers don't make more use of purely syllabic verse-lines. Russian — which is more stress-timed than English — made great use of syllabic verse inspired by French and Polish, before it gave way to German-inspired syllabo-tonic verse. But it didn't have to happen that way. English-speaking poets, and usually readers, tend not to feel like syllabic verse works very well. But — like the perception that major chords sound "happier" than minor chords — that is only because the stuff they are exposed to has conditioned them to feel that way about English. Our non-use of syllabic verse-lines is just as much a result of historical accident as the fact that both English and Portuguese are written in the Roman alphabet. It is a myth that syllable meters are less intuitive to English speakers, or less congenial to English, than the accentual or syllabo-tonic meters we do use. English-speaking schoolchildren, when first exposed to iambic pentameter, often find it easier to count out the syllables than to both (a) keep track of where the verse-beats should technically go, and (b) keep that in mind as a contrast to where the word-stresses in the line actually seem to fall. Syllabic meters don't need a language with syllable-timed speech rhythm to work. Modern Brazilian Portuguese is syllable-timed, whereas European Portuguese is stress-timed. That didn't stop Pessoa from keeping count of syllables, even the syllables that were no longer pronounced in his own speech. No surprise there. Syllabic meters not only do not depend on how audible or prominent the syllables are relative to each other, but they don't even require a syllable to be actually pronounced in order to "count" metrically. Parisian French lost its post-tonic /ǝ/ hundreds of years ago, and yet traditional French versification essentially requires the poet to act as if that vowel were still pronounced, (and many styles of French singing they still are, or are only pronounced optionally as the lyricist's rhythm requires.)
Prosody is a matter of what's in your head, not just what's in your mouth. And there's a good deal more in your head than how you talk.
More on form.
With Dante's Comedia, the terza rima is a form invented by the author, so integral to the enterprise that to abandon it is to abandon an important element of narrative motion and structure. I am willing to go to the finish-line for Dante's terza rima. But to insist on full ottava rima in translating Os Lusíadas, when Camoes doesn't make too much of a point about proper and improper rhymes, is not hung up on formal punctilios, and is using not a form of his own design but the conventional stanza of his day. would be overkill, and perhaps even a bit dense since modern audiences don't much care about that sort of thing. But to abandon the form seems also unnecessary, and unwise since it is part of how Camoes organizes narrative rhythm, and the way thoughts are expressed is quite clearly chunked according to where rhymes fall. In traditional Portuguese rhymed verse (leaving aside popular songs and the like) every line must rhyme with at least one other line somewhere nearby (though "nearby" could mean "in the next stanza.") If you see a line-end, you're expecting its mate before too long. This has not been true of English for some time. We are very used to quatrains where only the second and fourth line rhyme. A paired-down ottavo, of the form ABCBDBEE, seems like it will do just fine. The expectations of modern English readers allow for a relaxation here that frees me up to attend to other matters more.
But there is another point on which modern English readers' expectations are — to use the technical term — a motherfucking nuisance, and that is when it comes to poetic syntax and the traditional English poetic register. Specifically, modern readers have a problem with shifting word order around to suit one's needs or whim. Compared to earlier readers, mainstream modern English readers have an unsolved stick-related proctological issue. Ironically, this comes hand in hand with the belief that rigid adherence to prose syntax, and a taboo placed on the poetic register, are in some sense a form of freedom and disinhibition. The "unnatural" inversions, for which older rhymed translations in English are chastized only seem unacceptable to Modern English speakers because they have been taught to feel that way about anything written or translated too recently. A certain ideological work is needed to harmonize the cultivated high regard (or at least tolerance) for the likes of Spenser and Milton, who freely use syntactic scrambling to various degrees, with an aversion to "unnatural" language (which really is code for "archaic language" as the term is used even by people who have no problem with e.e. cummings.) We convince ourselves that the traditional poetic register is only appropriate for things written before 1900, after which date it is seen as unnatural. Whereas of course our own sensibilities about what language is or isn't worth taking seriously in a poem are completely "natural" and a function of the intrinsic properties of the language. While one must for practical reasons humor modern audiences on this matter (at least up to a point), I see no reason to pretend to find it less absurd than I do. Languages do not have essences. It is only the high-end kool-aid we sip that makes it seem like they do. And I do not like the buffoonery that this type of thinking leads to. A similar sleight of handwaving allows our modern arbitri elegantiarum to enjoy Bernart of Ventadorn or Sappho as great poetry, while also enjoying their anger at Bob Dylan's Nobel prize. They will have their cake and eat it too. So they vomit up the first cake before scarfing down the second, while shooing away anyone so boorish as to mention the puke now perfuming the patio, or so impertinent as to ask if this cake tastes better or worse when mixed with the upchucked aftertaste of its predecessor.
One translator describes the difference between English and Portuguese syntax as that of "a syntax of inflected nouns and verbs, and a syntax where meaning depends on word order" to justify avoidance of syntactic scrambling. This is overstated and underthought. True enough that word-order is more important in English syntax than Portuguese, but the difference between O homem mordeu o cachorro and o cachorro mordeu o homem is exactly that between "The man bit the dog" and "the dog bit the man." Anybody who's read Milton or Spenser knows that English syntax can sustain a good deal of scrambled word-order without obscuring meaning, and Gerard Manley Hopkins shows just how the opacity caused by word-scrambling can itself be used for artistic effect in English poetry just like any other linguistic property.
Literary Arabic is more inflected than Portuguese, with distinctions not only of gender, number and person, but also case. And yet syntactic scrambling, though a feature of traditional Arabic verse, is never quite as extensive as it is in medieval European or Renaissance vernacular verse, let alone Latin verse. Persian is less inflected than Arabic, and yet its traditional poetry also has far more scrambling. Russian is almost as inflected as Latin, and yet Russian verse doesn't use nearly as much syntactic contortion as can be found in the sonnets of Quevedo.
There are reasons for these difference in poetic habit, but they are not reducible to the structure of the language in which the poetry is composed. Syntactic scrambling is a natural thing to employ in poetic language if you've been raised on Latin and Greek. Greek verse word-order is free, but in Latin poetry it is often pushed to the breaking point and sometimes well beyond it. Anybody who knows Latin well is aware of the expressive potential of scrambled word-order in poetry, even if they're composing in a popular genre for the unlatin'd. It is against the backdrop of a European high culture steeped in Latinity that the medieval and Renaissance European poetic habit of syntactic scrambling is to be understood.
Camoes doesn't scramble syntax nearly as much as his more palpably baroque contemporaries, and by comparison to them has a more direct flavor. But there are plenty of places in the Lusiadas where word-order is so scrambled that the correct syntactic relationship is hard — and a few times impossible — to apprehend with certainty. As Ivo Castro puts it, Camoes is given to "a inversão da ordem mais frequente das partes da oração, o que resulta em frases de grande efeito pela dificuldade de entendimento que criam a uma abordagem inicial.” That Camoes feels no obligation to stick to the syntactic behavior of prose or ordinary speech is not at all peculiar, nor is it due to the character of the Portuguese language. He was like any other European poet of his day who did so because the audience was willing and more than able to deal with it. Modern English readers do not want to deal with it. I cannot change that, and have far better things to do than to try.
A modern translator working with a Renaissance poem about a historical event also has the option of dipping into different historical layers of the language if they wish. Modern readers will — no matter what translation choices are made — look at the history of European expansion very differently than Camoes did. The poem will mean different things to us. There is no point pretending this is not so, and no sense in not tapping into that as a resource when translating. Some English translators of the Lusíadas translate gentío as "Gentoo" (rather than "Indian" or "Hindu") in certain context to suggest a certain mood, or maintain a certain feel. There is a sense in which "Hindu" is out of place in the Lusíadas, for much the same reason that "Negro" or "Black Man" would be out of place in the mouth of Huckleberry Finn. I didn't feel the need for Gentoo because I didn't feel like sending readers zooming to the dictionary. But I do not feel like there is much to be gained in always keeping the name of the prophet Muhammad in its current standard English form. Mahomet will do fine for Camoes' Mahamede. Translating Mahometo and Mahomético as "Muslim" or "Islamic" would feel like cowardice. There is a value, as Landeg White's semi-free verse translation has demonstrated, in downplaying or bleaching out those aspects of the Lusiadas that make modern readers uncomfortable, and centering its role as a dramatization of cultural contact in a time of exciting discovery when so much of what Europeans thought they knew was coming untrued. But I don't think you need to bleach out as much of the former to get all that the latter has to offer. It might not be pleasant for the reader, but words like Mohamedan, Mahomet and Mohamedanism will feature in the Lusiadas as I translate.
One other thing that seems worth doing in translating the Lusiads is respecting Camões' linguistic liberties, just as it is worth respecting those of Dante.
When modern English speakers hear or read something like stultilocution or crazify or psychochasm they often protest that it is "not a word." This attitude is very far removed from English-speakers during the Elizabethan period, when coining words based on existing morphological patterns, or outright borrowing from other languages like French or Latin, was very much a part of educated linguistic behavior.
As in England so in Portugal, linguistic attitudes during the European Renaissance were not like those of today. A number of words common in Portuguese are used for the first time in the Lusiads. Including, in the opening passage, the words grandíloquo, devastar and tuba. It is traditionally said that Camoes borrowed most of these novencides from Latin. He certainly did borrow Latin words. But the direct source was more commonly Castilian, which Camões knew well and in which he composed a number of poems. (The earliest attested use in Castilian of grandiloquo is about a hundred years earlier. And devastar is even older.) Terms like grandíloquo and devastar would have seemed like high-flown cultisms to his audience. So far the best solution I have come up with is to translate the former as a nonce compound grandtongued, and the later by dipping into Renaissance English by using havoc as a verb.
