Although it is not my intention to attempt to add anything of significance to the interminable debate between the literalists and the imitators or adaptors or free renderers, I would like to say something about one aspect of my own specific experience of translating the poetry of Francisco de Quevedo who lived between 1580 and 1645. It is possible that what I have to say may be of interest, and even of use, to lunatics like myself who expend vast amounts of time and energy in trying to carry across into their own language the past poetry of another language.

To begin with … The major problem I encountered in translating Quevedo's poems was what to do about the prosodic elements of the Spanish. The French poet and critic and translator of Shakespeare, Yves Bonnefoy, in a short essay he wrote to preface his translation of Yeats's 'Sailing to Byzantium', succinctly states the problem and then proceeds to offer the following as his own solution:

To translate the poetry of the past according to rules [he means the rules of traditional prosody] that are now empty of substance would be … a betrayal, and one has to attempt it with the only line that is possible for us: one that is free, faithful to personal truth, submitting to laws like every life but not to conventions, and remaining content with evoking, potentially only, the initial regularity, through the occasional surfacing of the old metres which we can accept only when they are born as if from themselves.

Behind Bonnefoy's view of the problem and his proposed solution to it, is his understanding that the regular and formal prosody of tradition is authentic only when and insofar as it reflects the regularity of the accepted norms that inform the Weltanschauung of the society to which the metricist belongs. For instance, a translator of poetry of the past who himself in his own personal and societal life does not subscribe to such regular norms, is playing false in mechanically copying the metrical regularity of his original. According to Bonnefoy:

The translator cannot set himself the task of a regular verse line without failing in one of the dialectics that structure true poetic creation. When Yeats wrote 'Sailing to Byzantium', the metric structure he chose was a forma formans, something generative, since it determined the as yet uncertain 'content' of which it became, in return, an aspect. Form and content are thus a single act, wherein a freedom is inscribed. But when the significations are, in advance and entirely, defined as is the case when one translates, a form decided a priori can only be external to them, a form of adjustment demanding ingeniosity, removing from the translator's sphere of decision his obsessions and myths which I believe to be the only things c capable, however, of rekindling the darkened glow of the work.

It seems to me that these acute observations of Bonnefoy are of crucial importance to any translator whose concern is the poetry of the past. What I propose to do no is to demonstrate, as best I can, how I have taken cognisance of Bonnefoy's analysis of the problem and how I have tried to accommodate it in translating the poetry of Quevedo.

I believe it goes without saying that the effects of metre on readers are extremely difficult to calculate. In English it seems to me that it is wellnigh impossible to have any two readers scan a poem in exactly the same way. Often, indeed, the differences, even when the scansion is carried out by 'experts', are very great, so great that one might justifiably wonder if the 'experts' are reading the same poem. Little wonder, then, that Hopkins' metrical experiments are still the subject of debate and various interpretation. The root of the problem seems to be that the metres of classical Greek and Latin, which European literary culture used as its model, are 'out of sorts' (as it is often put) with the European languages that felt obliged to accommodate them. It is surely ironic that Roman poets sometimes complained that Greek metres, which they felt obliged to accommodate, were an imposition for them and distorted Latin into syntactic patterns out of keeping with what they felt to be the genius of their language. Hopkins, expert in classical metres, deeply experienced the lack of relevance of these metres to the English language; and he devoted a great deal of energy and a very great deal of talent to supplying what he considered a more 'natural' or 'native' metrics.

I hope I have made it clear from the outset that I rejected the strategy of reproducing (or attempting to reproduce) the regularity of Quevedo's verse line, agreeing as I do with Bonnefoy that such regularity is not meaningfully separable from the world-view of which it is an aspect of expression. Nevertheless, that regularity, it must be admitted, is an important ingredient of the original work and cannot simply be ignored because of its seemingly irreducible problematic nature. What should then be done?

Some years ago I came upon what I then considered a few interesting remarks made by Gerald Bullett in his Introduction to his fine anthology Silver Poets of the Sixteenth Century (London, 1947). Bullett was weighing the charge of crude versification against Wyatt, and reproduced a scansion (carried out by strict metrical procedure) of a Wyatt sonnet ('The longe love that in my thought doeth harbar') made by a Professor Padelford, which reduced, so far as Bullett was concerned) to aural chaos:

Now no one contends that Wyatt's sonnets are models of correct versification; admittedly his rhythms are often awkward and blundering. But Padelford's pointing of this particular specimen is quite fantastically wrong … This talk of trochees and iambs belongs rather to the grammar books than to literary criticism. The jargon of (a classical and so largely irrelevant) prosody is no necessary part of an English poet's equipment, useful though it may be to a student of Greek verse. Many of the loveliest of English lyrics have been written by men to whom such terms were Greek in more senses than one.

The t'rum-t'rum method of scanning blank verse is derived presumably from the fundamental error that syllable-counting is the basis of English metre. It is a method that can be applied with consistent success only to the worse examples. Surrey [Bullett quotes some lines by surrey which , like Wyatt's, have been rhythmically mangled by Professor Padelford] may or may not have aimed at strict uniformity of stress in his blank verse, but mercifully he did not achieve it. Either inexpertness or native sense - most probably an alternation of the two - saved him from the flat pedestrianism of the lesser Elizabethans and their undistinguished successors. He can be clumsy, and he can be metrically tedious; but he escapes the consistent tedium which must result, in English, from an exact undeviating identity between the basic metre and the actual tempo: in the best blank verse one is conscious or half-conscious of both, and much of the energy of the verse is generated by the tension between them, or, in other words, by the variety of the patterns woven on a basic metrical structure.

Now this analysis, though a little crudely phrased and somewhat too facilely dismissive of changes that no doubt took place in the pronunciation of English down through the centuries, seems to me to be generally valid. It can be applied to Shakespeare's verse, and it explains how Shakespearian actors can often succeed in eliminating the basic metrical pattern (which is iambic) as they play on the tempo to achieve a conversational rhythm which they believe (fallaciously, I think) makes Shakespeare more acceptable to a modern audience with its presumed antipathy to verse (such an audience it is to be hoped will scarcely notice that they are hearing verse). At any rate, I think it is true that in the best blank verse of the Elizabethans and Jacobeans, the basic metrical beat serves as the regulating, though sometimes barely heard rhythm for what one might call the interpretative or dramatic rhythm, language in the mouth of a credible speaking man.

Eliot in his essay "Reflections on "Verse Libre" has this in mind when he remarks,

…At the beginning of the seventeenth century, and especially in the verse of John Webster who was in some ways a more cunning technician than Shakespeare, one finds the same constant evasion and recognition of regularity. Webster is much freer than Shakespeare, and that his fault is not negligence is evidenced by the fact that it is often at moments of the highest intensity that his verse acquires this freedom. That there is also carelessness I do not deny, but the irregularity of carelessness can be at once detected from the irregularity of deliberation … We may therefore formulate as follows: the ghost of some simple metre should lurk behind the arras in even the 'freest verse; to advance menacingly as we doze, and withdraw as we rouse. Or, freedom is only freedom when it appears against the background of an artificial limitation.
(T.S.Eliot, Selected Prose, pp.89-90, Penguin Books, 1953)

By taking up these observations of the unlikely combination of Bullett and Eliot, there seems to me the possibility of doing what Bonnefoy recommends: a means of suggesting the regularity of Quevedo's verse line while at the same time allowing oneself the freedom to deal adequately with Quevedo's obsessions and myths, the 'essential' stuff of his poetry from which the original must be regenerated by the translator.

So I adopted this procedure and attempted to render Quevedo's sonnets with an irregular five-stress line. However, it soon became apparent to me that Quevedo's latinate concision was being diluted and lost; I was unconsciously allowing myself too much scope for paraphrase. Then I tried a four-stress line and although I found it sometimes difficult, the results I achieved satisfied me better. The constraints imposed by the four-stress line produced a greater compression and strictness of syntax that reasonably reflected Quevedo's tightly controlled Spanish.

The question of rhyme still remained. Rhyme, always fraught with danger in English and so facile in inflected Spanish as to have a different effect, I decided from the beginning to leave severely alone unless it occurred fortuitously with good effect. Instead of rhyme, however, and in keeping with my intention of working prosodically only by suggestion, and not practice, I used assonance, a major resource of system and pattern in Spanish. Consciously working with assonance came fairly easily to me since poetry in the Irish language uses assonance instead of full rhyme, and a number of Irish poets working in English, above all Austin Clarke, have experimented extensively with assonance. For years I was familiar with their work and I had myself as a poet experimented occasionally with assonance as an alternative to full rhyme. Nevertheless I was determined not to apply mechanically an assonantal principle that would result in awkward-sounding rhythms. Whatever syntactic tortuousness there would be in my translations of Quevedo's poetry, this would not be the result of any mechanical contrivance on my part, but would reflect the tortuousness often to be found in Quevedo himself as he strove for latinate concision.

For all the dogmatic perversity of his views on the audible reading of poetry, the American critic and poet Yvor Winters knew that there was a basic beat and a huge variation on that basic beat, though he was intolerantly and stupidly limited in his notion of the degree of variation permissible, clinging as he did to irrelevant concepts of spondees, trochees and other such figments of the imagination .

Rhythm and metre, it should be observed, are quite distinct from each other, in spite of the fact that many critics (myself among them) sometimes use the two words as if they meant the same thing. Metre is an arithmetical norm, the purely theoretic structure of the line; rhythm is controlled departure from that norm. The iambic pentameter norm, for example, proceeds as follows:

One two, one two, one two, one two, one two.

Yet no other line in the language corresponds exactly to the line just given; and to achieve another as regular one will have to resort to the same repetitive structure with a new pair of syllables. Every other line will depart from this one for these reasons: no two syllables ever have the same degree of accent - that is, so far as versification is concerned there is no such thing as in inherent accented or unaccented syllable, but syllables which count technically as accented can be recognised as such only with reference to the other syllable or syllables within the same foot; secondly, although quantity or syllable-length has no part in the measure, it is, like accent, infinitely variable and it affects the rhythm; and thirdly, feet of other types may be substituted for iambic feet, at least within reason. As I have said, rhythm results from the proper control and manipulation of these sources of variation.
(Yvor Winters, The Function of Criticism, London, 1962, p.82)

T.S.Eliot had something similar in mind when he wrote:

… At the beginning of the seventeenth century, and especially in the verse of John Webster who was in some ways a more cunning technician than Shakespeare, one finds the same constant evasion and recognition of regularity. Webster is much freer than Shakespeare, and that his fault is not negligence is evidenced by the fact that it is often at moments of the highest intensity that his verse acquires this freedom. That there is also carelessness I do not deny, but the irregularity of carelessness can at once be detected from the irregularity of deliberation …

We may therefore formulate as follows: the ghost of some simple metre should lurk behind the arras in even the 'freest' verse; to advance menacingly as we doze, and withdraw as we rouse. Or, freedom is only truly freedom when it appears against the background of an artificial limitation.
(T.S.Eliot, Selected Prose, Penguin Books, London, 1953, pp. 89-90)

Recently, and after finishing the bulk of my translations, I was re-reading J.L. Gili's Lorca (Penguin Books, 1967), and I noted in his Introduction the attention he drew to Stephen Spender's (as Gili put it) 'beautiful rendering into sonnet form' of Lorca's 'Adán' (for which Spender used Gili's plain prose translation). So, re-reading Spender's rendering I found, to my surprise and satisfaction, that Spender had consciously or unconsciously used a four-stress line and that he had not allowed the pursuit of rhyme, well achieved nonetheless, to produce deviation from the 'sense' of the original.

It is relevant enough to my own efforts in translating Quevedo to quote in full Spender's fine version:

Morning by tree of blood is moistened

where the newly delivered woman groans.

Her voice leaves crystals in the wound

and in the windows a print of bones.

While the light comes in secure and gains

white boundaries of oblivious fable

in the rash from the turmoil of the veins

towards the clouded coolness of the apple.

Adam dreams in the fever of clay

of a child which draws nearer galloping,

with the double throb of his cheek its way.


But another obscure Adam sleeping

dreams neuter seedless stone moon far away

where the child of light will be kindling.

Of it will be immediately apparent that not every reader would read the poem with precisely these stresses, but it should nonetheless be obvious that there is an irregular iambic background beat and that most of the lines have four stress rather than a regular five. It should also be apparent that Spender has managed quite a few rhymes, as befits the sonnet, without drawing too much attention to them.

Now before finally looking at the application of these ideas to an actual translation of my own, I must reiterate that I did not bind myself mechanically to the employment of a four-stress line or to a deliberate pattern of assonance. If I found that a three-stress line or a five-stress line adequately served the 'sense' and caused no rhythmic awkwardness, I felt free to take that option. And in the use of assonance, I was not attempting to reproduce a sonnet rhyming-pattern such as abbaabbacdcdcd, but merely to suggest the use of rhyme in the original. Using a very crude form of phonetic symbolisation, the following should give some indication of the sort of prosodic procedures that were involved in my translations [transcriptions would be EE as in seen, y as in I, e as in men, a as in day, u as in moon or you, and so on]:

But first of all, here is the original Spanish of Quevedo:

Oír, ver y callar remedio fuera
en tiempo que la vista y el odio
y la lengua pudieran ser sentido
y no delito que ofender pudiera.

Hoy, sordos los remeros con la cera,
golfo navegaré que (encanecido
de huesos, no de espumas) con bramido
sepulta a quien oyó voz losnjera.

Sin ser o&iacutedo y sin oír ociosos
ojos y orejas, viviré olvidado
del ceño de los hombres poderosos.

Si es delito saber quién ha pecado,
los vicios escudriñen los curiosos:
y viva yo ignorante y ignorado.

Now here is my version in English:

&nbspEE EE EE y e EE

To hear, to see, keep quiet, would be A
a u y e y a EE

a cure in time when sight and speech A
a EE EE EE e

and hearing would be simply senses B
a y e

and not a crime to give offence. B
oo ai o e EE e

Today, the oarsmen's ears deaf B
a y ai a EE

with wax, I'll sail the gulf that, bleached A
o o EE o

with bones, not foam, buries with roar D
a oo EE y oo

the man who hears the siren's lure.
EE y EE

Unheard, unhearing, with eyes and ears A

oo y e
unused, I'll live forgotten C
y au au e

by the frown of powerful men. C

y oo oo

If it's a crime to know who's sinned,

o oo a y

let those are curious examine vices,
o a o
a
and let me live, unknowing and unknown.