3

The Critic as Translator

MUCH HAS BEEN written about translation in general and the translation of Chinese poetry into English in particular. I do not intend to revive all the old controversies about how to translate Chinese poetry, or to repeat everything that I have written on the subject, but to try to clarify a few basic issues concerning the translation of Chinese poetry by discussing them within the conceptual framework of this book. However, in order to make the discussion intelligible to readers not familiar with my previous writings, it will be necessary for me to reiterate certain points that I have made before.

I should like to begin by drawing a distinction between the poet as translator (or poet-translator for short) and the critic as translator (critic-translator) and pointing out that they have different aims and different readerships.¹ The poet-translator is a poet or poet manqué whose native Muse is temporarily or permanently absent and who uses translation as a way to recharge his own creative battery. His primary aim is to write a good poem in English based on his understanding or misunderstanding of a Chinese poem, however he may have arrived at this. A critic-translator is a critic writing in English about Chinese poetry. His primary aim is to show what the original poem is like, as a part of his interpretation. As for their respective readers, the poet-translator aims at those who cannot read Chinese but enjoy reading poetry in English, whereas the critic-translator may have several kinds of readers: English-speaking students of Chinese poetry struggling with the original texts and needing guidance and help, specialists who wish to compare someone else’s interpretations and evaluations of Chinese poetry with their own, and those who cannot read Chinese but are seriously interested in learning more about it for comparative or other scholarly purposes rather than simple enjoyment. The first kind of readers, including most book reviewers for the popular press and for journals not devoted to Asian studies, are in no position to know, and perhaps do not care, how far a translation resembles the original, and when they say a work is a good translation, what they really mean is that it reads well in English. The other kinds of readers need to know how far a translation resembles the original and not simply how well it reads in English.

The two kinds of translators are, of course, not mutually exclusive. Ideally, they should coincide: the same translator should be a competent critic of Chinese poetry and a competent poet in English, but in actual fact there has rarely been a translator who translated directly from the Chinese unaided and who is also an established poet in English in his own right. Gary Snyder seems to be the only exception, but even his translations are not free from errors that are obviously due to misunderstandings of the original texts and not to poetic license. Other translators of Chinese poetry form a whole spectrum, from the freely inventing Pound at one extreme to the uncompromisingly prosaic Karlgren at the other. On the whole, poet-translators, who may know some Chinese or none at all, such as Pound, Amy Lowell, Witter Bynner, and Kenneth Rexroth, have either reworked existing translations by scholars or had collaborators. Unfortunately, the collaborators or “native informants” they chose have not always turned out to be well-informed about Chinese poetry. Then there are some translators who appear to aim primarily at producing poetic English versions of Chinese poetry, such as Arthur Waley, David Hawkes, A. C. Graham, Burton Watson, and Jonathan Chaves, yet none of them is well known as a poet in his own right. Apart from a few early poems that he wrote as an undergraduate at Cambridge and one or two self-parodies, Waley did not publish his own poetry; Graham has published at least one original poem; the others, as far as I know, have not published original poetry. Turning to critic-translators, such as Hans Frankel and Stephen Owen, they naturally also wish to produce readable English versions of Chinese poetry, but their primary concern is to demonstrate certain features and qualities of the original poems, rather than writing good poems in English. It is, of course, with the last kind of translator that I identify myself.

Since the poet-translator and the critic-translator have different aims and different readerships, they naturally differ in the ways they translate. Two examples should suffice to illustrate this point. First, when faced with a cliché in Chinese, a poet-translator would probably try to improve it and not replace it with an English cliché, whereas a critic-translator would hesitate to improve the original but feel obliged to point out that it is a cliché. Second, in the case of allusions, with which Chinese poetry is replete, a poet-translator would avoid cumbersome footnotes but either paraphrase or incorporate whatever explanation he thinks necessary in the translation itself, whereas a critic-translator would preserve all the allusions and then add notes or commentaries to explain their significance and poetic functions.

Perhaps we can now lay to rest the ancient dispute between “literal” and “free” translations, or at least discuss the question in different terms. Rather than asking whether a translation should be “literal” or “free,” we should ask what linguistic structures fulfil what poetic functions in the original, and what linguistic structures in English can fulfil similar poetic functions. A poet-translator would probably not be overconcerned with the first question but would address himself to the second question, using any linguistic structures in English that he thought could best fulfil the desired poetic functions. By contrast, a critic-translator would consider it his duty to show how certain poetic functions are fulfilled by certain linguistic structures in the original, whether such structures can be reproduced in English or not, and whether, if reproduced, they can fulfil similar poetic functions in English or not. He may have to choose between two alternatives: either to reproduce the original linguistic structures, if possible, no matter whether they are poetically effective in English or not, while informing the reader how they work in the original, or to substitute different linguistic structures in English, which he hopes will fulfil similar poetic functions, and then inform the reader that these functions are fulfilled by other linguistic means in the original.

It should be obvious from the above discussions that the interrelations between linguistic structures and poetic functions are complex and changeable, and that it is a fallacy to think that by reproducing the original linguistic structures of poetry one will automatically reproduce the original poetic effects. Such a fallacy seems to underlie the tendency that I have termed “barbarization,” by which I mean the attempt to reshape the English language so as to preserve the linguistic structures of Chinese poetry and its underlying ways of thinking and feeling. The opposite tendency, which I have called “naturalization” refers to the attempt to turn Chinese poetry into English poetry without violating existing conventions of the English language.² For those who advocate naturalization, “the first requisite of a translation is that it should not sound like a translation,” as Marianne Moore put it,³ but to those who favor barbarization, a translation should sound like a translation and not an original English poem. Let us widen the scope of our discussion for a moment and consider briefly translations from and into other languages. Historically, most successful English translators of poetry from other languages, from Chapman through Dryden and Pope to FitzGerald, were naturalizers, although barbarization was not absent, a notable example being the King James Version of the Bible. On the theoretical level, barbarization found an advocate in Robert Browning, if the account given by John Addington Symonds is accurate: “Browning’s theory of translation. Ought to be absolutely literal, with exact renderings of words, and words placed in the order of the original. Only a rendering of this sort gives any real insight into the original.”⁴ This, of course, assumes that one can draw exact equations between words of one language and those of another, and that the same word order in one language will mean the same thing in another. If only this were true! In the twentieth century, Walter Benjamin quoted with approval Rudolph Pannwitz’s observation that German translators should not try to turn Hindi, Greek, or English into German but do the reverse, so as to expand and deepen the native language by means of the foreign one.⁵

To turn back to translations of Chinese poetry: the earlier translators, such as James Legge and H. A. Giles, put their translations into traditional English meters and adopted pseudoarchaic diction, so that their translations now generally read like parodies of polite Victorian verse. More recent translators, from Arthur Waley to present-day translators (with a few exceptions like Alan Ayling and Duncan Mackintosh), have eschewed traditional English meters and rhymes but adopted sprung rhythm verse or experimented with other kinds of modern English verse. However, most of them share a common aim: to write English poetry. No one can object to such a laudable aim, but the question is, how far should one go in naturalizing Chinese poetry, with its different linguistic structure, different modes of thinking and feeling, and different ways of expression? Should one turn unfamiliar concepts and attitudes, exotic imagery, recondite allusions, and ambiguous syntax into familiar and easily comprehensible ones? Surely it is going too far to translate shunü (“virtuous girl”) as “Nymph” and junzi (“lord’s son” or “young lord”) as “Shepherd,” as one recent translator did,⁶ so that a Greco-Roman world is substituted for the ancient Chinese one? By the same token, should one not also substitute “rose” for “lotus,” “golden tresses” for “cloudy hair,” and “Diana” for “Chang’e”? Is it too much to expect readers of translations of Chinese poetry to learn the symbolic values and emotional associations of images and the significance of allusions? If all that such readers want is English poetry of a familiar and conventional kind, why should they bother to read translations of Chinese poetry at all?

As far as verse form is concerned, David Hawkes has put the case for naturalization well:

Of course, no translation from the Chinese, whatever the genre, can be formally like the original. Five-stressed verses in sprung rhythm are no more like Chinese pentasyllabics than iambic pentameters or stanzas in common metre. A translator can only choose or invent a form in which he can best express to his own satisfaction the “feeling” that he is given by the original. The measure of his success has very little to do with the number of formal elements that the original and the translation have in common.⁷

Granted that the “feeling” one gets from any verse form in English can never be demonstrably “like” that which one gets from a Chinese verse form, it still seems to me that the translation should bear some formal resemblance to the original. For instance, to turn a compact antithetical couplet from a Chinese poem in Regulated Verse into several loosely constructed lines of free verse is to change the character of the verse completely, but to turn it into two compact and paratactic lines of English verse would preserve at least some of the original “feeling.”

Not all poet-translators and those who aspire to be such are naturalizers; some of them, from the Imagists to the “Syntaxists” (those who believe that translations of Chinese poetry should follow the original syntax), are barbarizers, in practice or in theory, if not both. There are two main arguments for barbarization: first, that it will revitalize the English language and initiate a new kind of poetry and a new poetics; second, that it is only by barbarization that one can preserve the true nature of Chinese poetry. Since we are concerned here with the interpretation of Chinese poetry and not with its effect on modern American or British poetry, we need not consider the first argument but have to consider the second.

I have no wish to deny that, as Wai-lim Yip and others have asserted, Chinese poetry enjoys a high degree of syntactic fluidity, or that a translator can unwittingly impose a Western mode of thought and perception on Chinese poetry by introducing such grammatical features as tense, number, conjunctions, articles, and pronouns, which are absent in the original.⁸ After all, it was I who first suggested that the absence of tense, number, subject, and so on gave some Chinese poetry an impersonal, timeless, and universal character.⁹ However, I think we should not exaggerate this aspect of Chinese poetry, or rather, we should not pay attention exclusively to this kind of Chinese poetry. There are intensely personal, as well as narrative and even discursive, poems in Chinese. Similarly, we should not exaggerate the intuitive and nonrational character of Chinese thought: one cannot ignore the long tradition of Confucian rationalism, which found expression in some poems. Furthermore, even the most intuitive kind of poetry, which gives the impression of being impersonal, timeless, and universal, cannot literally present what Yip calls “pure experience” or “pure Phenomenon” unmediated by language, because language is mediation between the speaker and the world. A poet may wish to create the illusion that the reader is in direct contact with experience or phenomenon, but actually no one can write poetry without some degree of conceptualization. On the practical level, we may formulate the question thus: is it either absolutely necessary or always possible to follow Chinese syntax in translations of Chinese poetry, so as to preserve its underlying mode of perception?

The answer to the first part of the question is that it depends on what semantic and poetic functions the original syntax fulfils. Sometimes it is necessary to keep the original syntax, such as in the line

which first presents a sense impression, then locates it, and finally identifies the object that causes it, so that the reader feels as if he were experiencing the scene directly. The effect of the original can be largely preserved if we translate it as

Green, green: the riverside grass,

where the only concession to English grammar and idiom, the addition of the definite article, does not alter or detract from the original effect. But if we translate the line as

The grass by the river is very green,

then it becomes an assertion, and the reader will feel that he is being told something instead of experiencing it. In other cases, however, the semantic and poetic functions of a line can be preserved, or even better served, by syntactic changes in the translation. The last line of Du Fu’s “Stone-moat Village Officer,” which was quoted in the preceding chapter, reads in the original:

If we should translate this as

it would not bring out the point that the speaker said goodbye to the old man alone because the old woman had left, apart from being unidiomatic. That is why I translated it as

And said goodbye to the old man alone.

Irving Lo’s version, “Only the old man was there to see me off,”¹¹ can be justified on similar grounds.

Indeed, it can be positively misleading at times to follow the original syntax and word order. For example, the phrase shan shang ren means “man on mountain”: to follow the original word order and translate it as “mountain on man” would turn it into nonsense.

If a critic-translator provides the reader with the original text, transliterations, and word-for-word versions, as various recent translators, including myself, have done, then even those who cannot read Chinese will be able to see what the original syntax is, and some concessions to English grammar and idiom can be made in a more “readable” version.

As for whether it is always possible to reproduce Chinese syntax, the answer is no, for sometimes a line in Chinese can be syntactically construed in several ways simultaneously, and no translation can preserve such syntactic ambiguity. A famous example is Wang Wei’s

which can be construed as “sun’s color chills green pines,” or “sun’s color chills among green pines,” or “sun’s color is chilled by green pines.” The poet is not concerned with the question which object is chilled by which, but with presenting a single experience that fuses the impressions of pale sunlight, chilliness, and green pines. In the original, the word leng (“chill”) has a pivotal role, pointing both forward to qing song (“green pines”) and backward to ri se (“sun’s color”).¹³ It is simply impossible to do this in English: one has to decide whether “chill” is a transitive verb or an intransitive one, whether it is active voice or passive voice, whether it is a verb or an adjective, and whether the line is a sentence or two noun phrases. In such cases, the critic-translator can only describe what he cannot reproduce.

Earlier on I referred to the absence of tense and number in Chinese poetry; one or two points in this connection should be clarified. By the “absence of tense” I meant, of course, the absence of tense inflections in classical Chinese verbs, not the absence of any awareness of temporal relations between what is spoken of and the time of speaking. If we accept the definition of tense given recently by a linguist as “the semantactic category that establishes the relationship which holds between the situation or event talked about and the time of utterance,”¹⁴ then we obviously have to admit that tense is present in at least some Chinese poems, with the proviso that “time of utterance” must be understood in a double sense: from the author’s point of view, it refers to the historical moment when the poem was written, but from the reader’s point of view, it refers to the time of reading, when the reader, identifying with the speaker of the poem or at least imagining the speaker as speaking, revives the moment of writing. Thus, when reading a poem, we have to take any temporal deictic term there may be as if it referred to our present rather than the author’s. For example, the expression jinri (“today”) has to be taken as if it referred to our “today,” even if the author or some conscientious commentator tells us the exact date when the poem was written. The historical date of writing can, of course, be relevant to the interpretation of the poem, but at the moment of reading we have to temporarily forget that the poem was written so many years ago, but feel as if that date were now. It is a paradox that although all poetry is timeless in one sense, as I suggested in the introduction, in another sense poetry explores different ways of perceiving time. In fact, an enquiry into the way or ways in which a poet orients himself to time in a given poem provides one of the most helpful means of entry into the world of that poem, as we shall see in the final chapter.

As for the “absence of number,” this refers not to the absence of numerals but to the fact that Chinese nouns do not have singular and plural forms. It is, of course, perfectly possible to indicate number if desired. For instance, in the Quatrain “Sitting Alone at Jingting Mountain,” Li Bo uses several words to specify number:

The contrast between zhong (“many, multitudinous, crowd, flock”) and gu (“solitary”) emphasizes the contrast between the multitude of birds and the solitude of the cloud. This probably symbolizes the contrast between the common crowd of people and the solitary speaker himself. The solitude of the cloud (speaker) is further emphasized by the word du (“alone”). This sense of solitude is attenuated by the feeling of communion with the mountain, which is brought out by the words xiang (“mutually”) and liang (“two, both”). Yet some feeling of isolation remains, as suggested by the word zhi (“only”): the speaker and the mountain may have each other for company but they are cut off from everyone else.

Sometimes one can tell the number of a noun from the context or from conventional associations. For example, any experienced reader seeing the word yuanyang (“mandarin duck”) would assume it to be dual, since the male and female of this species conventionally symbolize husband and wife or a pair of lovers. More often, Chinese poets leave number unspecified because it is poetically irrelevant. Despite the objection that Burton Watson raised some time ago,¹⁶ I still think that in Wang Wei’s lines

it is of no consequence whether we take niao (“bird”) as singular or plural.¹⁸ In any case, I do not believe that poetic imagery has to be visualized in order to be effective. How are we supposed to visualize the following?

And even when number is specified, we need not visualize the image. When Romeo says of Juliet’s eyes:

are we supposed to visualize two big eyes twinkling in the sky? In Chinese poetry, the effect of imagery often depends on symbolic significance and emotional associations rather than visual appeal. Besides, as Roman Ingarden pointed out, it is in lyric poetry that “spots of indeterminacy,” which need to be filled in by the reader’s imagination, are of the greatest importance.²¹

What has been said above about syntax also applies to other aspects of the linguistic structure of a poem, such as prosodic features. Some prosodic features of Chinese verse, notably tonal patterns, simply cannot be reproduced in English. Other features like alliteration and rhyme can be reproduced, but it is questionable whether they have the same effect in English as in Chinese. End-rhyme, which is ubiquitous in classical Chinese poetry except for a few very early poems, can be reproduced in English, but this is often achieved at the cost of distortion of meaning, unnatural inversions, omissions, or padding. What is more, English rhymes, especially masculine rhymes in couplets, tend to have a jingling and comic effect, which is not the case with Chinese rhymes. That is why so many rhymed translations of Chinese poetry sound like doggerel. (It is, of course, an entirely different matter when a translator of Chinese fiction or drama deliberately uses doggerel to indicate the poor quality of the original.) This does not mean that translations of Chinese poetry should never use rhyme, but only that one should consider what effect the original rhymes produce, and experiment with rhyme as well as other devices like slanting rhyme and assonance to achieve a similar effect. But however successful he may be in this respect, a critic-translator would still feel obliged to describe the prosodic features of the original poem and their poetic effects.

Whereas translators of Chinese poetry have always paid great attention to imagery and, more recently, to syntax, few have paid sufficient attention to two other aspects of poetic language: the level of diction and the poet’s general tone in a poem.²² Chinese poets did not employ a uniformly “poetic” diction, but varied from the archaic to the colloquial. Even the same poet may use different levels of diction according to the theme and genre of the poem. True enough, as Burton Watson has observed, that it is difficult to know how far apart the literary and spoken languages may have been at any given period, but we do have some notion as to the degree of formality or colloquialism of the language of a poem. When Cao Cao (155-220) wrote in the meter of the Book of Poetry and incorporated lines from it verbatim in his own poetry, we can be quite sure that he was not writing in the colloquial language of his time. I am not suggesting that classical Chinese should be rendered into Anglo-Saxon or pseudoarchaic English, but only that translations should reflect to some extent the degree of formality or colloquialism, elegance or plainness, sophistication or simplicity, of the original. Presumably, contemporary English is capable of such distinctions, and not all contemporary American and British poets write in a uniformly colloquial style. Hence, to be noncolloquial is not necessarily the same as being noncontemporary, as some recent translators appear to think. Such translators would condemn any translation that does not use a strictly contemporary colloquial idiom as pedantic or pompous, without asking whether the original is formal, or colloquial, or consciously archaistic. Similarly, a translator should consider the tone of the poem: whether it is solemn, or lighthearted, or intimate, or ironic, and so on. The translation should adopt a similar tone and not aim at a “poetic” tone of its own. One should not forget that it is the interactions among various linguistic elements that produce the total effect of a poem, and although these elements are not all equally translatable, some consideration should be given each.

In brief, the critic-translator stands in the same relation to his reader as the original author stands in relation to the critic-as-reader. The critic-translator’s task is to bridge the gulf between the world of the poem and the world of the reader, and to describe, if he cannot reproduce, the original linguistic structure that embodies the poetic world. Since the kinds of readers for whom the critic-translator writes are presumably seeking knowledge and not just pleasure, it is more important for the critic-translator to instruct than to please. To him, translation is not an end in itself but part of the interpretative process. Finally, it may be added that there can never be a definitive translation of any poem, just as there can never be a definitive interpretation of any poem. However, this does not mean that all versions and perversions are equally acceptable, any more than that all interpretations and misinterpretations are equally valid.