Article body

1. Setting the Context

The use of rhyme to translate classical Chinese poetry into English is an unresolved aesthetic issue. This topic remains relevant today in that while most native English translators have moved away from rhymed translations (as exemplified in the past by H. A. Giles, John Turner, etc.), many native Chinese translators (especially those in mainland China) continue to turn classical Chinese poems into rhymed English. In part, these contrary practices are rooted in a disparity of aesthetic sense. Since English rhymes less naturally than Chinese, English translators may feel no great sense of loss in giving up rhyme, but as virtually all classical Chinese poetry is rhymed, Chinese translators may feel a strong need to retain rhyme in poetry. Xu Yuanzhong typifies this view when he argues that “the best way to regain poetry is to recreate it” through “creative transformation”: “a poetic translation should be as beautiful as the original in sense, in sound and, if possible, in form” (1987, pp. 5‑6).[1]

Perhaps an early distinction needs to be made between the function of rhyme in writing and in translating poetry: the one creates synergic aural effects out of the source language sound system; the other tries to transpose these effects to a target language with a different phonetic matrix and prosodic system. Lefevere states that “[a]lthough the arguments against translating poetry into rhymed and metered verse are persuasive, rhyme can play an important part in the original poem: it marks a completion, a rounding of the line, and acts as a further ‘marker’ in the development of the poem as a whole. Furthermore, the sound effects produced by the succession of rhymes undoubtedly heighten the illocutionary power of the poem” (1992, p. 71). At best a synergic fusion of sense and sound, rhyme is a key element of poetic music that serves salient functions. Structurally, it is at once a dividing and integrating factor, distinguishing units within a poem and linking them into an identifiable whole. Semantically, rhyme can be “a principal means of elaborating or ironising sense” (Lennard, 2005, p. 190), depending on how sound and sense enhance each other in different languages. Emotively, rhyme recurrences set up pleasing resonances that deepen a poem’s artistic appeal. And prosodically, regular rhyme echo “affords the satisfaction of the return to the keynote in a melody” (Deutsch, 1965, p. 117): rhyme is a modulator of pace and rhythm, while rhyme change—which may include both phonetic and tonal change in a language like Chinese—can be quite useful in marking a turn of sense and rhythm in a long poem.

At the same time, the significance and operation of rhyme in versification varies among languages. Moving beyond general observations and experiential insights currently available, this essay will present some empirical evidence on the rhyming resources and practices of English and Chinese in order to cast new light on the linguistic and aesthetic issues involved in using rhymed English to translate classical Chinese poetry (problems in using rhymed Chinese to translate rhymed English poetry will not be discussed here). It should be noted that in this essay, rhyme refers to end rhyme occurring at the end of verse lines. In a largely monosyllabic language like classical Chinese[2] where rhyme works easily and cleanly, it is not a “[mere] convention” that “lines of verse rhyme at the end rather than at the beginning” (Fraser, 1970, p. 60), for end rhymes evoke deeper effects of linkage and resonance by virtue of the finality of their position. The case is less compelling for syllabically irregular English, where rhyming is not as clean and as natural an artistic device in poetry composition.

2. Rhyming Resources of English

One may begin with some experiential insights offered by two distinguished English translators of classical Chinese poetry. Arthur Waley has stated that “rhymes are so scarce in English (as compared with Chinese) that a rhymed translation can only be a paraphrase and is apt to fall back on feeble padding” (1962, p. 9). Similarly, David Hawkes has observed that “it is impossible to use the same rhyme for very long in English without running into serious difficulties, and at the same time introducing a heavy emphasis into the rhyming word which is not present in the Chinese. Moreover, the effort of sustaining a rhyme in English verse generates a tension which often finds relief in laughter” (1964, p. 99). Such statements call for further inquiry and testimony. In English versification, rhyme “consists of the repetition, in the rhyming words, of the last stressed vowel and of all the speech sounds following that vowel: láte‑fáte; fóllow‑hóllow” (Abrams, 1999, p. 273), with a “difference in the consonantal sounds that immediately precede the accented vowel sound” (Allen and Cunningham, 1998, p. 1). The most common types of rhyme are one‑syllable (single/masculine) and two‑syllable (double/feminine) rhymes (ocean‑devotion). Three‑syllable or triple rhymes (lyrical‑satirical) and even longer rhymes are possible, but their effects are better suited to light or humorous than to serious poetry (e.g., comparison‑garrison in Byron’s Don Juan, Canto I, 17). Similarly, the fashioning of mosaic or composite rhymes involving two or more single words (see us‑flee us, rowing now-snowing now) is generally more appropriate to light verse.

It has been noted by the editors of an English rhyming dictionary that rhyme is “an alien importation” “not native to English,” “which is often hostile to rhyme” (ibid., pp. 832, 853). This may sound like a stern judgment, but most critics will agree that “English is less rich in rhymes than many other languages”: for instance, “love in English [rhymes] perfectly only with the undignified word shove or the trivial word glove (itself a disguised rime riche). The word God… has its aptest rhymes in odd… and sod” (Fraser, 1970, pp. 61‑62).[3] Clement Wood’s Complete Rhyming Dictionary (1936) lists about 60,000 words thinly spread over about 7,800 rhyme groups, which average out at eight words per group. As one can expect, double and triple rhyme groups contain even fewer words than single rhyme groups.[4]

Given the limited manoeuvrability of rhyming in English, many overused rhymes soon become hackneyed (trees‑breeze, fire‑desire). In fact, it is easy to see that in a language with inadequate rhyming resources, rhyme is an especially forceful and restricting determinant of expression. Once a poet or translator has settled on a key rhyming word, the choice of final words for the rhyming lines is confined to a small group of rhyming mates. The verbal artist directs his/her thoughts and their expression towards a fitting (and hopefully natural) use of these words, working within self‑defined limits that at once inspire and constrain the creative imagination. One may contend that where rhyming resources are limited as in the case of English, the translator (following the poet’s cue) can claim a degree of “poetic license” to fashion lexical groups or twist language out of its normal grammatical order for the sake of rhyme. On the other hand, there is the constant danger of stretching this license to concoct too many unnatural inversions and strained formulations,[5] which will result in a forced, artificial style bound to affect artistic quality. Though clever niceties may impress for a short while, one will recall that in the living language of Shakespearean drama, the non‑rhyming blank verse lines contain few instances of warped construction or ungrammatical poetic license.

One added complication arising from the syllabic irregularity and limited rhymes of English is that even in using single rhyme, accented syllables often have to be rhymed with unaccented or weakly accented syllables, e.g., sea‑harmony, ring‑wedding. While such praxis is accepted in English versification if not done too often, the rhyming effect is in fact diluted if the words are read naturally, and a drastic twisting of the normal word accent may become necessary to preserve the aural resonance. Take for instance a few verse lines in Christopher Marlowe’s (1564‑1593) “The passionate Sheepheard to his love”, where natural word accents have to be wrenched in order to make the rhyming and metric patterns work:

2.3

By shallow Rivers, to whose falls,

2.4

Melodious byrds sings Madrigalls....

6.1

The Sheepheards Swaines shall daunce and sing

6.2

For thy delight each May-morning,...

This excerpt shows both couplets (2.3/2.4 and 6.1/6.2) rhyming. As the closing feet of lines 2.3 and 6.1 end on a stressed syllable (masculine ending) while the closing feet of lines 2.4 and 6.2 end on an unstressed one (feminine ending), secondary accents have to be artificially added to the final syllables of madrigalls and morning for the rhymes to work. The problem is smaller for madrigalls, which is a dactyl (one stressed syllable followed by two unstressed syllables) that allows a minor accent, thus turning the word into an amphimacer that keeps the basic iambic metre. On the other hand, morning cannot be read naturally as morNING, and any attempt to turn the trochee into a spondee will produce a triple‑accent molossus sequence that upsets the iambic rhythm. Though this is “conventional in the folk ballad (for example, ‘fair ladíe,’ ‘far countrée’), and is sometimes deliberately used for comic effect” (Abrams, 1999, p. 160), accent wrenching is not a satisfactory and sustainable way of rhyming serious poetry.

As a matter of fact, despite the considerable length of some English rhyming dictionaries, the number of usable rhymes is materially reduced by at least three factors. First, English words can end in a large number of ways; in particular, the consonantal clusters (rather than open vowels as in Italian) in which English words often end make rhyming mates rather difficult to find.[6] Thus ordinary words like bulb, wasp, film, gulf, wolf, eighth, width, breadth, depth, month, warmth, lounge, orange, revenge, plagued, spoilt (with two post‑vocalic consonantal phonemes), sixth, twelfth, amongst (with three post‑vocalic consonantal phonemes) have no rhymes.[7] Second, the resulting dispersion of the lexis over a wide range of rhyme groups means that many rhymes consist of only a handful of items that are quite inadequate for literary use, e.g., calm, balm, palm; arm, charm, harm (note that while the two groups of words belong to one rhyme in British English, the articulated /r/ turns them into two rhymes in American English, which further reduces the availability of rhyming mates). Indeed, there are a considerable number of words that either have no rhyme or that rhyme only with one word (mountain‑fountain). Third, identical‑sounding syllables (or rime riche[8]) are not acceptable rhymes regardless of spelling or meaning: bay‑obey, stair‑stare, well‑well (noun‑adjective) are seen as identities and not true rhymes, since the consonantal sounds preceding the accented vowels must differ.[9] This means that even rhyme groups with long lists of member words generated from inflectional suffixes (e.g., -es, -ing for verbs, -es for nouns) or derivational suffixes (e.g., -al, -ism, -ity, -ive, -ize, -ness, -ous, -y, etc.) are more restrictive than they appear on the surface.[10] For as noted above, if a polysyllabic word with a dactylic ending is put in a rhyming position, the second unaccented syllable tends to be given a secondary accent. Thus a suffix serving as a rhyming syllable cannot take other words ending in the same suffix as rhyming mates (e.g., loveliness‑laziness do not work together).

The relative paucity of rhyme words makes rhyming a less natural feature in English versification compared to other languages; it also entails a long‑term strategy of employing multiple rhymes even in a short poem. But since even then it remains difficult to work with only true rhymes, various types of near rhymes have been increasingly conscripted in place of true rhymes since the beginning of the twentieth century in order that the sound‑repetition resources of the language may be enriched.[11] A. C. Graham, for one, points to “the part‑rhymes to which the English ear has grown accustomed since Wilfred Owen’s experiments in dissonance” (1965, p. 24). One type of near rhyme is assonance or vowel rhyme, with identity of the final accented vowel sound and difference in the following consonant and/or vowel sounds, e.g., main‑game (/mein/‑/geim/), plenty‑trendy (/plenti/‑/trendi/), silver‑deliver (/silvə/‑/dilivə/).[12] A yet more widely used type of near rhyme is consonance, with difference in the accented vowels and identity of all the following consonant and vowel sounds, e.g., given‑heaven (/givn/‑/hevn/), shadow‑meadow (/ʃaedou/‑/medou/).[13] Both consonance and assonance may appear visually as eye rhymes, words with rhyming syllables spelt alike yet pronounced differently, including words long accepted as rhyming mates such as earth‑hearth (/ə:θ/‑/hɑ:θ/) or love‑move‑prove (/lΛv/‑/mu:v/‑/pru:v/).[14] Witness also Shakespeare’s (1564‑1616) Sonnet 18:

Shall I compare thee to a summer’s day?

Thou art more lovely and more temperate.

Rough winds do shake the darling buds of May

And summer’s lease hath all too short a date.

Phonetically and aesthetically, temperate (/tempərət/) and date (/deit/) are consonant eye rhymes rather than true rhymes. A similar case is seen in Byron’s (1788‑1824) use of maid‑staid‑said in Don Juan (Canto V, 18). In fact, eye rhymes can be wide apart in pronunciation, like the assonant finger‑singer‑ginger (/fiŋgə/‑/siŋə/‑/ʤinʤə/), or rough‑through (/rΛf/‑/θru:/), which are not even close enough to qualify as consonance or assonance.[15] Other pairs of words such as bare‑are (/bεə/‑/ɑ:/), real‑steal (/riəl/‑/sti:l/) are hardly more satisfying, since the first word in each pair is disyllabic, while the second word is monosyllabic.

Although rhyme is a consciously adopted device, it should ultimately convey an effect of unobtrusive naturalness in serious poetry, or else it will undercut its artistic effect on the deeper sentiments. Obvious rhyming will draw attention to its own technical cleverness and detract from its proper appeal; strained or irrelevant words forced by rhyming needs are even more dearly bartered. No doubt the adoption of consonance, assonance and eye rhymes in English poetry is an attempt to extend the resources of rhyming, but such efforts also reflect a basic shortage of rhyming facilities in the language. This may have been why Wood thinks that “rhyming is an even more unnatural convention of poetry than meter” (1936, p. 20).

3. Rhyming Resources of Chinese

Wood’s judgment does not hold the same validity in Chinese poetry. While rhyme can be “a shelter for mediocrity” (ibid., p. 22) in versification in any language, it has deeper intrinsic aesthetic functions in Chinese than in English. Forming easily in Chinese and far more than a mnemonic tool or an adorning device, rhyme is a natural way of enhancing the condensed affective expression that is poetry. While rhyme did not form the basis of ancient Greek and Latin poetry[16] or the earliest English poetry (e.g., Beowulf, 8th century),[17] it has been a staple element of Chinese poetry from its beginnings in the second millennium BC, or about three millennia before the same feature took root in English poetry.[18] Whereas the latter needs to use different rhymes due to a paucity of rhyme words, Chinese poems (except for very long ones) typically employ one rhyme: “ancient‑style” poems using 10 characters from the same rhyme group for 20 verse lines are fairly common, and poems of 30 or more lines using 15 or more characters from the same rhyme cannot be called rarities.[19] In addition, it may be noted that while “mono‑rhyming” Chinese can always choose to diversify its poetic resonances by using multiple rhymes, “poly‑rhyming” English will be hard put to sustain the same rhyme for very long without running into difficulties or producing unintended effects.

The importance of rhyme in Chinese poetry can also be attested by the fact that Chinese rhyming dictionaries have existed at least since the 3rd century.[20] A full account of Chinese rhymes—including the diverse artistic effects of the different tones[21] amplified in rhyming position—is beyond the scope of the present inquiry, but a sketch of some key developments will be useful here. Thus, Lu Fayan’s 陸法言 Qie yun《切韻》 (with a preface dated 601), the rhyming basis for the golden era of Tang (618‑907) poetry, listed about 12,000 characters distributed among 193 rhyme groups, according to an official literary sound system (Middle Chinese) that took into account diachronic (inherited) and synchronic (dialectal) factors (Li, 1985, pp. 52‑56). During the High Tang, Sun Mian 孫愐 expanded Qie yun into Tang yun 《唐韻》 (with two editions finished in the 730s and 750s), which served as the standard rhyming reference from the mid‑Tang up to the early Song (960‑1279), with about 15,000 characters in 195/204 rhymes and certain groups combinable into larger ones.[22] Guang yun 《廣韻》, the oldest extant rhyming dictionary, compiled in 1008 by Chen Pengnian 陳彭年 (961‑1017) and his team, assembled 26,194 characters distributed among 206 rhymes combinable into 113 groups; an abridged version of an expanded edition of Guang yun compiled in 1037 (《禮部韻略》) removed scores of obscure and rarely used items, trimming the vocabulary to 9,590 characters.[23] By the first half of the 13th century, rhyming dictionaries had largely settled on a framework of 106 rhyme groups (平水韻), which remained the norm through the succeeding eras of imperial China. For instance, the authoritative Peiwen yunfu 《佩文韻府》, finished in 1711 by Zhang Yushu 張玉書 (1642‑1711), Chen Tingjing 陳廷敬 (1639‑1712) and others under Emperor Kangxi’s 康熙 (r. 1662‑1722) edict, had a vocabulary of 10,258 characters distributed among 106 rhymes,[24] with the “level‑tone” groups containing more entries than the three “oblique‑tone” rhymes, especially the “rising‑tone” and “departing‑tone” rhymes (Wang, 1979, pp. 323, 348). Mixed at times with dialectal sounds in praxis, this work remains the standard reference for writers of classical Chinese poetry up to the present day.[25]

It should be noted that while the phonetic qualifications for rhyming are basically the same in Chinese as in English, there is an element of flexibility built early into the standard rhyming system of classical Chinese poetry. Now the Middle Chinese syllable is made up of an initial (consonant) plus a final. Rhyme concerns only the final, which structurally consists of one or two medial vowels, a nuclear vowel and an ending, or any combination of the three elements.[26] In the early Tang, some officials proposed that certain close‑sounding rhymes be combined in praxis to ease the undue constraint imposed by a strict adherence to minutely differentiated rhymes (Wang, 2003, p. 54). Thus in due course rhyming in classical Chinese poetry slightly relaxed to the point of requiring the identity or close similarity of the nuclear vowel and the following sounds: the medial does not matter (e.g., /-a/, /-ua/, /-ia/ belong to one rhyme), and in a few exceptional cases a small degree of imperfect rhyming (e.g., /-eu/, /-iεu/; /-am/, /-iaem/)[27] is accepted within the standard rhyme scheme (Tang, 2002, pp. 43‑45; Wang, 2003, pp. 22‑23) to facilitate poetic expression. This has no doubt helped to liberate, stabilize and sustain rhyming in classical Chinese poetry from an early stage.

It may not be very realistic to rely on the figures in comprehensive rhyming dictionaries to measure the facility of rhyming in classical Chinese poetry, for they are rather thick works that include many obscure and rarely used characters. In terms of gauging a practical sense of utility, I will take two slimmer, more easily available reference works to give a concise and faithful picture of the availability of classical Chinese rhymes. I refer to Yu Zhao’s (fl. 1800) rhyming dictionary Shiyun jicheng,[28] and an even more highly distilled list of “commonly used characters” given in Wang Li (1900‑1986)’s Gudai Hanyu.[29] In addition, since the most widely written form of classical poetry since Tang times is “regulated‑style” verse (including quatrains) that almost always uses level‑tone rhymes, the summary illustration here can be further condensed by focusing on the number of characters listed in the 30 level‑tone rhyme groups in the two works. The traditional division of the rhymes into Parts I and II is not relevant, and the 15 rhymes in Part II will be numbered R16‑R30 for the sake of clarity:

Table

-> See the list of tables

It can be seen that the character count of Wang’s list is less than half (44.5%) of that in Yu’s dictionary: the average number of rhyming characters per group is 143.6 in Yu’s work (range 41‑455, median 116.5) and 63.9 in Wang’s list (range 12‑171, median 54.5). Eight groups are generally categorized as “broad” rhymes (A), seven as “narrow” rhymes (C), four as even less manoeuvrable “perilous” rhymes (D); the other eleven are seen as “neutral” rhymes (B) (Wang, 1979, p. 44).[30] While the categorization of rhyme groups may not look entirely scientific in statistical terms (e.g., R1=A vs R17=B, R27=B vs R25=C, R15=C vs R18=D), it may be noted that apart from the raw number of characters available in each rhyme group—

  • Yu: A=170-455, B=69-183, C=62-112, D=41‑106

  • Wang: A=63‑171, B=40‑87, C=29‑56, D=12‑34[31]

—the characters’ semantic content, in terms of its affinity with life experiences and usability within the lyric horizon, also bears on the overall perception of what constitutes a broad or narrow rhyme group. The judgment is quantitative as well as qualitative, and relative if taken cross‑culturally; in some cases a narrow Chinese rhyme may already be the equivalent of a broad rhyme in English. And it should not be forgotten that the tonal language of classical Chinese poetry, i.e., Middle Chinese, consists of four tonal categories that amply extend the possibilities of rhyme (especially in ancient‑style poetry) and the potential synergy between poetic sense and poetic music; this is a prosodic aspect which non‑tonal English is not equipped to cover adequately in translation. In sum, it can be stated with a credible degree of clear evidence that Chinese rhymes more easily and naturally than English.[32]

Yet one is still obligated to ask: are there restrictions on rhyming in Chinese comparable to those in English as outlined in the previous section? First, classical Chinese has a simpler syllable structure than English: syllables do not end in consonantal clusters that are hard to match phonetically, but in open vowels or single consonants (/-m/, /-n/, /-ng/, /-p/, /-t/, /-k/).[33] Second, and partly for this reason, Middle Chinese has a far smaller number of rhyme groups (106) than English (7,800 in Wood’s rhyming dictionary), and so does not spread its lexical resources thinly in terms of rhyming potential. There is extra flexibility in rhyming if needed: in ancient‑style poetry “neighbouring rhymes” (e.g., R1/R2, R6/R7, R11/R12, R14/R15/R16) within all tonal categories can rhyme with one another, and in addition, corresponding rising- and departing‑tone syllables may also rhyme with one another;[34] in regulated-style poetry, characters from consonant neighbouring rhyme groups can likewise be used for the optional rhyme in the first line. Such neighbouring rhymes have a phonetic affinity and prosodic unity generally higher than consonance in English (which allows any degree of difference in vowel sound): Wang (1979, pp. 331‑334) observes that the 30 rhyme groups form 15 clusters for the purpose of exercising the added freedom offered by the use of neighbouring rhymes, but even within a cluster not all characters can work as neighbouring rhymes due to varying degrees of phonetic difference. This self‑imposition also shows that Middle Chinese has ample rhyming resources to utilize—as does the fact that classical shi poetry rhyme does not have to rely on the weaker resonances offered by assonance.[35] Third, in syllabically irregular English, rhyme is concerned with sound rather than meaning,[36] syllables rather than words (e.g., eight rhymes with -late in relate), which may be one reason why identities are not accepted as rhymes. But in monosyllabic classical Chinese, rhyme takes account of sound and meaning at the same time;[37] identities are perfect rhymes, and the only restriction is that a word cannot rhyme with itself. All these factors make rhyming a much easier matter in classical Chinese poetry.

One added reason why Chinese rhymes with such clean resonance is that grammatically it is an isolating and analytic language with little inflectional morphology; “most words in Chinese have one immutable form, which does not change according to number, case, gender, tense, mood or any of the other inflectional categories familiar from other languages” (Li and Thompson, 1990, pp. 824‑825).[38] Classical Chinese thus remains basically monosyllabic morphologically (just as modern Chinese remains largely monosyllabic and disyllabic), free from problems of unevenly accented rhyming or wrenching inherent in English poetry. Thus even if one is able to produce a rhymed English translation without incurring the problem of wrenching, the nature of the Chinese morpheme makes for a cleaner and more cogent rhyming effect than, for example, having a monosyllable rhyme with a polysyllable in English.[39]

In sum, the monosyllabic, isolating‑analytic and tonal nature of classical Chinese, its simple syllabic structure that facilitates a concentration of rhyme groups, all constitute significant prosodic and morphological differences from English. They suggest that Chinese poetry is intrinsically better suited to rhyming than English versification. It is no surprise that no English rendition can rhyme as naturally as an original Chinese poem and achieve the same prosodic resonance that is part and parcel of the latter’s artistic appeal.[40]

4. Rhymed English Translations of Classical Chinese Poems: Examples and Effects

To flesh out the conceptual observations made above, it will be useful to examine some rhymed English translations of classical Chinese poems. The five renditions below were produced by accomplished translators whose command of the target language is beyond question:

 

P1:贈別 杜牧

多情卻似總無情

唯覺樽前笑不成

蠟燭有心還惜別

替人垂淚到天明

T1: Old Love Du Mu

Old love would seem as though not love today:

Spell‑bound by thee, my laughter dies away.

The very wax sheds sympathetic tears

And gutters sadly down till dawn appears.[41]

 

P2:送元二使安西王維

渭城朝雨浥輕塵

客舍青青柳色新

勸君更盡一杯酒

西出陽關無故人

T2: Seeing Off Yuan Second on a Mission to An[x]i Wang Wei

The light dust in the town of Wei is wet with morning rain;

Green, green, the willows by the guest house their yearly freshness regain.

Be sure to finish yet another cup of wine, my friend,

West of the Yang Gate no old acquaintance will you meet again![42]

 

P3:靜夜思 李白

床前明月光

疑是地上霜

舉頭望明月

低頭思故鄉

T3A: Thoughts on a Silent Night Li Bai

Before my bed a pool of light—

Can it be frost upon the ground?

Eyes raised, I see the moon so bright;

Head bent, in homesickness I’m drowned.[43]

T3B: Night Thoughts

As by my bed

The moon did beam,

It seemed as if with frost the earth were spread.

But soft I raise

My head, to gaze

At the fair moon. And now,

With head bent low,

Of home I dream.[44]

 

P4:贈孟浩然李白

吾愛孟夫子

風流天下聞

紅顏棄軒冕

白首臥松雲

醉月頻中聖

迷花不事君

高山安可仰

徒此揖清芬

T4: To Meng Haoran Li Bai

O Master Meng my friend! How I love thee

Whose spirited ways to all the world are known!

White‑head’d thou seek’st to lie beneath pine tree.

As in fair youth thou spurned Rank and Gown.

Beneath the moon too oft thy cup thou’dst fill,

And be rather charmed by flowers than the King to serve.

Thy Virtue fragrant, like a lofty hill,

I can but homage pay that thou deserve.[45]

A minute analysis of the renditions or a comprehensive evaluation of the implications of verse translation will not be necessary here. Instead, my cursory remarks will merely focus on the use of rhyme in the renditions, a strategy essentially involving an exchange of sense and syntax for an aural effect that is the basic appeal of employing rhyme. Two translated versions of P3 are quoted for the sake of a more detailed case study below.

  1. The renditions adopt a strategy of using single rhymes (mostly monosyllabic words) based largely on iambic metre, in an attempt to keep the rhyming effect as clean as possible.

  2. The rhyme schemes of the translations are ababcdcd for the regulated‑style poem (T4, cf. abcbdbeb for P4), and for the quatrains, aabb (T1), aaba (T2), abab (T3A), and abaccddb (T3B, cf. aaba for P1, P2 and P3). Note that the original poems all employ one rhyme in the even‑numbered lines, with the option of rhyming the first line as well (as in P1‑P3). Thus apart from T2, which tries to follow the original rhyme scheme by employing one rhyme, the sample renditions all use multiple (2 or 4) rhymes to produce rhyming and general artistic effects different from those of the original poems (note the contrast between T1 and T3A in rendering the same original rhyme scheme). This point is especially worth noting in the case of the quatrain, since the unrhymed third line often effects a turn in the poem’s sense, sensibility or ambience: in P1 the candles embody an empathic deepening of the poem’s affective appeal; in P2 the farewell cups mark a transition from the descriptive first couplet to the emotive crescendo in the last line; in P3 the lifting of the eyes carries the shift from the descriptive‑narrative first couplet to the lyrical longing at the end. The artistic function served by the unrhymed third line is thus forfeited by rhymed renditions T1 and T3, which represent at once a tailored act of naturalization and an ironic process of foreignization. On the other hand, a rhymed translation of the T2 type, which preserves the artistic marker here by using one rhyme, is bound to run into problems of sustainability for longer poems.

  3. The renditions have taken liberties with sense and syntax in order to engineer rhymes:

    1. Semantic addition/padding: T1 – “today”, “dies away” (ll. 1, 2); T2 – “yearly… regain”, “again” (ll. 2, 4, which end up being 16 and 15 syllables in length); T3A – “drowned” (l. 4); T3B – “spread”, “soft”, “and now” (ll. 3, 4, 6); T4 – “that thou deserve” (l. 8).

    2. Semantic relocation: T1 – “tears” (from l. 4 to l. 3); T3B – whole quatrain irregularly expanded into 8 lines (l. 1→ll. 1‑2, l. 2→l. 3, l. 3→ll. 4‑6, l. 4→ll. 7‑8); T4 – reversal of l. 3 and l. 4, “thy virtue fragrant” (from l. 8 to l. 7).

    3. Syntactic/word order inversion (minor instances not noted): T2 (ll. 2, 4); T3B (ll. 3, 8).

    4. Use of archaic forms: T4 – “thee” (l. 1), which leads the rendition to use other archaic forms like “thou” (4 times) and “thy” (twice).

    5. Use of imperfect eye rhymes: T3B – “now/low” (ll. 6. 7); T4 – “known/gown” (ll. 2, 4).

In his study of various English translations of Catullus’s Poem 64, André Lefevere lists rhymed translation as one of seven different strategies employed, where the rendition “enters into a double bondage” of metre and rhyme to produce a “caricature” of the original (1975, pp. 49, 61). Lefevere’s stern conclusion need not be applied to the generally fine translations quoted here, but the skeletal observations above do point to various shades of awkwardness arising from the use of rhyme.[46] Needless to say, the main concern about the translations is not any surface “infidelity,” but how the liberties taken due to the self-imposed “bondage” actually affect or alter the poems’ artistic appeal. It should therefore be useful to examine one case in some detail for the sake of illustration. P3 will serve as a fitting touchstone, and the juxtaposition of two renditions will highlight some of the issues involved.

This most beloved Chinese poem is deceptively simple and effortlessly crystalline; Turner admits “[i]t is the most difficult poem I ever tried. It is so simple and natural that translating it is like trying to dye a rose‑leaf” (Turner, 1989, p. xxvi). Effortless simplicity and naturalness, soft touch and deep feeling are indeed the poem’s core artistic qualities and the yardstick by which the renditions are to be gauged. Like others in a similar vein, Turner explains that he regularly uses rhyme “in an attempt to preserve the singing or musical quality in Chinese,” whose “rhyming and metrical systems are more purely wrought and exquisite than any other” (ibid., p. xxv). This worthy goal often makes him go overboard in his rhyming efforts. First, the eight‑line rendition in T3B doubles the quatrain’s length in a laboured construction that is inconsonant with the magical simplicity of the original poem; as Graham observes, “the gift of terseness is the least dispensable literary qualification of a translator from Chinese,” or else “some of the sparest Chinese writers seem windbags when read in English” (1965, p. 19). Second, the rendition has altered beyond recognition the evocative pithiness and formal integrity of the original, since the quatrain is unevenly expanded partly to meet the exigencies of rhyme. Third, the rhyme engineering is unnatural and untidy as well: instead of fashioning a more regular and pleasing aural pattern (such as ababcdcd or aabbccdd, as per some of the rhyming couplets used in T3B), Turner’s rhymes are haphazardly placed among consecutive, alternating and even widely separate lines. The second pair of rhyme words end up being a remote six lines apart, for once the word “dream” is easily settled in line 8, the translator has to fashion a rhyming mate at a spot that comes closest to fitting the sense of the original—in this case “beam.” Fourth, the translator has to insert words like “beam,” “spread” and “and now” to construct rhymes, yet they are all terms of action with an assertive flavour that disturbs the soft, subdued tone on which the poem’s emotive depth depends. And when all is said and done, Turner admits that “rhyme presents its own special problems too… every time I looked at [my own translation], the Cockney rhyme of now, low hit me between the eyes” (Turner, 1989, p. xxv‑xxvi). The limited rhyming resources of English have cramped his valiant effort, and imperfect eye rhyme waters down the poem’s musical quality that he sets out to preserve (the same point applies to the use of known/gown in T4.)

Xu’s version (T3A) presents fewer problems at first sight: it is less arduous on the whole, and the formal integrity of the original is preserved. The translation shows a consistent effort to make the poem more “literary” than it is by removing the unusual repetition of three words in the original, i.e., by substituting “a pool of” for “bright moon” (l. 1) and “eyes” for “head” (l. 3).[47] In this case, however, the repetition (taking up 15% of the textual space) is part of the poem’s simple, unmediated lyric flow, adding to the unadorned spontaneity and purity of the sentiment, and thus should be preserved. In terms of the rendition’s rhyming, which differs from the original, the rhyme scheme does not allow the third line to function as an aesthetic turn (as noted above), but remains regular and pleasing to read in the target language. Like Turner, however, Xu has to insert new words and manipulate word order to make the rhymes work: “bright” is moved to the end of line 3, which calls for the addition of “so” to fit the iambics and thus to inject an unwanted tone of emphasis into the sense. Similarly, following the cue to use “ground” as a rhyme word, the translator ends up fabricating “drowned” that over‑dramatizes the poem’s sentiment of homesickness. Both additions impose a sense of laboured assertiveness that betrays the effortless grace and soft aura of the original, ultimately debilitating its natural appeal. While there can be no guarantee that any rendition will manage to transport the artistic magic across languages, gratuitous padding or laboured sculpturing is likely to weaken the effect further; the liberties taken with sense and syntax for the sake of hammering out rhyme (and iambics for T3A) are dearly exchanged.[48] Liberation from this self‑imposed bondage will at least allow an unrhymed translation to approach the crystalline simplicity of the original in tone, language and ambience with a better sense of balance:

Bright moonlight before the bed;
I thought it was frost on the ground.
Looking up, I see the bright moon;
Looking down, I think of home.[49]

Whether rhymed or unrhymed, translated poetry should read like poetry, which requires regard for the syntactic, cognitive and associative operation of the target language. This is partly why most poets and translators agree that translating a poem is invariably to rewrite it, but continue to find themselves stranded on the horns of a perennial dilemma. A “minimalist” approach of lexical translation with minimal syntactic interference, valid and sensible in intent, will run the risk of fragmentation when the target‑language version is actually read and felt.[50] On the other hand, the use of rhyme to rewrite the original will almost invariably recast it out of shape, especially when rhyme is not one of the stronger suits in English versification. Waley, the outstanding poet‑translator, has always considered it “impossible not to sacrifice sense to sound” (1941, p. 1). Lü Shuxiang notes three problems entailed by the use of English rhyme to render classical Chinese poetry: padding rhymes; inverting word order; adding, deleting and altering the original sense (Lü and Xu, 1988, p. 20). Lefevere writes in remarkably similar terms: “Translators who translate with rhyme and meter as their first priority often find themselves neglecting other features of the original: syntax tends to suffer most…, and the information content is almost inevitably supplemented or altered in none too subtle ways by ‘padding’: words not in the original added to balance a line on the metrical level or to supply the all‑important rhyme word” (1992, p. 71). Indeed, even faithful believers of rhymed translation acknowledge the pressure of rhyme‑padding at the inevitable expense of “harming the meaning,” and the need to minimize the damage (Ma, 2000, pp. 149‑160). While the profile and degree of the difficulties involved vary according to praxis, there is no doubt that they are intractable problems inherent in rhymed translation. James Liu testifies to a drastic change of heart in his rendition of classical Chinese poetry:

As for rhyme, I formerly advocated reproducing the original rhyme schemes in translation of Chinese poetry, and tried to put this in practice, with unfortunate results,… I now realize the virtual impossibility of keeping the rhymes without damage to the meaning, and no longer wish to insist on the use of rhymes. Thus two of the most important elements of Chinese versification, tone‑pattern and rhymes, have to go.

1969, p. 42

Not all poetry gets lost in translation. But if imagery is a poetic element that can traverse time, space and culture with relative ease, phonological attributes are not transportable across languages. A non‑tonal language like English can do nothing about Chinese tone‑patterns, and a language with limited rhyming resources can do little in translation to preserve rhyme in a poetic language with rich rhyming resources where rhyme functions easily as an integral element of poetic music. While there can be no total transfer from source language to target language under any circumstances, this absolute limitation does not justify the extra liberties which a translator must take with sense, tone and appeal in order to fashion rhyme. Yet the fact remains that the sensible option of free verse (or prose) translation forfeits the rhyming resonance and vivid rhythm of the original poem. Waley is speaking from his modern English sensibility when he claims that “lack of rhyme will not be generally felt as an obstacle” (1962, p. 9). For the source‑language reader, rhyme is the earliest and most basic aesthetic attribute in classical Chinese poetry, bringing an irreplaceable sense of unity and resonance to a poem: at once a dividing and binding element that marks off larger sense groups, rhyme unifies them prosodically into a musical whole. In the end, rendering rhyming Chinese into unrhymed English means writing off the inevitable loss without the recompense of a viable option. James Liu did not give up rhymed translation without a sense of regret.

5. Concluding Remarks

Poetry translation is the slippery, shifting act of seeking the most favourable “aesthetic exchange rate.” Among other things, translators have to judge whether the use of rhyme (and metre) will militate against other aspects of the poem—sense, tone, syntax, ambience and appeal. The merits and demerits to be weighed are partly general and partly specific, varying according to the source and target languages and poetics, sometimes even according to the original poet. It is worth noting that while English poetry itself has moved away from the use of rhyme, the use of rhymed English to translate classical Chinese poetry has shown greater resilience. One will no doubt continue to find rhymed English renditions of classical Chinese poetry,[51] and the debate over rhymed or free verse translation will go on.[52] Those who insist on using rhymed English to translate classical Chinese poetry are rightly emphatic about the importance of poetic music within the aesthetics of the source language; they are inherently circumscribed in trying to surmount the ill effects on the artistic whole once the element of rhyme is reasserted in the target language. In thinking too much of one element and too little of others, a rhymed translation typically loses more than it preserves or recreates. Straining to follow one set of methodological criteria is almost bound to cause a warp in the artistic structure at the expense of the organic whole, especially since prosodic and phonological grids show a far greater degree of incompatibility across languages than matters of imagery or theme. Besides, different languages have different linguistic resources for versification, and English versification has long been confronted with specific and general difficulties in using rhyme, whether in composition or in translation. As a comparative inquiry into one (and merely one) aspect of versification, what this essay has tried to do is to offer some empirical evidence on the constitution and application of rhyme in English and Chinese poetry, along with verifiable linguistic and aesthetic observations, which should help to elucidate a translation issue likely to remain in the foreseeable future.