Fenollosa, Pound and the Chinese Character

In 1954 Ezra Pound published his translation of the third of the Chinese Classics under the title "The Classic Anthology Defined by Confucius." (Harvard University Press) It was immediately recognized that these translations, or "translucences", as some call them, stood in a different category from the numerous ones that had preceded. Richard Wilbur is quoted on the dust-jacket as naming Ezra Pound "the first translator of our age." I.A. Richards salutes "Mr. Pound at his best." Achilles Fang, in the Introduction to the volume, notes that "Pound now emerges as a Confucian poet."

There is here an intriguing element of mystery. Pound is not a professional student of the Chinese language, like Legge, Giles, Waley, Karlgren, and others who have tried their hand at translating the Book of Odes into English. We do find the dust-jacket proclaiming that "Pound's translation ... is the culmination of forty years of Chinese study," but it is doubtful whether Pound himself would wish to make that statement. We shall see that he expressly disassociates himself from "scholarship", which he is likely to view as smothering the art of translation, rather than making a contribution to it. The question that we pose may then be stated as follows: Does the superiority of Pound's translation lie in the end-product, the superior style and poetic quality of his English, or does it lie at the source, a deeper penetration into the mind and art of the Chinese poet who furnishes the raw material for the translation? An answer to this question is of interest to all who study Chinese, whether for finding their own pleasure therein, or for giving pleasure to others through translation.

It does not appear that Pound has anywhere offered an exposition of his method in translating Chinese. There is no doubt, however, that he has been continuously stimulated by a short essay entitled "The Chinese Written Character as a Medium for Poetry", composed in the main by the late Ernest Fenollosa. Pound received the manuscript of this essay after the author's death in 1908. He edited it, and in some introductory paragraphs dated 1918 described it as "a study of the fundamentals of all aesthetics." According to "A Preliminary Checklist of the Writings of Ezra Pound" by John Edwards (New Haven, 1953) the essay was first published by Pound in 1920 (No. 26). It was republished by him in 1936 (No. 52) as the first issue of a projected "Ideogramic Series" which seems not to have been continued. In the preliminary Note to Pound's translation of "The Analects" that appeared in the Hudson Review in 1950 (No. 635) the essay is referred to in the following paragraph: "During the past half-century (since Legge's studies) a good deal of light has been shed on the subject by Fenollosa (Written Character as a Medium for Poetry). Frobenius (Erlebte Erdteile) and Karlgren (studies of sacrificial bone inscriptions)." We may not be justified in saying that Pound learned from Fenollosa how to translate Chinese, but we are certainly justified in assuming that he held him in high regard. And this leads us naturally to an examination of the latter's ideas.

Fenollosa's essay is a small mass of confusion. Within the limits of forty-four pages he gallops determinedly in various directions, tilting at the unoffending windmills. This follows, perhaps inevitably, from the fact that one of them is logic. On pages 61, 64, 75, 76, 77, 78 (I refer to the edition in the Square Dollar Series, 1951) he thrusts his random spear at logic that "has abused the language," logic that "deals in abstractions," logic that establishes "classifications." These are but preliminary skirmishes, for the central sin of logic, as it relates to Fenollosa's various theses, is that it has spawned "grammar". It is hard, at this time, to determine where the camp of his enemies lay, since he never names them, but he attributes to "professional grammarians" definitions of the sentence (60-61) and formulations of the copulative (63-65) that seem highly repugnant, at least to him. It is noteworthy that, fifty years later in 1958, the same quixotic tournaments persist. Fenollosa was not clear whether the grammarian was one who described how a language operated or one who prescribed how it should operate. His conceived enemy was the latter, just as it is the nemesis of every would-be writer from sixth grade through freshman college composition who feels himself worsted in contest with a grade-dispensing authority that penalizes every deviation from its own established norm. In so far as Fenollosa was fighting to protect poetry from what he viewed as the stifling palm of a grammarian's commandment, one may sympathize full-heartedly with him. But in our age this issue is a corpse, and the palm is lifeless. No linguist or grammarian elects himself a dictator, nor is he antagonistic to poets. Contrariwise, their functions complement -- the one to create, the other to record.

Having eased himself of his rancor toward prescriptive grammarians, if such there were, Fenollosa moves happily in his essay to the post of descriptive grammarian of Chinese. Whereupon many curious things happen. Fenollosa claims the sentence form to be "forced upon primitive men by nature itself." (p. 62) This form "consists of three necessary words," (p. 63) the agent, the act, and the receiver, in that order. So "the form of the Chinese transitive sentence, and of the English, exactly corresponds to this universal form of action in nature." As an example: "Farmer--pounds--rice." Fenollosa's prose at this point is exceptionally eloquent, and well worth reading. Since nature is not static, but in constant flux, its movement is an unending transfer of power from one point to another. The essence of this movement lies in the transfer, but the transfer is not possible without two terminals. A transitive sentence, --actor, action, receiver--"brings language close to things, and in its strong reliance on verbs it erects all speech into a kind of dramatic poetry."

It should be noted that what has chiefly happened here is the substitution of Nature for the sixth-grade schoolmarm, and that Fenollosa has relapsed into the role of prescriptive grammarian, however much he may detest it. As a symbol of authority, Nature may be more awesome or more sympatisch than the minion of the Board of Education, but in principle we have the same thing. Sentences must contain a subject, verb, and object, in that order, not because Miss Cherivitsky says so, but because Nature has so ordained. Try to beat that! It is highly ironic that the oriental language with which Fenollosa was best acquainted was Japanese, which disobeys these laws of Nature. The three "necessary" elements are expressed in Japanese in the perverse order of subject, object, verb. In two cavalier sentences Fenollosa disengages himself from this embarrassment. Japanese, and some other languages, can behave differently because they have "little tags and word-endings." (p. 63) And with this denouement of a shift back to the descriptive, the curtain comes down on the comedy. Japanese words do have little tags and word-endings. And the skeleton word-order is subject, object verb. But unless Nature deliberately contrived outlandishness for the Japanese and Germans and others whom we have sought in recent decades to confound, there is little left of that universal order.

After settling the natural form of the sentence, Fenollosa discusses parts of speech, and introduces the topic with two brilliant sentences that place him still in this particular regard ahead of our time. His statement is: "Every written Chinese word is properly ... an underlying word, and yet it is not abstract. It is not exclusive of parts of speech, but comprehensive; not something which is neither a noun, verb, or adjective, but something which is all of them at once and at all times." (p. 68) If Fenollosa had had the linguistic training to work out the implications of that statement, or if anyone of us now had the ability, a new descriptive grammar of Chinese could be initiated. Scholars, and grammarians as well, who deal with written Chinese, expecially [sic] poetry, are quite persuaded to follow Fenollosa in the view that parts of speech do not exist. But it is difficult to describe, perhaps even to imagine, such a linguistic condition in terms of another language like English where word-classes are still of some importance. Fenollosa hints that Chinese, unlike English, has "a common word underlying at once the verb 'shine', the adjective 'bright' and the noun 'sun'." (p. 68) Pound, in a very perceptive note, suggests that English could use 'to shine' (verb), 'shining' (adjective), and 'the shine' (noun). It might prove awkward to apply this on a thorough-going scale in English, but at any rate we have a true and fruitful line of thought.

Fenollosa characteristically shifts position as soon as he has taken it. "One of the most interesting facts about the Chinese language is that in it we can see not only the forms of sentences, but literally the parts of speech growing up, budding forth one from another." At that he proceeds to discuss Chinese nouns, verbs, adjectives, pronouns, and even prepositions and conjunctions, with the purpose of showing that all other classes are derived from verbs. Having acted alternately as descriptive and prescriptive grammarian, Fenollosa here assumes the role of historical grammarian. The question whether at any particular period, say the period of T'ang poetry, written Chinese has or has not parts of speech is one to be settled by a descriptive process. If it is found to have parts of speech, then the fact that these arose at some remote period by differentiation from a single class of verbs is quite irrelevant. It would be of historical interest to discover the stage in time at which all elements of the Chinese written language were verbs, but this is not possible, first, because there can be no definition of what constitutes a verb in classical Chinese, and second, because we have no recorded texts in which there is not already a highly differentiated class of grammatical particles.

Such theses of Fenollosa, and others that we shall omit from discussion,--theses tossed out in confusion and self-contradiction, --are designed to preach the simple principle that strong transitive verbs add vigor and vitality to poetry. If this was a new principle in Fenollosa's time, he must be given credit for its forceful presentation, but it is hard to see why it was necessary to shore it up with questionable Chinese props. Fenollosa states that "the great number of these [Chinese] ideographic roots carry in them a verbal idea of action ... a large number of the primitive Chinese characters ... are shorthand pictures of actions or processes." (p. 59) "In translating Chinese, verse especially, we must hold as closely as possible to the concrete force of the original, eschewing adjectives, nouns, and intransitive forms wherever we can, and seeking instead strong and individual verbs." (p. 65-6). If the poetic principle is solid, its application need not be limited to translation from the Chinese, but should be fundamental in all poetic creation or translation alike. But, of course, if Chinese in its own structure relies heavily on strong transitive verbs, --and this has been a central theme, --then it makes more than usual demands on the translator so to represent it. And with no trace of humor Fenollosa presents us with a typical line of Chinese poetry "MOON - RAYS - LIKE - PURE - SNOW." (p. 56) Of the five words three are indisputable nouns, one is an adjective, and the fifth, 'like', is not much more than a variation of the copula. There is no verb in the lot, let alone a "strong individual verb." But the line is typical.

We have now reached a point of despair. The complete Chinese poem of twenty syllables will be reproduced later, when it will be found that eleven of the twenty are defined as nouns, three as adjectives, two as copulative variants, one as intransitive verb, two as of doubtful classification, and only one, 'admire', as a transitive verb with fairly low dynamics. A paraphrase of the whole poem, which appears to be the joint effort of Fenollosa and Pound, reads as follows: (p. 86)

     The moon's snow falls on the plum trees;
     Its boughs are full of bright stars.
     We can admire the bright turning disc;
     The garden high above there, casts its
        pearls to our weeds.

There are perhaps two places where the "strong verbal action" shows through: 'falls on' and 'casts'. Unfortunately, neither one has any counterpart in the Chinese original. 'Admire' is a pretty wistful sort of verb. 'Turning' is intransitive. 'Full of' is properly a passive. (Being poet neither by inclination nor training I ask in childish wonder why the second line as English is not rendered. "Bright stars fill its boughs". And I return to this question later.) On the other hand we have ten nouns and three adjectives. Then the notion that Chinese poetry is overloaded with strong transitive verbs, and that its translation requires use of strong transitive verbs, is, to put it mildly, shot to hell. But if the thesis has failed of support in the arguments so far adduced, it is possible that it may be substantiated through a consideration of the Chinese written character. This is the reason for the title of the Essay, and to this subject we now turn.

We have already quoted Fenollosa's remark that "a large number of the primitive Chinese characters are shorthand pictures of actions or processes." He goes on to say, "but this concrete verb quality, both in nature and in the Chinese signs, becomes far more striking and poetic when we pass from such simple original pictures to compounds." (p. 59) The facts behind these statements are the following. The Chinese unit of writing is a "graph" or "character." In printed texts all characters occupy equal space and appear equally independent, though they vary greatly in complexity from a single stroke to a conglomeration of thirty or more. Many of the simpler characters can be described in such a way as to show pictorial origin still recognizable despite change or distortion. Such forms, of which Chalmers in 1882 compiled a list of 300, are sometimes called "primitives." Another list that comes early to the attention of the student of Chinese consists of the 214 "radicals " or "keys " by which characters are generally arranged in dictionaries. Eighty percent of these are made up of eight strokes or less. The pictorial significance of many of these forms is a matter of debate that has been somewhat clarified in recent decades through the archaeological discoveries of inscribed shells and bones dating from as early as 1400 B.C. Apart from these, chief reliance has been placed on an etymological dictionary compiled around 100 A.D., in which 9353 characters were listed and analyzed with respect to their original shapes and meanings. Contrary to impressions current among westerners, only 364, or 3.9 percent of the characters, could at that time be traced to a pictorial origin.

Simple forms, as may be expected, recur in the more complex characters. Thus in a dictionary of less than 8000 characters the pictograph for 'mouth' is found as "key" in 378, and probably occurs, in one way or another, in 500 more. The pictograph for 'tree' has about the same distribution. Quite contrary to Fenollosa's view, the vast majority of the primitives whose pictorial origin is determinable are pictures of objects, and are translatable as nouns. Since the number of primitives is small, it follows that most Chinese characters appear as composites. The percentage in the etymological dictionary mentioned was 94.8. Nothing then seems simpler to the western mind than that, given the meanings of a few hundred primitives, the meaning of a composite is derivable from the sum of its parts. Once this view is adopted, the reading and translation of Chinese becomes a game that any number can play, and with infinite variety. For the association of a fish, an eye, and a roof, can suggest different things to different people. (That is why the Rohrschach tests are effective.) The odds against the meaning of a character being equal to the sum of the meanings of its parts are about fifty to one. But the two-percent chance seems sufficient to keep the game going and the players happy.

We are fortunate to have a complete record of such a game, played by Fenollosa and Pound as partners, and appended to the Essay on pages 84 and 85. This record is reproduced below. The Chinese text is a poem in four lines of five syllables each. The English words in capitals, except for two mistakes, are dictionary definitions, while below them are the analyses that should in theory produce the definitions.

The results of the analysis shown above cannot but be disappointing to those who have read Fenollosa's essay in hope and expectation. There is hardly a case where the dictionary meaning of a character shows any intelligible dependence on the meanings of the parts. And this despite the best efforts of the players to twist meanings to that end, an enthusiasm that has led in many cases to misreading or misinterpretation, as shown in the following notes:

1 2 3 4 5
6 7 8 9 10
11 12 13 14 15
16 17 18 19 20
  1. Correct. This is the only simple pictograph out of the twenty.
  2. The element at lower right is a pictograph of a bird, but there is no justification for equating this with 'flying'. There is, to be sure, an association of ideas.
  3. The elements 'woman' and 'mouth' are present, but it is not apparent how these suggest the meaning 'like'.
  4. The right hand part has nothing to do with sky except as it may describe its color. The word for which it stands can be rendered into English only by a periphrasis 'the color of nature'. In the spectrum it is blue, green, or even black.
  5. The upper part is rain. The lower, as it stands, is pictographic for 'hand'. The graph for 'broom' contains this element, but so do many other graphs which might equally well be suggested.
  6. On the contemporary level the right-hand part is unanalyzable. There is no element 'crooked', nor 'female breast, but it does include the graph 'mother'. The relation of all this to the meaning 'plum' is, in any case, mysterious.
  7. The element at lower right may easily be confused with 'spoon', but here it represents a man turned in the opposite direction from the normal.
  8. The right-hand element stands for 'by means of' or 'in order to', never 'try'. It is hard to see how this adds up to 'resemble'.
  9. The elements are all correctly identified.
  10. The graph in the text has been wrongly written. When corrected to the graph for 'star' it will consist of 'sun' and 'grow'.
  11. The two strokes here additional to 'mouth' are not 'hook', but a sign of exclamation.
  12. The element described as 'fire' is, in its present form, 'rice'. There is nothing identifiable as 'girl', while 'descending through two' is pictographic for two men back to back. Corrected analysis would hardly lead to 'admire (be in love with)'.
  13. This graph is not analyzable on the contemporary level.
  14. The meaning of this graph is 'mirror'. The elements have been correctly identified, but their total relevance is obscure.
  15. The upper right-hand part is not another 'carriage', but is supposed to be pictographic for 'yoked ox'.
  16. The meaning 'blend' is not represented here. There is an element 'shed' (a building) and an element 'court'.
  17. This is an indicative graph for 'top'.
  18. As the graph now looks it is 'king' and a dot, but it is not clear why this should mean 'jewel'.
  19. The meaning 'weeds' for this graph is the result of some misunderstanding. The upper element is 'grass' and the lower 'square', while the whole means 'fragrant'.
  20. The upper half of this graph stands for a musical instrument consisting of suspended stones. The lower half is made up of 'sun' under 'grain', not 'tree'.

[Fenollosa left the notes unfinished; I am proceeding in ignorance and by conjecture. The primitive pictures were "squared" at a certain time. E.P.]

bright + feathers
woman mouth sun + azure sky rain + broom
cloud roof or cloth over falling drops
sun disc with the moon's horns Bright, vide note on p. 42. Upper right, abbreaviated picture of wings; lower, bird=to fly. Both F. and Morrison note that it is short tailed bird Sky possibly containing tent idea. Author has dodged a "pure" containing sun + broom Sweeping motion of snow; broom-like appearance of snow
tree + crooked
female breast
man + spoon under plants abbreviation, man + try = sun + knife
probably actual representation of blossoms.
Flowers at height of man's head. Two forms of character in F.'s two copies
does what it can toward Bright here going to origin: fire over moving legs of a man
(be in love with)
heart +
girl +
descending through two
gold + to erect
legs (running)
carriage + carriage
tenth of cubit
I suppose it might even be fish-pole or sheltered corner Present form resembles king and gem; but archaic might be balance and melting-pots (?) Bent knuckle or bent object revolving round pivot
to blend + pace, in midst of court king and dot plants
Note: plain man + dot = dog I.e. growing things that must be destroyed Specifically given in Morrison as fragrance from a distance. M. and F. seem to differ as to significance of sun under growing tree (cause of fragrance)

What then is wrong here? For something must be frightfully wrong. Just a complete misunderstanding of what Chinese characters are, how they were created, and how they function as speech symbols. A few illustrations may make this clear:

3. The third character of the poem is divisible into a pictograph 'woman' on the left and a pictograph 'mouth' on the right. At the earliest times that we know the word for 'woman' was nyo, and the word for 'like' happened also to be nyo, or something very close to it. Scribes then began by writing the pictograph 'woman' to represent phonetically the word 'like', leaving context to determine which was meant. This process is familiar to every child as rebus-writing, in which the picture of a stick of wood stands for 'would', or the picture of a bee plus the number four makes up 'before'. In Chinese a refinement was introduced by adding the sign for 'mouth', which in scores of characters gives a signal to be interpreted as follows: This character represents the sound nyo, which is, of course, the word for 'woman'. But in this case I mean the sound nyo that is not 'woman', and you will recognize it as the word 'like'.

7. The lower half of this stands for the sound hwa and hence the word for 'transformation' 'metamorphosis'. The word for 'flower' in archaic times was hhwa (voiced h as initial), and was written with a very different character. Sometime later there came into the language a word for 'flower', hwa, whether by dialect mixture or a sound shift we do not know. But scribes represented it phonetically with their existing character for 'metamorphosis', and later, for the sake of clarity, added above it the symbol for vegetation which was already present in the character for hhwa 'flower'. It will be noted that this process differs somewhat from that illustrated in 3, because there is a possibility that the two words 'flower' and 'metamorphosis' are really only one word in different extensions of meaning.

16. A true example of what has just been suggested for 'flower' is seen here. If the three strokes at top and left of this character be removed, the remainder stands for the sound dieng and the word 'court '. Now there is considerable physical difference between a 'tennis court' and the 'Supreme Court', although we get along in English with one word for both. Similarly in Chinese the word dieng came to mean both an open court in someone's front yard, and the court of the king, which was presumably enclosed. It then occurred to some fussy scribe to add the three strokes that picture a roof and a side wall, so that a palace dieng might be distinctive. He did not realize that such artificialities rarely work in language. Since there was only one word dieng, it could not matter to any but the most meticulous which way it was written. And, amusingly enough, in the poem we are examining, the symbol for dieng, though decked out for the eye with roof and wall, stands clearly for the dieng of the 'garden' variety.

A consideration of the last example will make clear why the approach of many westerners to Chinese is unrealistic. It is not the purpose of this article to teach Chinese, and it may seem to the reader that we have already become too technical. But it is impossible to say anything on the subject without emphasizing and reiterating that characters are symbols for sounds, and through sound are symbols for words. They are not a code for the deaf and dumb, nor a collection of pictures to entrance the eye. It is basic to the philosophy of Fenollosa, Florence Ayscough, Amy Lowell and other translators to believe that the quality of a line of Chinese poetry is chiefly determined by a picturesque choice of characters, and that lithe thought-picture is not only called up by these signs as well as by words, but, far more vividly and concretely." (p. 58) But the fact is that such images as appear through the sort of analysis illustrated above are not present in the mind of the Chinese reader, because he has never thought of them. They were unknown to the compiler of the etymological dictionary of 100 A.D. It is more than likely that they were unknown to the Chinese poet himself, who used the characters as arbitrary symbols for the words of his poem.

In the introduction to his translation of Tu Fu, Professor William Hung directs a gentle criticism at the professed method used by Florence Ayscough and Amy Lowell. "The basic assumption of the method was that the etymological derivations of the Chinese ideographs composing the lines in a poem were of great importance. Suppose there are two characters having the same meaning in current usage. Why should the poet choose the one instead of the other? The ladies believed that the choice was determined according to how well the "descriptive allusions" or the "undercurrent of meaning" would enrich the "perfume" of the poem." (p. 9) Professor Hung is of opinion that "a poet's discrimination between synonyms is very frequently concerned with the difference in sound values." We might go further to argue that sound values are the primary consideration. The assembling of twenty characters, however strong their perfume, does not make a Chinese poem.

In the specimen of poetry here reproduced, a number of formal requirements are shown. In the first place, the words are arranged in four lines of five syllables, each line syntactically complete. In the second place, the second and fourth lines rhyme, and the rhyming syllables are in the level tone. In the third place, the first and second lines show a word to word parallelism or contrast, so that the grammatical structure of both is identical. In the fourth place, with the dichotomy of tones into "level" and "oblique", the tone of each syllable in the first line is matched in the second line by a syllable of the opposite tone class. This tonal opposition appears also between the third and fourth lines. In the fifth place, the tonal pattern of each line is different. And finally, the poem has a unity of thought.

The tonal pattern of the poem is as follows :

/ / _ _ /
_ _ / / _

We do not suggest that these rigid conditions are ideal for the creation of poetry, but merely that such conditions were generally imposed on the T'ang poet. And it is quite beyond reason to imagine that after satisfying these conditions the poet had time to worry over the pictorial stimulation that particular characters might furnish to future western readers. Perhaps the supreme irony is the fact that we have no real knowledge of precisely how a T'ang poet, for example, committed his poem to paper. The modern translator looks at a printed text, and the one thing he may be confident of is that this text does not correspond to the original writing. Handwriting uses abbreviations, sometimes of generally accepted usage, sometimes personal to the writer. But even if it should correspond, there is another conclusive disproof of the notion that the form of the characters was of prime importance. One of the typical situations in which poetry was composed was at an earnest cocktail party, after sufficient spirit had been infused. Then the host might propose a topic, and ask his guests in turn to chant (orally) an extemporaneous poem on the subject. Of the great poet Po Chü-i no biography omits to mention the fact that his poems were appreciated by illiterate old women, and were constantly on the lips of fishmongers in the market-place.

What this all amounts to is simply that Chinese poetry was composed in a language, as all poetry must be. And a poem of the eighth century A.D. can be properly understood only if one knows the language of the eighth century A.D. The assumption of the "etymological" translators--Fenollosa, Pound, Ayscough, Lowell, and others--is that the meaning, connotation, allusion, perfume, concreteness of a given Chinese character has remained immutable from pre-historic times. But this is inconceivable. The important question is, "What was the word represented by a particular character in the eighth century, how did that word sound, and what were its connotations?" To discover this is the effort of philology.

Only in rare combination can philologists double as poets, or poets as philologists. The philologist is concerned with excavating expression from a foreign language, the poet with perfecting expression in his own language. The combination that succeeds is then a combination of both. Despite the trumpeting of Fenollosa to announce a new visual interpretation of Chinese poetry, there is no evidence that he ever followed his own call. The poems in Cathay, translated by Pound "from the notes of the late Ernest Fenollosa, and the decipherings of the Professors Mori and Ariga" (1915), are given a conventional interpretation. The first excursion by Pound alone is to be found in the translation of "The Great Digest" (1928, Edwards No. 36), where three pages of "Terminology" explore the possibilities of pictorial analysis divorced from accepted meaning. The results are exciting and unreal. In "The Unwobbling Pivot" (1947) a certain amount of this analysis continues, but in the "Analects" (1950) it is barely discernible. Those who take the trouble to compare this with Legge's translation (1861) will find that Pound has in large measure taken over philologist Legge and dressed up the English that was sadly unpoetic. In "The Classic Anthology" (1954) the English of Pound has loosed itself completely from any Chinese mooring. And in the Rock-Drill Cantos (1956), particularly no.85., the Chinese has become a decoration with no intelligible meaning.

To elaborate on the foregoing statements would require more space than can be allowed at present. For anyone who grants that Chinese is a language, elaboration is unnecessary. Chinese poetry, like any other, is to be sung, chanted, whispered, recited, muttered, but not (God forbid!) to be deciphered. The association of ideas that results from the dissection of a given character may produce a poetic thought. But this is a new thought, and it may completely overshadow the thought that was in the mind of the writer. In the "Terminology" prefaced to Pound's translation of "The Great Digest" a Chinese character meaning 'sincerity' is analyzed as "the precise definition of the word, pictorially the sun's lance coming to rest on the precise spot verbally." This is sheer imagination in the style of Edward Lear. What is "the sun's lance"? Even if there were an etymological basis for this fantasy, to use it in translation would be comparable to a Chinese insistence on always rendering the English word 'sincerity', as "a state of being without wax." The first line of the Analects reads, "Having studied something, constantly to practice it, is this not a joy?" Pound has "Study with the seasons winging past, is not this pleasant?" "Seasons" is impossible. The thought of "winging past" comes by isolation of a portion of the character meaning "practice". Six sentences later the same character occurs, and Pound translates it "practice". Either the thought of "winging past" failed to materialize, or it was found impossible to work it into the context. But this represents a totally irresponsible attitude toward the Chinese language. When it suits the translator's whim, he may construct any number of bright images from the bits that he thinks he has discovered in the character. When he is tired, he falls back on the simple word that the character symbolizes.

The character in this case is pronounced shyi, southern China ziq, time of Confucius zip. If there is a language, then zip has always had a specific meaning, not necessarily the same, since language grows. But this meaning cannot be found by theorizing, any more than one might determine that "minimum" means "milk" because it begins and ends in m. All Chinese literature we have, including the Analects, indicates that zip means, and has always meant, "practice". In the Analects zip occurs three times, twice in association with 'learn'. The repeated idea is that learning is fruitless unless one puts it into practice. Pound sacrifices this rather important precept for the sake of a pastoral where the seasons go winging by. Undoubtedly this is fine poetry. Undoubtedly it is bad translation. Pound has the practice, but not the learning. He is to be saluted as a poet, but not as a translator.