The Color of Nature and Other Hobbyhorses in Ezra Pound's Cathay

by Timothy Billings



Shortly after the release of my critical edition of Ezra Pound’s Cathay in late 2019, there appeared the oddest review of it in Make It New which I thought not worth the bother of responding to at the time on the grounds that the edition speaks for itself, the truth will out, and all that. But in recent months I have been nagged by the thought that since the points in that review all relate to questions of East Asian language and culture which many of the otherwise learned readers of this journal might not have the background to adjudicate on their own, I should probably step forward to correct the record. It also seemed like an occasion to explore further a few ideas that could not easily be accommodated in the footnotes of an edition, as well as to take a moment to reflect on what we are really doing when we review each other’s work.

I suspect that most readers have seen book reviews that seem to be more about the reviewer than the book under review, though they may be hard to recognize as such unless one already knows something about both. Although the reviewer in this case, Andrew Houwen, seems almost begrudgingly to concede that the new edition is “indispensable,” he says almost nothing about why that should be the case—indeed, he says very little about the contents of the book at all—in favor of making a number of complaints that are not only misrepresentations of the edition, but also mostly mistaken in and of themselves, all related to Japan in some way. The reviewer does at least acknowledge the importance of the edition’s shift of critical perspective away from the typical comparison of Pound’s versions with Chinese originals he never saw, onto Ernest Fenollosa’s notebooks which Pound plundered so gleefully and gloriously. But the aim of the review was apparently not to explain the contributions of the edition along with an assessment of its strengths and weaknesses, or even to give readers an idea of what it covers and how, but rather to stake a territorial claim of expertise in “Ezra Pound’s Japan” in anticipation of the reviewer’s own book on that topic, which has just been released.

My edition of Cathay, like everything else in this vale of tears, is not flawless, I will not hesitate to admit, despite my slavish labors to make it so. Since publication, I have discovered a few typos in its nearly one thousand footnotes which somehow neither I nor the copyeditor caught in production. Such unforced errors are lamentable, but they are not systemic; and therein lies an important distinction that is sometimes overlooked by reviewers.

In the second paragraph of page 18 of my introduction, as the reviewer points out, I refer to Herbert Giles as “Giles Herbert.” That’s a mortifying error, to be sure, a real head-smacker, a unique eruption of the mild dyslexia I have more or less completely overcome with extreme meticulousness, but it is an occasional error, not a systemic one. What the reviewer fails to mention is that I refer to Herbert Giles 86 times over the course of the book—and to his son, Lionel Giles, once—but it is only in that one paragraph that I absent-mindedly invert his names.

The reviewer, however, presents this mistake as a characteristic failure of the edition to insist on “right names,” as he puts it wittily enough, the chief evidence for which is that I refer to Fenollosa’s teacher, the eminent sinologist Prof. Mori Kainan, by his family name “Mori,” instead of “Kainan” which is, the reviewer writes, “his gō 号 a traditional literary or artistic pseudonym often used in Japan until around the Second World War. A cursory examination of Japanese-language literary scholarship shows that, after first mention, he is always referred to as ‘Kainan.’” All well and good, but I did not write my commentaries in Japanese, and the convention in English-language literary scholarship, of course, is to refer to people, after first mention, by their family name. Honestly, I don’t know what else to say. The reviewer then concedes: “It is perhaps defensible to maintain the use of ‘Mori’ because Fenollosa always refers to him as such, but at the very least a note acknowledging this issue would be helpful in explaining such basic information regarding the Japanese context of Cathay’s development.” All of this strikes me as less of a meaningful critique of the edition, and more of an attempt to suggest that I don’t really know what I’m talking about. If the reviewer had made a cursory examination of Chinese-language scholarship on Li Bai (Li Bo) 李白 he would also know that the Tang poet is typically referred to as Taibai 太白 which is his zi 字 or “style name.” Indeed, that’s what Prof. Mori himself calls him in his own Chinese preface in a Japanese work called Rishi kōgi 李詩講義 (Lectures on Li’s Poetry), posthumously published in 1914, which the reviewer complains I do not cite in my edition even though he seems not have examined it very carefully himself. I’ll return to that work in a moment, but the point here is that I don’t refer to Li Bai as “Taibai” for the same reason I don’t refer to Mori as “Kainan”: it was not ignorance, or negligence, or a lack of respect on my part, but simply that I was writing in English.

As for Rishi kōgi, that collection of lectures just mentioned, the reviewer scolds me for not citing it because of “the clearly important role that published lectures by Kainan on Li Bai’s poetry ought to play in an understanding of Cathay.” It’s true that I don’t cite that volume—in truth, because I didn’t know about it at the time, and I was delighted to learn of its existence—but I have now looked at it only to discover that it is not at all what the reviewer claims it to be. These so-called “published lectures by Kainan” are in fact notes taken in shorthand by another Kanshi 漢詩 poet (that is, a Japanese poet of classical Chinese verse) known as Aranami Engai 荒浪煙厓 (literally, “Rough Waves Smoky Cliffs,” a pseudonym, of course) (1870–1954) during his own lessons with Mori, which were then posthumously reconstructed for publication by yet another famous Kanshi poet named Morikawa Chikukei 森川 竹蹊 (1869–1917), who explains all of this in his foreword with apologies for the many reconstructions, omissions, and remaining uncertainties in the text. From my survey of these reconstructed lectures, where the selections do overlap with Fenollosa’s selections (all from the same Qing-dynasty Chinese anthology, which I had already identified from the notebooks as his source), they unsurprisingly present the same readings, which is to say there is nothing new there. Each Chinese poem is paraphrased in Japanese, sometimes with dilations to make implications more explicit, along with a few comments on backgrounds and form, and explanations of allusions—all of which we already find in Fenollosa’s notes in English. Indeed, Fenollosa’s notes would be more helpful in explicating the Rishi kōgi than the other way around. For one, Fenollosa’s notebooks include word-for-word glosses which the posthumously edited student lecture notes do not. If I ever revise the edition, I don’t see anything there worth the trouble of citing, though of course I know that fine-toothed combs always produce some crunchy nits. Apparently, Mori did intend to write up and publish his lectures at some point before his premature death at the age of forty-eight (which I assume because he wrote that Chinese preface, unless it is reprinted from some other place I don’t know about), but that’s not what the Rishi kōgi is. A good project for someone highly proficient in Japanese might be to locate those handwritten notes by Rough Waves Smoky Cliffs (if they still exist), decipher whether anything of consequence had been edited out or altered by Morikawa, and then explore whether any of that has any bearing at all on Fenollosa’s notes behind Pound’s versions. That’s beyond what I could do (and undoubtedly beyond the reviewer’s abilities, too, unless he has a hidden proficiency in fin-de-siècle Japanese shorthand he hasn’t boasted about), but that’s why we work collectively to produce knowledge. I’ve done my part with my proficiency in classical Chinese; there remains, as always, more work to be done by others with other proficiencies.

In the next sentence the reviewer digresses for a moment to discuss a different work: “Bush, Billings, and the bibliography draw attention to Kainan’s Tōshisen hyōshaku, though the book does not directly cite it, or any other Japanese-language source for that matter. The Tōshisen hyōshaku would not be a particularly helpful source for a critical edition of Cathay: out of more than four hundred poems, only three appear in Cathay.” Just to get the facts straight: Cathay contains six poems, not three, from the Tōshisen, and I cite them all, but only to inform readers which poems to look for if they should wish to seek out Mori’s comments on them in Japanese, which of course Pound never saw. (It also includes the poem on the Yellow Crane Tower by Li Bai’s contemporary Cui Hao 催顥, which I translate in the notes as an essential pre-text to the original for “The City of Choan.”) I agree that this anthology is not “a particularly helpful source,” which is one reason I do not bother with it. But what readers must understand is that the work in question, as the name suggests, Tōshisen hyōshaku 唐詩選評釋 (The Annotated Selection of Tang Poetry), is Mori’s edition of the Tangshi xuan 唐詩選 (Selection of Tang Poetry), a famous late Ming (16th-century) anthology which became the predominant anthology of Tang poetry in Japan. Mori’s edition reproduces the Chinese text and provides Japanese translations and a few comments like those in Fenollosa’s notes. Chris Bush and I both mention it separately in our introductions because it was an important work in the Japanese reception of Chinese poetry which helped establish Mori’s reputation as an expert on Tang poetry. The reviewer probably believes that I never cited the edition because he did not recognize that I cited it using the conventional sinological abbreviation for the Chinese name of the original anthology for which Mori’s is the Japanese edition (i.e, TSX for “Tang-shi-xuan”), as is clearly spelled out in the abbreviations given in the bibliography. I also cite textual variants in TSX, such as, for example, in the opening to note 14, on page 133, about a word in the poem that became “The River Song,” where I observe: “Miao and TSX read xiao 嘯 (whistle; roar; incant; wail, etc.), but TSSC gives xiao 笑 (laugh) in a slightly variant form.” In other words, the Ming edition that Mori edited (TSX) and the Qing edition Mori lectured from (TSSC) have two different characters (homonyms of each other), which I then discuss for the rest of the note. Admittedly, if you don’t know what the Miao text is or the TSX or the TSSC, then you won’t be able to make sense of that first sentence (and different readers will use different parts of the edition in different ways), but that’s what the abbreviations in the bibliography are for. For the record, the six poems of Pound’s originally from this anthology are: “The River Song” (part one only), “Taking Leave of a Friend,” “Old Idea of Choan,” “Leave-Taking at Shoku,” “Separation on the River Kiang,” and “The City of Choan.”

For what it’s worth, it’s also not true that I never cite Japanese-language materials, though they are generally beyond the scope of the edition which is already a dizzyingly complex layering of textual material. In note 8, on page 110, for example, I cite a series of Japanese articles on the composition of Chinese characters by Ōhara Kyōzō 大原 京蔵 published in the spring of 1898 which seems to have influenced an annotation of Fenollosa’s I spotted in the notebooks. As I suggest in that note, these articles of Ōhara’s “warrant further study as an influence on Fenollosa’s ideas about Chinese characters” (Cathay 110). Someone with even tolerable proficiency in Japanese could make an interesting project of examining them together with Fenollosa’s reading notes along with The Chinese Written Character as a Medium for Poetry. That’s an article I would love to read. Such an article would also ideally examine Teikichi Kojō’s 古城 貞吉 (1866–1949) Shina Bungaku Shi 支那文學史 [A History of Chinese Writing] (Tokyo: Keizai zatsushi, 1897), which I also cite by way of suggestion (30 n. 21). 

For, Oh, for, Oh, the Hobbyhorse is Forgot

After having slighted the importance of Mori’s edition of the TSX, the reviewer then circles back around to those reconstructed lectures: 

Kainan’s Rishi kōgi, by contrast, which are freely available in digital form via the National Diet Library’s online catalogue, examine in much greater detail what Pound would translate as “The River-Merchant’s Wife: A Letter”, “The Jewel Stairs’ Grievance”, “The River Song”, and “Poem by the Bridge at Ten-Shin”. An examination of Kainan’s extensive commentary on these poems would provide valuable points of comparison with Fenollosa’s notebooks. (My emphases.)

First, to get the facts straight, the reconstructed lectures in Rishi kōgi also cover the originals for “South Folk in Cold Country” (39), “Ancient Wisdom Rather Cosmic” (54), and “The River Song” (part two) (459). Second, to suggest that there is “extensive commentary” that would provide “valuable points of comparison with Fenollosa’s notebooks,” with “much greater detail” is simply a falsehood. But the reviewer then proffers an example (which does not actually exist), suggesting that my treatment of “The River Merchant’s Wife” is flawed because I do not cite Rishi kōgi which “analyses this poem at length, including the depiction of the ‘bamboo horse.’”

Some readers will recognize that the reference here is to the “bamboo horse” crux which has been well known since Wai-lim Yip drew attention to it in his monograph on Cathay half a century ago. The two characters zhu 竹and ma 馬 separately mean “bamboo” and “horse,” but they combine to form the compound zhuma 竹馬 meaning “hobbyhorse.” In the word-for-word glosses, Fenollosa duly gives the separate characters as “bamboo” and “horse,” but in the paraphrase he gives the compound as “bamboo stilts” because the same two characters in Japanese (not Chinese) form a compound that means “stilts made of bamboo,” a sense it has apparently had for centuries and continues to have to this day. A plausible explanation for the disparity is that the Japanese word takeuma (bamboo + horse) was already in use to mean “bamboo stilts” before the Chinese characters were borrowed to represent it. One graphic signifier, two “sound images,” as Saussure liked to call them, for two different signifieds in two different languages. (That’s what happens when you start sampling and remixing graphs across languages.) Thus, the hobbyhorse crux is one of the extremely rare instances where some kind of Japanese linguistic bias enters Fenollosa’s notebooks, but which has all too often been improperly invoked as if it were a systemic error characteristic of Mori’s lack of proficiency in classical Chinese. 

Here is the so-called lengthy analysis of the bamboo horse crux in Rishi kōgi:

今, 我夫となった人は, まだ子供の事であって, 其時分は, 竹馬に乗って我家 へ来た, 幼友達であるから, 我床を遶って, 頻に青梅をば手に摘んで, 共に戯れて遊んだことであった。 (315)

Now, the person who became my husband was at that time still a child who came to my house riding bamboo stilts [竹馬], therefore a childhood friend, and we would frolic and play together circling my bed [床] and picking blue [青] plums.

That’s it. That is the sum total of the discussion of the hobbyhorse crux in Rishi kōgi. (Note that there are three cruces in this passage, which I shall discuss in turn.) Most readers of this journal probably recognize this passage as a paraphrase of the third and fourth lines of the poem that became “The River Merchant’s Wife: A Letter,” as Pound renders them: “You came by on bamboo stilts, playing horse / You walked about my seat, playing with blue plums” (119). I cannot know whether the reviewer read the text carelessly and mistakenly imagined that it must contain a lengthy analysis, or whether he read the text carefully and yet disingenuously suggested that it contains what he knew it does not. But I am sure of one thing: It is not a lengthy analysis of anything.[1] 

Moreover, this supposedly lengthy analysis provides no resolution whatsoever for the crux. When Mori in his lecture said “takeuma” (assuming that this is what Rough Waves Smoky Cliffs heard and jotted down, and what Morikawa in turn reproduced as 竹馬), was Mori envisioning a pair of stilts or a hobbyhorse? How could we know? What we can observe is that there was apparently no further attempt to disambiguate the term, as though there were no awareness of a potential misunderstanding between the Chinese and the Japanese terms. The mere use of the word whose meaning is in question reveals nothing, although it must be stressed that almost any Japanese person would say that the word in that sentence means “bamboo stilts” quite simply because that’s what takeuma typically means in Japanese. Ask a Japanese speaker. Indeed, when I read the passage above with the help of Prof. Kawai Shoichiro 河合 祥一郎, a distinguished scholar at the University of Tokyo and a celebrated translator of Shakespeare into Japanese, Prof. Kawai was surprised to learn that what he has always known as takeuma 竹馬 in Japanese means something other than “bamboo stilts” in Chinese. Prof. Mori, as a specialist in Chinese poetry, may have known otherwise, but if he did, he also ought to have known that other Japanese readers, listeners, or interpreters would not.

I believe that one of the most significant insights of my work on Fenollosa’s manuscripts is the observation that Mori and his interpreter, Ariga Nagao, read through each poem in its entirety two times with two distinct aims, first providing word-for-word glosses of individual characters with little or no regard for their context, then providing accurate paraphrases for each line properly interpreting everything according to context, along with other brief commentaries on form and appreciation. Classical Chinese is highly elliptical and polysemous, and context is crucial to determining sense. Pound, however, often favored the acontextual glosses over the contextualized paraphrases, which led to many of the various “mistranslations” sinologists continue to complain about. (Pound used the same process, mutatis mutandis, in his handling of the glossary to Sweet’s Anglo-Saxon Reader for his translation of “The Seafarer,” a poem that has nothing to do with Japan, of course, and is one of the many elements of the edition not mentioned by the reviewer.) Let me stress this point once again because of its importance: The word-for-word glosses do not represent Mori’s final interpretations of the poems, but are merely a sort of lexical map or glossary for each line, whereas the paraphrases are his actual “decipherings”; but Pound, like every other scholar after him who looked at the notebooks, seems not to have understood this structure (or to care, if he did), and so drew his poetic material from the storehouse of words in the glosses which were, in many cases, “incorrect” according to context but then later fully corrected when Mori paraphrased the whole line. In the case of the hobbyhorse, Pound drew his readings from both the glosses (“bamboo” + “horse”) and the paraphrase (“bamboo stilts”) even though they differ, combining them into what I call “a uniquely Anglo-Sino-Japanese line”: “You came by on bamboo stilts, playing horse” (Cathay 127 n.9). Some of the curious beauty and value of Cathay derives from precisely such cultural and linguistic fusions, which anyone can appreciate now that there are legible transcriptions of Fenollosa’s notoriously illegible notebooks at hand. 

The Beds i’th’East are Hard

The second of the cruces in the passage from Rishi kōgi above is also illustrative here. The graph 床 has puzzled many a reader and translator over the years since it typically means “bed” (in both Chinese and Japanese), but the opening scene of the poem obviously depicts two children playing outdoors. An obsolete sense is clearly operative here, but what? Arthur Waley ventured “trellis,” but admitted that it was “only a guess” (cited in Cathay 127). Mori & Ariga, as recorded by Fenollosa, however, sagely render it as “seat” in both the gloss and also the paraphrase. The reading refers to either or both of two glossing traditions explaining it as: 1) a huchuang 胡床 (literally, “barbarian bed”), a kind of chair or folding stool which was still a new import to China during the Tang dynasty; or 2) a jingchuang 井床 (literally, “well bed”), an old word for “well curb” or “well side,” a protective wall or railing around a well, including one that could serve as a bench or seat (Huo et al., 65; Cathay 127). The latter is the more likely reading. I find the same compound in a parallel couplet from a ninth-century poem by Yan Qian 彥謙: “Fig vines hang from my reading studio; / The parasol tree droops over the well-bed” (薜荔垂書幌, 梧桐墜井床) (Cathay 127). If this is the correct reading, the sense of the original couplet would be: “You came up riding on a bamboo horse, / Then circled the well curb tossing green plums.” Since such a well would ideally be located in the courtyard of a house where a tree was customarily planted—such as a parasol tree or perhaps a plum tree—the lines might imply movement from the front gate in line two to the inner courtyard in line four while also hinting at the erotic telos of the marriage bed in their furniture future (Cathay 127). Fenollosa’s notebooks show that Mori was clearly aware of the false friend and also of one or both of the glossing traditions. Pound might easily have given us a boy on stilts horsing around the girl’s bed, but Mori was learned enough never to give him a chance. In the reconstructed lectures in Rishi kōgi, however, all we find is the word toko 床 (bed) without any disambiguation, suggesting that the word means nothing but what it typically means in Japanese: bed. In short, Ariga and Fenollosa in the notebooks thus tell us more about what Mori was thinking here than Rough Waves Smoky Cliffs and Morikawa do in Rishi kōgi.

The passage also contains one of those tasty little nits I expected to find, the interpolated phrase “therefore a childhood friend” (幼友達であるから). I understand this comment as one of Mori’s typical little asides, explaining that to ride bamboo stilts is a metaphor for “a childhood friend,” as in the Japanese phrase chikuba notomo 竹馬の友 (literally, a “bamboo stilts friend”). Prof. Kawai tells me that the compound is only ever pronounced with its Sino-Japanese reading of chikuba (instead of takeuma) in this set phrase, which may hint at a closer etiological connection with the Chinese term, though in Japanese it traditionally has the gender restriction of referring to boys. The analogous term in Chinese, qingmei zhuma 青梅竹馬 (literally, “green plums, bamboo hobbyhorse”), traditionally describes a heterosexual romance that begins in a childhood friendship, something like falling in love with the boy or girl next door. The proverbial expression derives from this very poem by Li Bai (Jiang 122).

Disguised Compliments

The reviewer’s unfounded complaints about my handling of the hobbyhorse crux are, however, merely a particular instance of the more general—and utterly dumbfounding—complaint that my edition makes Mori “the victim of unjustified criticism regarding his interpretations of the Chinese poems.” This should come as a surprise to my readers, as it did to me, since I believe it is one of the more salient arguments I make in the edition, and repeatedly demonstrate in the critical apparatus, that Mori’s deservedly high reputation as an expert on Chinese poetry has been continually underestimated in a century’s worth of commentary scapegoating him for supposed errors of translation in Cathay which actually result from Pound either generously exercising his poetic license or misreading Fenollosa’s cribs. 

Let’s review some background here. As is well known, since Mori’s command of English and Fenollosa’s command of Japanese were simply not good enough for them to have a lesson together without an interpreter, they enjoined the help of a former student of Fenollosa’s with an interest in Chinese poetry and an exceptionally good command of English, Ariga Nagao, who became a prominent professor of international law. The notebooks suggest that Prof. Ariga provided for them what we now call consecutive interpretation, translating into English what Mori said in Japanese, sentence by sentence, phrase by phrase, and even word by word when glossing individual Chinese characters. I deduced from certain phrases and errors in Fenollosa’s notes that Ariga was often interpreting Mori’s speech very closely, even literally at times (for example, saying “This is my idea” to refer to Mori’s idea, not to his own); and Fenollosa was apparently taking dictation from Ariga just as closely (for example, writing “This is my idea” before clarifying in parentheses “Mr. Mori”) (Cathay 19). That is to say, Fenollosa wrote down what Ariga said instead of recasting the ideas into his own words, probably because it was swifter to do it that way. This insight is important because it means that when Pound borrows a word or phrase directly from Fenollosa’s notebooks—which he often does, as the critical apparatus of the new edition shows—what he is doing is borrowing poetic diction from a Japanese law professor. That’s a savory ingredient in the delicious hybridity of the Cathay poems, but it poses conceptual and practical problems for the commentator who wishes to respect that complexity while also discussing the hundreds of little translingual transformations that take place from Li to Mori to Ariga to Fenollosa to Pound. Every time we invoke the agency behind, say, a word in one of Fenollosa’s notebooks, we’re not really talking about Fenollosa’s ideas or even his language anymore. In the past, it has been easy for commentators to dismiss the cribs as faulty on the grounds that Fenollosa was a mere beginning student of Chinese poetry, but I have shown that Fenollosa’s notes represent a relatively reliable transmission of Mori’s considerable sinological learning. It would therefore be regressive and misleading to continue to refer to individual notations in the cribs simply as Fenollosa’s. For this reason, I adopted a unique naming convention for the edition, as I explain thus: 

When discussing the cribs in this edition, I use the compound agent “Mori & Ariga” where the idea is indistinguishable from the diction. Of course, Fenollosa also (literally) had a hand in setting down the language, but his part goes without saying since the formula would quickly get unwieldy (“Mori-Ariga-Fenollosa write . . .”). Like any student’s lecture notes, most of what we find there is not, strictly speaking, the student’s writing. (Cathay 29 n.16)

In other words, when discussing an idea in the cribs without regard for the particular language in which it is expressed, I refer to “Mori” alone; and when discussing Ariga’s language without regard for the idea behind it, I use “Ariga” alone. But the latter case is extremely rare for what should be obvious reasons, and I believe I do so only when speculating about whether certain of Fenollosa’s errors seem to have resulted from Ariga’s accented English (such as “ship’s intestines” for “sheep’s intestines”), or when discussing one of the notebooks prepared by Ariga alone where everything is Ariga’s (Cathay 19). Otherwise, when I talk about the contents of the notebooks, I refer to “Mori & Ariga” because it is almost always impossible to make a distinction between whatever Mori said in Japanese and how Ariga translated it into English as recorded by Fenollosa.

This background is relevant when I raise the question of whether the Japanese cultural and linguistic contexts for Fenollosa’s study of these Chinese poems may have exerted some unique influence on the cribs which fell into Pound’s hands, in particular whether Mori & Ariga’s lessons on these millennium-old poems may have been affected by the Japanese language itself, which uses many Chinese graphs to refer to words in Japanese, an example of which we have just seen in the “bamboo horse.” That it is a natural question to ask of any translingual, transcultural work, it seems to me. And it proves to have been enlightening in the case of Mori & Ariga’s two-stage reading method I mentioned earlier which structures the glosses and paraphrases of Fenollosa’s cribs as two different modes of approach to the Chinese text with two different goals. I believe that constitutes a groundbreaking insight into Pound’s materials and how he used them—the sort of thing, you know, one might mention in a book review—but Mori is certainly not to be faulted for following a culturally specific pedagogy that Pound did not understand. Nor do I ever fault him for it. That’s how discourse works. If we have learned anything at all from the past century of literary and cultural study it is that reading practices are contingent upon discourses that are particular to their cultures and subcultures, and that language and cultural conventions play at least some role in shaping hermeneutic outcomes. We need not accept the strongest and most absolute forms of those ideas to recognize that they are true to some extent, and that their effects are sometimes material and significant.

Therefore, by posing the question of whether there might have been some Japanese linguistic bias in the notebooks stemming from the unique relationship Japanese has with Chinese, one that might be more visible from the triangulated perspective of English, I am not attributing ignorance or stupidity to Prof. Mori, but only asking what I routinely ask of all translations, including my own. Faux amis existed long before Facebook. They are fascinating articulations in the conjunction of languages that can be very interesting to bend back and forth in criticism. If humility is something one may boast about, let me say that I confessed my own naïveté in my introduction for all to see, having been humbled by Mori’s impressive knowledge as it is profusely represented in Fenollosa’s scrawlings, which forced me to revise my initial assumptions by the time I began to answer that question in writing: “I naïvely expected to find many errors in the cribs stemming from Japanese linguistic bias, on the foolish assumption that Japanese teachers must necessarily make plenty of ‘Japanese’ mistakes. But I found only two or three significant instances among the hundreds of glosses for these poems, and a thin smattering of insignificant ones” (Cathay 25). It’s true that when I began work on the notebooks I did underestimate Prof. Mori—as quite a lot of scholarship on the topic had encouraged me to do—but it did not take long for me to see what an exceptional reader of classical Chinese poetry he was, even more impressive in various ways than Giles (a.k.a. “Herbert”). I have learned a great deal from Prof. Mori. He is my teacher, too. The edition shows this. But I also explore what I believe to be a judicious critique of those few instances where the mediating language shapes what it mediates.

Here is the note in my edition which the reviewer found so objectionable:

In only a handful of places do Mori & / or Ariga seem to suffer from a Japanese linguistic prejudice in their readings which find their way into Pound’s poems. The most notable is qing 青, which means “green, blue, or black,” which is almost always “blue” in Japanese, hence Pound’s “blue” mountains and “blue” willow-tips. (See the note on this tricky chromonym to line 1 of “The Beautiful Toilet.”) The other is yu 玉, which means “jade” or some other precious stone, but which almost always means “jewel” in Japanese, hence Pound’s “jewel” stairs, “jeweled” flutes, etc. (See line 38 of “The Exile’s Letter,” which contains both of these biases, but see also line 41.) The compound zhuma 竹馬 (bamboo hobbyhorse) was long ago adopted into Japanese as an idiom for “bamboo stilts,” and as such entered “The River Merchant’s Wife.” (I can’t help but wonder whether this was Ariga’s “English” error alone as he interpreted for Mori, or if it just never occurred to either of them.) Unfortunately, it has become the poster-child howler for Mori’s incompetence ever since Yip’s study, but it is an almost unique case. (Cathay 31 n. 33)

That still doesn’t sound like “unjustified criticism” to me, but rather more like the open-eyed defense it was meant to be. I must point out that this is the one and only place in the edition where I write “Mori & / or Ariga” instead of “Mori & Ariga” because I was introducing a varied list in a longer note and wanted to suggest that in some of these cases the mistake might be Ariga’s alone since (as I have made abundantly clear) all the English recorded by Fenollosa was spoken by Ariga, and none of it by Mori who spoke only Japanese. Perhaps one could try to distinguish instances where a correct reading by Mori was incorrectly translated by Ariga on the one hand, from instances where an incorrect reading by Mori was correctly translated by Ariga on the other, but there’s usually too little evidence to make any such claims, and what I am doing there anyway is listing in a footnote some of the instances I found that readers might wish to explore more fully on their own. And yet I also couldn’t resist suggesting in that parenthetical aside that the hobbyhorse crux in particular, which is so well known, really looked to me like Ariga’s error of English translation alone, though there is not enough information to do more than guess, suspect, wonder.

On that point, the reviewer condescends to agree: “Billings is right, though, in ‘wondering’ whether the misreading of 竹馬  (takeuma, or zhuma in modern Mandarin) as ‘bamboo stilts’ in ‘The River Merchant’s Wife: A Letter’ is ‘Ariga’s “English” error alone.’” My cautiousness in “wondering” is thus presented as a failure of certainty, which the reviewer rushes in to save me from, allowing his greater decisiveness to claim my idea as his own. But, honestly, who can know for sure what Mori & / or Ariga were thinking at that moment in the lecture? If I am cautious enough to express uncertainty about such a question, while also suggesting that this one error does not look to me like Mori’s, how is my tentativeness an assault on Mori’s memory? (That’s what I learned in grad school to call responsible scholarship.) Let me stress again, if the only evidence we have from Fenollosa (or Rough Waves Smoky Cliffs, for that matter) suggests that Mori simply said takeuma without acknowledging in any way that it means something different in Chinese, how could we possibly know whether he was thinking something other than what that word typically means in Japanese, i.e. bamboo stilts? (Ariga, as the notebooks record, was not thinking anything else.) I would not be surprised to learn that a Japanese expert on Chinese poetry like Mori might know the Chinese sense of zhuma 竹馬 (hobbyhorse), but I also know from personal experience that this is precisely the sort of mistake (i.e., the simplest kind) that is easiest for any comparatist to make. I would be no more surprised to learn that it never occurred to Prof. Mori to think the Chinese compound 竹馬 meant anything other than what it meant to Prof. Ariga or to Prof. Kawaii when they intuitively read it in the Japanese sense, which is not to undermine the extent of Mori’s sinological knowledge, but to acknowledge the fallibility even of a master while also attributing the mistake (if that’s what it was) to a linguistic problematic as opposed to an individual shortcoming. Chinese commentators typically do not annotate zhuma 竹馬 probably because its meaning has been more or less self-evident to every child in China since the Han dynasty, at least until the Xbox arrived.[2] Perhaps Mori &/or Ariga never bothered to explain it for the same reason, but with a different “self-evident” sense in mind. I do not know. I present the textual evidence for readers to decide on their own, and suggest paths of thinking without dictating the itinerary, as I believe it is the responsibility of an editor of such a volume to do. 

Gruesome Nature

At last, I come to the title of my essay and the third of the cruces from Rishi Kōgi. What the reviewer singles out for his most severe objections are my comments (in the footnote already quoted) about the notorious color word qing 青, or “colour of nature” as it is so poetically defined by Robert Henry Mathews in the Chinese-English dictionary Pound was so fond of using (164). What exactly is the color of nature? It’s the color of willows and unripe plums; of moss, mountains, and mists; of oxen, eyebrows, grass, and the heavens above. Or, as Mathews goes on to say, more prosaically: “green, blue, black. A drab neutral tint” (164). According to Donald Sturgeon’s word-cloud of the approximately 49,000 poems in the massive Qing-dynasty compilation known as Quan Tangshi (The Complete Tang Poetry), qing 青 is among the most frequently used characters in Tang poetry.[3] It is what linguists and social scientists have come to call “grue,” a portmanteau word defined by Carole Biggam in The Semantics of Colour as “a macro-category consisting of green+blue and, sometimes, including grey” (206). The word was apparently coined in the 1950s by the American philosopher Nelson Goodman to describe an unknown color in an induction problem, that might be either green or blue upon inspection, but the term was then adopted by color researchers to name a basic color term that appears in many languages (Cohnitz and Rossberg, online; Kay and Maffi, “4.3. Map 134A: Green and Blue,” online). The Chinese grue is unusual but not unique for including black. Pound would later use Mathews’ turn of phrase in The Cantos when ideogrammatizing the homonym qing 情 (feelings) in the compound enqing 恩情 (loving kindness) to produce “his feelings have the colour of nature,” followed by the harangue: “And as Ford said: get a dictionary and learn the meaning of words” (XCVIII, 709). And so we do: qing 情 (feelings) is composed of xin  忄= 心 (heart) and qing 青 (grue). You might already know that from Wellen (70).

A number of studies have shown that macro-color categories in many languages are multi-focal: “In a grue category, for example, there may be one focus in the green area and another in the blue area, and one of the foci may be of lesser salience than the other” (Biggam 61). This is exactly what we see in the Chinese and Japanese grues, with an inverted salience of focalization on green and blue despite their representation by the common graph 青. In other words, the Chinese qing 青 has a greater focal salience in green (i.e., primarily meaning green and secondarily meaning blue), whereas the Japanese ao 青 has a greater focal salience in blue (i.e., primarily meaning blue and secondarily meaning green).

It may be challenging to wrap one’s mind around the idea of a multi-focal macro-color if you don’t speak a language that has one, but for speakers of those languages the difficulty really only arises in translation. Brent Berlin and Paul Kay in their seminal early study of basic color terms popularized the example of Russian’s two words for blue, sinij (dark blue) and goluboj (light blue), which are overlapping but distinct basic color terms, where sinij is sometimes understood to contain goluboj and at other times to be distinguished from it as a separate color (36). By this loose analogy, we might think of the English “blue” as a sort of macro-color relative to sinij and goluboj, which does not prevent English speakers from distinguishing between the two hues of light and dark blue. (We just don’t have “basic color terms” for them.) Biggam explains it thus:

It should not be imagined that a community using macro-colour categories cannot visually distinguish between the two hues; they just regard them as two varieties of the same colour, since the one merges into the other and the community has no compelling reason to regard them as fundamentally different. They are very likely, of course, to have non-basic terms appropriate to each separate hue so that those who need to, such as painters, can adequately communicate a specific colour area. For example, Zulu speakers can specify with phrases: “grue like the sky” and “grue like the grass.” (61-62)

It follows from the same logic that the precise color signified by a grue color word in a text will be almost entirely contingent upon context (or convention), which is to say, by its association with a thing of an already known color. This creates an intriguing paradox insofar as the thing modified modifies the modifier (only partially, but significantly), and the meaning of the modifier then rests on a collective understanding of the color of the thing modified. In Chinese, as in Zulu, those two foci are typically designated, as Victoria Bogushevskaya observes, by “the standard combinations qīng tiān 青天 (blue sky) and qīng căo 青草” (green grass) (63). A grue sky is a blue sky, but grue grass is green grass.

Where this gets interesting for our discussion of Cathay is that the greater salience of focalization of the Japanese grue (ao) in the blue region leads Mori & Ariga to define the Chinese grue (qing) as “blue” over and over again, even in a few instances where it seems that the word ought to mean green. This is why I point out in the edition that the graph 青 “typically” or “almost always” means “blue” in Japanese. Not always, mind you, but almost always, typically, which is why it is repeatedly given as such in Fenollosa’s notebooks. The reviewer, however, sprays us with half a dozen Japanese terms in which ao 青 is combined with green things to mean “green” (of course, as also in Zulu) as though such a catalogue would overturn the simple fact that it means “blue” more often and in general use, and as though my putative ignorance of such combinations led to my supposed mistake. Here is the reviewer’s complaint in full, partially quoted earlier:

Although Kainan is rightly given more prominence than in previous English-language discussions of Cathay, he is also the victim of unjustified criticism regarding his interpretations of the Chinese poems. Billings claims, for instance, that it is the ‘Japanese linguistic prejudice’ of ‘Mori &/or Ariga’ that causes the translation of 青 (qing in modern Mandarin, ao or sei in Japanese), which describes the ‘grass’ in ‘The Beautiful Toilet’, to be rendered as ‘blue’ because this character is ‘almost always “blue” in Japanese’. In fact, 青 frequently refers in Japanese to what English speakers would consider to be ‘green’ or ‘fresh’, such as 青信号 (ao shingō, a green traffic light), 青年 (seinen, a standard term for ‘young person’), 青梅 (aoume or ōme, green ume fruit widely sold in Japan), the common female name 青葉 (Aoba, lit. ‘green leaf’, several of whom I teach, all in no doubt it means ‘green’), or indeed 青青 (aoao, ‘lush green’, as in 青青とした草原, aoao to shita kusahara, a ‘lush green field of grass’), the very term used in ‘The Beautiful Toilet’. Countless other examples could be added. To any Japanese speaker, let alone a renowned scholar and practitioner of classical Chinese poetry such as Kainan, it would therefore be an absurd mistake to think that the 青青 in ‘The Beautiful Toilet’ means what Fenollosa’s notebooks interpret as ‘blue blue’. The error must therefore have occurred in the translation into English (and thus to be attributed to Ariga or another Japanese translating for Fenollosa) rather than between the Chinese and Japanese.

Perhaps what baffles me most about this complaint is that reviewer’s main idea here is, in fact, my idea, not a corrective to my idea: that glossing qing 青 as “blue” is an error of translation from Japanese to English, which I make unmistakably clear in the edition. Ariga is the only one speaking English. (How many times must I repeat this?) I never claim that Mori thought that grass is “blue,” as we say in English, even in a Chinese poem, but rather that between Mori & Ariga, a Japanese linguistic bias around a very small number of words manifests itself in the notebooks. It is contrary to both the spirit and the letter of my critical apparatus to claim that these critical observations of mine are an attack on Mori’s scholarly competence. 

And the fact remains that the notebooks do say that the grasses on the riverbank are “blue blue.” As are the willows. Indeed, qing 青 appears 16 times in the poems Pound translated for Cathay, and in 15 of them it is glossed in the notebooks as “blue” and “blue” alone. That’s what I call statistically significant. In the one remaining instance (for “Mulberry Road”), it is initially glossed as “green” but then emended to “blue,” and it is “blue” that is then used for the paraphrase. See Figure 1.


what it modifies

probable meaning



Beautiful Toilet

青青p. 90, line 1


river bank grass


(and lush)

blue / blue*


River Song

p. 103, line B0


willow color




River Song

青青p. 104, line B3


willow color


(and lush)



River Merchant

p. 119, line 4





Exile’s Letter

p. 172, line 38


green &/or blue



Exile’s Letter

p. 177, line 54






Exile’s Letter

p. 162, line 5





Taking Leave

p. 225, line 1


blue &/or green



City of Choan

p. 237, line 5





Old Idea of Choan

p. 268, line 2





Ancient Wisdom

p. 311, line 7


blue, green, or black; east

blue / name


[proper name]

Mulberry Road

p. 262, line 7

silk (strands)

blue, green, or black

green* / {blue}


Light Rain (Hirai)

青青p. 216, line 2


willow color


(and lush)

blue / blue

green green*


Figure 1. Instances of qing  in Fenollosa’s notebooks for poems in Cathay. An asterisk indicates the reading Pound chose, if any.

And it’s not just Mori & Ariga. The last row on the table is taken from Fenollosa’s notes with another Japanese teacher, the English-speaking Hirai Kinza, who also gives “blue” for his one-word gloss on the color of the willows, though he changes it to “green” when translating the entire line. Once again, the Chinese graph 青 in isolation is translated by Japanese scholars as “blue” undoubtedly at least partly because of the linguistic interference of the Japanese word ao 青. Indeed, the connotative value of blue in ao is apparently so strong that it sometimes overrides the proper sense in some compounds in the minds of native speakers. When I asked Prof. Kawai to translate aoume 青梅 (green plums)—that “fruit widely sold in Japan,” as the reviewer puts it—as it appears in the passage from Rishi kōgi, Prof. Kawai rendered it as “blue plums” (see above). Even though he knows perfectly well how green ume are, having seen them his entire life, the sight of that word ao 青 led him instinctively to the English word “blue.” I hope it is not really necessary to say that Prof. Kawai is an extremely learned scholar with a fine literary sensibility and an exquisite command of English, who, as a translator of Shakespeare, has a deep understanding of translingual issues.

Oddly, among the supposedly “countless examples” of compounds in which ao means “green” when combined with well-known things that are known to be green, the reviewer names the “green traffic light” (ao shingō 青信号). But, in fact, like the linguistic nuances we have been discussing, traffic lights in Japan have historically differed slightly in color from those elsewhere, and have been the topic of amusing human-interest stories in recent years, such as “The Japanese Traffic Light Blues: Stop on Red, go on What?” in The Japan Times (2013), and “Red for Stop, Grue for Go: How Language Turned Traffic Lights ‘Bleen’ in Japan” from the podcast 99% Invisible (2018). (“Bleen” is an alternative portmanteau word for blue+green.) Although they have now mostly been replaced, many of Japan’s “green” lights were once evidently as bluish a blue-green as would be allowed by international standards—a concession, it is widely believed, to the common understanding of the word ao 青 in Japan as “blue.” Nevertheless, a rare few legacy lights remain here and there around the country which truly look more bluish than greenish. (The internet’s favorite example is the one in Nikkō, shown in the articles just cited and still visible in the Google Maps Street View linked in the note.[4]) Historically, it is possible that ao may originally have been closer to qing as a true grue, but its meaning shifted toward blue in the Heian period, about a millennium ago, with the introduction of midori (green) as a new color word (Richarz, online). Thus, as Saussure might have predicted for us, like the Russian macro-color sinij which sometimes signifies both “dark blue” and “light blue,” but is typically used to signify “dark blue” in opposition to “light blue” (goluboj), so also ao sometimes signifies both blue and green (especially in certain predictable combinations or classical formulas), but is typically used to signify blue in opposition to green (midori).

Let me stress that the aim of the table in Figure 1 is not to demonstrate that Mori & Ariga really were wrong, but rather to show that Japanese scholars really do choose “blue” over “green” when translating qing 青 into English, even occasionally when the context suggests that “green” might be the better choice. But only the examples in the first four rows and the final row are unambiguously green. The “jade” in “The Exile’s Letter” (shorthand for the original Chinese poem) could theoretically be green or blue-green. The “cow” in “Old Idea of Choan” is probably black, or just “dark,” as Pound deftly renders it in one of the very rare instances in which he intuitively corrects Fenollosa’s cribs. The color of the “gate” in “Ancient Wisdom” is impossible to know: early glosses note that people called it qing because of its color—though, as usual, without specifying exactly what that color is (blue, if I had to guess)—and yet its meaning may be partly overdetermined by the word’s association with the cardinal direction East and the fact that it was the Southeast entrance to Chang’an (modern-day Xi’an). And it’s just as impossible to know what the color of the silken strands in “A Ballad of the Mulberry Road” were meant to be, though black and green are probably good guesses. The “sky” in “The City of Choan” is obviously blue, as are the “clouds” in “The Exile’s Letter,” though we understand that they must have only hints of blue showing through them, and the compound is a metaphor for high office. In sum, if Ariga was a little careless for the sake of convenience in consistently using the English “blue” to translate what was probably Mori’s ao, those readings were not therefore consistently wrong.

As for what color Mori himself had in mind, for the most part we can only wonder. But in one extraordinary instance the notebooks actually do tell us that he was thinking blue in opposition to green, though Hugh Kenner overlooked these lines when he examined the crib for this poem half a century ago, before the notebooks were publicly available. As some readers may recall, Kenner mused on this very problem with characteristic savvy in The Pound Era, wondering how grass and mountains might end up “blue,” and wondering (I believe that’s still the right word) whether Mori might have been exercising his “poetic judgment” in the matter. The passage is worth revisiting at length:

Time and again the only meaning of “correct” is “traditional.” We can sometimes say that a word cannot possibly mean what a translator has written in response to it; more often we can say that he has not written what readers of the original usually understand. The poem called by Pound “The Beautiful Toilet” opens in Cathay 

Blue, blue, is the grass about the river

but in all other translations “green, green.” The character in question is Ch’ing [qing], Mathews 1168, the 174th radical, glossed in Mathews’ Chinese-English Dictionary “The color of nature, green, blue, black,” and in Karlgren’s “Green; (various kindred colors;) blue, azure, greyish.” It is a component in compounds meaning “green grass,” “blue sky,” “young foliage.” In Mei Sheng’s poem (if it is Mei Sheng’s; the usual citation is to a collection called “Nineteen Old Poems”) custom seems to demand the reading “green,” no doubt because grass is present, but Fenollosa has only “blue, blue,” and one can perfectly well imagine a light in which river grass is bluish. Was Ariga’s mind’s eye on such a light when he told his pupil to write the English word “blue”? Or was Ariga uncertain of the demarcation between the English words “blue” and “green”? The Chinese mind sees one color as a shade of the other. When the same character occurs at the beginning of the poem Pound calls “Taking Leave of a Friend,” his “blue mountains” is more plausible than “green mountains,” yet an expert gloss by Professor Roy E. Teele again elects green. Does either “blue,” or one of them, reflect Kainan Mori’s poetic judgment? Is there a traditional way to understand either poem? If so, did Mori know it, or not? (216-17)

Kenner is partial to Pound’s “blue mountains,” though I think most readers would associate a forested mountain with green, as indeed I did when I lumped “mountains” together with the “willows” in my note. But I have thought about that poem quite a bit since the edition was released, and I have decided that Mori’s reading of “blue” may be the better translation, if one had to choose between “blue” and “green.” The evidence for Mori’s reading in this case appears on the verso page of the crib proper, where Fenollosa typically recorded Mori’s additional commentaries, in a description of the semantic parallelism of the poem’s opening couplet:

First two lines are in pairs

blue mt. + white water — white under sunshine

green would be too commonplace (Cathay 225)

Apparently, blue mountains and a white river were both counterintuitive enough to Mori for him to comment on them here, though there is no such nuanced aside in his edition of TSX, and the poem is not covered in Rishi kōgi (illustrating once again the value of the notebooks over other texts in preserving Mori’s teachings). My gut tells me, however, that Mori is not quite right about his explanation for the diction; and that hunch is confirmed by a word search of an online database of Quan Tangshi (Complete Tang Poetry) showing that qing 青 is by far the most common color-word used with shan 山 (mountain), several times more frequent than the alternatives: bi 碧 (light green or blue), cui 翠 (bright blue green), cang 蒼 (dark blue or green), 綠 (green), yu 玉 (jade green), or the extremely rare zi 紫 (purple) (Sturgeon). In other words, qing can’t have been chosen to avoid a commonplace if it is the commonplace. (And the commonplace may have had as much to do with the metrical convenience of the double ping 平 [level] tones in qīng shān 青山 [grue mountains] as it did with the color itself, given the strict prosodic requirements of the regulated verse form.) But the point is that Mori’s comment only makes sense if one assumes that qing 青 means blue and not green—i.e., the “commonplace” green that Mori thought Li Bai was trying to avoid. Note that Mori is not trying to disambiguate the term, but rather to offer an explanation for why the poet would choose such a surprising color as blue to describe mountains. Presumably, the word Mori used to describe those mountains was ao (grue, blue) and the word for the commonplace was midori (green). Ariga translated it. Fenollosa recorded it. In sum, Mori thought that qing 青 really did mean blue here in opposition to green, just as ao 青 typically does in Japanese (even though, once again, it can also sometimes include it).

The evidence thus shows that Kenner was right to suspect that this reading reflects Mori’s poetic judgment—fine poetic judgment, I might add—nudged in that direction, I would argue, by the linguistic bias of Japanese. Unfortunately, we don’t have the benefit of a commentary or even a paraphrase for the blue grass in “The Beautiful Toilet” since that poem was merely mentioned as an example in the course of a historical overview as opposed to one of the poems Mori & Ariga taught Fenollosa in detail. In the cribs for “The River Song,” the willows are “blue” both in the title of the poem that Pound notoriously conflated with the previous poem, and also in the second couplet of that second poem. Of course, nobody can say with certainty what color Mori saw in his mind’s eye when he called them ao (as I assume he did), or even what color Ariga himself saw in his mind’s eye when he conjured up the English word “blue,” all of which wondering Kenner has already beat us to, anyway. But we can say that willows are not blue (pace Thomas Minton and Josiah Spode), and it would be “foolish” to think otherwise (as the reviewer insists), at least if we are thinking referentially. If I had to guess, I would guess that both Mori & Ariga saw a bluish green in the graph 青 modifying willows, but I also recognize that it is an assumption to think that they saw anything at all—i.e., that they necessarily treated the poem imagistically rather than philologically, whereas qing may have been chosen less for its precise chromatic denotation than for poetic convention and the quality of the diction: qing 青 is generally a more “poetic” word than 綠 (green), especially in this context.

As for Pound, we know (I think) what he was thinking from his poem. Here is the crib for the first instance of the blue willows in “The River Song”:





(for first time)


What that means is:

[alongside] Dragon Pond, [the] willows’ color [is] newly green (龍池柳色初青)

In order to make phanopoetic sense of that egregious “blue,” however, Pound understands it as the color of the pond, not the willows, which he then motivates by introducing a reflection of the sky:

I looked at the dragon-pond, with its willow-coloured water

Just reflecting the sky’s tinge. (Cathay 103)

Strictly speaking, there are no willows here, only a blue sky reflected in willow-colored pond water. Pound’s solution to the oddity of “blue” here is ingenious, but results in a completely new sense. Let me stress again that this is one of the extremely rare instances where the cribs are directly misleading or show any Japanese linguistic bias. I wonder what Pound would have done if the crib had simply read “green”? (My guess is: something not quite as interesting as this.)

Here is the second instance of blue willows in that poem, only a few lines later, but this time with a paraphrase as well:








South of the pond (sunny) the willows are already half blue.

The line means: 

The color of the willows south of the pond is partly lush and green (池南柳色半青青)

Pound renders it as:

South of the pond the willow-tips are already half-blue and bluer. (Cathay 104)

Note how Pound’s “and bluer” strives to express the sense of that extra character in the reduplication of qingqing 青青 (which connotes a “lushness”), but otherwise he follows the paraphrase almost verbatim. He seems not to have made the connection between these two lines, which might have helped him sort them out. Indeed, not recognizing that the former is actually the title to a new poem, he failed to recognize that the willows announced in that title are the very same that reappear in the latter. Is it “absurd” to think that Pound imagined these willow tips as half-blue? Or even bluer than half-blue? I wonder.

As for Kenner’s question about a “traditional” reading, I have surveyed the traditional commentaries, and I have found that no Chinese scholar over the centuries has ever thought that a qing-colored mountain required a gloss, as I was unsurprised to learn. We can, however, turn to modern annotations by contemporary scholars who typically describe the qing mountains in this poem as green—which is not to say that they necessarily know better than Mori did what Li Bai was thinking in the early 700s. As those of us who have been put through the paces know very well, traditional Chinese pedagogy requires students to translate classical verse into modern vernacular Chinese (baihua 白話); and you’ll find any number of modern editions of Tang poetry for study or appreciation that provide just such intralingual translations, which serve as a valuable record of readings. Of the dozen or so editions I happen to have on my shelf, most merely repeat the term as if the color of a qing mountain were self-evident (as in Yu 2:933; Ma 188; Zhao 138; Lai 141; Zhan 289, who glosses the bamboo horse, but not this; and even Zheng 162, in an edition intended for primary school students, which I find profoundly humbling). Mori’s edition of TSX does the same (3:59.) Of those that do specify the color, qing is invariably described as some sort of green, such as: a) qingcui 青翠 (verdant, fresh and green) (as in Xu Zhengzhong 140; Qiu and Liu 101; and Xiao 138); b) cangcui 蒼翠 (green, verdant) (as in Huo et al. 208; Sha et al., 180); or c) congyu 葱郁 (verdant, luxuriantly green) (as in Li 135). 

Indeed, it may not be an exaggeration to say that the more salient focalization of the Chinese grue on green is itself a linguistic bias that may prevent modern Chinese readers from visualizing how much blue is also signified by qing in some contexts. Perhaps the most fascinating example comes from Xu Yuanzhong 許淵沖 (1921–2021), the much venerated late professor emeritus at Beijing University who wrote some 50 books on Chinese poetry. In a sort of demonstration on translation, Xu translates qing both ways. In one version, which he calls “more faithful to the word,” he gives “green hills,” while in another version which he dubs “more beautiful and poetical,” he gives “blue mountains.” It seems likely that he was looking over his shoulder at Pound, but the fact remains that he found Pound’s version (which is Mori & Ariga’s version) the “more beautiful and poetical” one. Unsurprisingly, the only Japanese translator of Li Bai’s work into English, Obata Shigeyoshi 小畑 薰良 (1888–1971), echoes Mori & Ariga’s “blue mountains” in 1922, though he is usually at pains to distinguish his translations from Pound’s, whose “extravagant errors,” as he put it, had prompted him to write his own, more accurate versions (94; vi). All of this is perhaps best summed up by the person who has probably thought about it more than anyone else, Huang Lirong 黃麗容, who writes in her monograph on color in Li Bai’s poetry: 青色可含藍,綠和黑色,但詩人實將青色指為藍,綠或黑色已不可知 (The color qing encompasses blue, green, and black, but when the poet is using qing to indicate blue, green, or black is now simply unknowable) (105). 

Remembered Hills

Many years ago now, Qian Zhaoming, who was ever keen to defend Pound’s aesthetic choices as philologically accurate ones, suggested an explanation for those blue mountains: 

in Chinese landscape painting mountains in the far distance are depicted as blue rather than green to show a “linear perspective.” This is because, as one Chinese art critic [Jin Xuezhi 金学智] explains, “the air is not entirely transparent. Therefore the densely-wooded mountains in the near distance are deep green [nong lü 浓绿] while those in the far distance look pale [greenish] blue [danqing 淡青] or [pale] violet [dan zi 淡紫]; further away still they are even paler until they become void of color and their outlines obscure.” (60, my additions)

Qian may touch his thumb to the scale when translating danqing 淡青 (pale grue) as “blue,” but its defining opposition with “green” is just what we saw in Mori. (For what it’s worth, Qian miscited the page, the year, and even the name of the Chinese journal in which this article appeared, but I take that to be an innocent mistake, not systemic carelessness or a diversionary tactic.) I find this idea extremely compelling in large part because of the ocular proof before me. The eponymously green mountains—the verts monts—of “The Green Mountain State” of Vermont look unmistakably blue from my window even now as I write this sentence, a pale blue with a greenish tint. And they are blue in Shropshire, too, as the Lad knows well: “Into my heart an air that kills / From yon far country blows: / What are those blue remembered hills, / What spires, what farms are those?”).[5] Just last week, on my way home from the hardware store, as if in prophetic response to my musings, I drove past a spot on the highway where I could see three ridgelines in a row receding into the distance: a green rise of low hills about 2 miles away; a slightly higher blue-green range beyond that about 5 miles away; and an unmistakably blue ridge of mountains beyond that about 15 miles away. Mathews’ “color of nature” used to seem to me to be just a little bit silly, to be honest, a little too figurative to be precise. But on a sunny summer day when the sky is clear blue overhead, and the grass is bright green underfoot, it’s thrilling to think there’s a single word for that. 

Part of what I love about Kenner’s wondering is how persuasively it suggests that there is no “correct” reading, only perhaps a “traditional” one, which in this case was never explicitly recorded. One thing we do know, however, is that qing 青 anciently meant “green,” its original sense related either to vernal growth or to green pigments; and although words evolve, as we all know, this one may continue to lean greenward a bit more for that reason. Bogushevskaya confirms this conclusion on multiple counts in “Chinese GRUE: On the Original Meaning and Evolution of Qīng 青” which summarizes the most important etymological, phonological, and archeological studies of the word; as does Peter Boodberg in his almost Joycean “cedules.” After discussing examples of even greater textual uncertainty in Homer and Sappho, Kenner ends his reflection on qing by suggesting that Mori may simply have missed something: “Yet today’s understanding of what many details in many poems may mean is steadied by glosses that come down with the texts, and perhaps did not always come to Fenollosa’s instructors in Tokyo. . . . Small or large errors committed in Tokyo cast shadows on Cathay’s page” (218). From what I’ve seen, however, Mori did have access to as many traditional commentaries as anyone else in the materials he was using; and the errors committed in Tokyo (those that really are errors) were small and only very few and far between. What was cast from Tokyo were not primarily shadows, but a radiance and a glory which—like the yao 耀 on Cathay’s cover—continue to shine from the pages within.



Backhaus, Peter. “The Japanese Traffic Light Blues: Stop on Red, go on What?” The Japan Times. 25 February 2013. <>

Biggam, C. P. The Semantics of Colour: A Historical Approach. Cambridge University Press, 2012.

Bogushevskaya, Victoria. “Chinese GRUE: On the Original Meaning and Evolution of Qīng 青.” L’analisi linguistica e letteraria. XXIII (2005): 61-76.

Boodberg, Peter. “Cedules from a Berkeley Workshop in Asiatic Philology (With a Postscript by S. H. Chen)” (Berkeley: Self-published, 1955): 1–39. 

Cathay, see Pound.

Cohnitz, Daniel and Marcus Rossberg, "Nelson Goodman." The Stanford Encyclopedia of Philosophy (Summer 2020 Edition). Edward N. Zalta, ed. <>.

Gilonis, Harry. Eye-blink. London: Veer Books, 2010.

Housman, A. E. “XL. Into my heart an air that kills.” A Shropshire Lad. [London, 1896]. Academy of American Poets ( <>

Huang Lirong 黃麗容. Li Bai shi secai xue 李白詩色彩學 [Color Studies in the Poetry of Li Bai]. Taibei: Wenjin chubanshe, 2007.

Huo, Songlin 霍松林, Shen, Wenfan 沈文凡, and Li, Bohao 李博昊, eds. Tangshi sanbaishou Mingjia jiangjie 唐詩三百首名家講解 [Authoritative Explanations on Three Hundred Tang Poems]. Changchun: Changchun chubanshe, 2008.

Jiang Xun 蔣勳. Shuo wenxue zhi mei: Pinwei Tangshi 説文學之美: 品味唐詩 [Speaking of the Beauty of Literature: Appreciating Tang Poetry]. Taibei: Youlu wenhua, 2017.

Jin Xuezhi 金学智.Wang Wei shi zhong de huihua mei 王维诗中的绘画美 [Pictorial Beauty in Wang Wei’s Poetry].Wenxue yichan 文学遗产 [Literary Heritage]. (1984) vol. 4: 55-66.

Kay, Paul and Luisa Maffi. “Number of Basic Colour Categories.” Dryer, Matthew S. & Haspelmath, Martin, eds. The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology, 2013. <>

Kohlstedt, Kurt. “Red for Stop, Grue for Go: How Language Turned Traffic Lights ‘Bleen’ in Japan” 99% Invisible. 26 November 2018. <>

Lai Fangling 賴芳伶, ed. Tangdai shixuan: Da Tang wenhua de qipa 唐代詩選: 大唐文化的奇葩 [Selection of Poems from the Tang Dynasty: A Florilegium of Tang Culture]. Taibei: Shibao wenhua, 2012.

Li Mengsheng 李梦生, ed. Qianjia shi quanjie 千家诗全解  [Poems by the Thousand Masters Fully Explained]. Shanghai: Fudan daxue chubanshe, 2007. 

Ma Dongyao 马东瑶, ed. Tangshi sanbaishou 唐诗三百首 [Three Hundred Tang Poems]. Beijing: Beijing Shifan daxue chubanshe, 2014. 

Mathews, Robert Henry. Mathews’ Chinese-English Dictionary. Revised American Edition. Cambridge, MA: Harvard UP, 1975.

Mori, Kainan 森 槐南. Rishi kōgi 李詩講義 (Lectures on Li’s Poetry). Tokyo: Bunkaidō shoten, 1914.

Obata, Shigeyoshi. The works of Li Po, the Chinese poet. New York: Dutton, 1922.

Pound, Ezra. Cathay: A Critical Edition. Ed., Timothy Billings. With a foreword by Haun Saussy and an introduction by Christopher Bush. New York: Fordham UP, 2020.

Sturgeon, Douglas, ed. Zhongguo zhexueshu dianzihua jihua 中國哲學書電子化計劃 (Chinese Philosophical Texts Digitization Project):

Qian, Zhaoming. “Translation or Invention: Three ‘Cathay’ Poems Reconsidered.” Paideuma, vol. 19, no. 1/2, 1990, pp. 51–75. <>

Qiu Xieyou 邱燮友 and Liu Zhenghao 劉正浩, eds. Xinyi Qianjiashi 新譯千家詩. Taibei: Sanmin shuju, 2014.

Richarz, Allan. “According to Japanese Traffic Lights, Bleen Means Go.” Atlas Obscura. September 12, 2017. <>

 Sha Lingna 沙灵娜 , He Nian 何年 et al. Tangshi sanbaishou 唐诗三百首全译 [Three Hundred Tang Poems Fully Translated]. Guiyang: Guizhou renmin chubanshe, 2008. 

Yu, Xianhao 郁賢皓, ed. and trans., Xinyi Li Bai shi quanji 李白詩全集 [The Complete Poetry of Li Bai, Newly Translated]. 3 vols. Taibei: Sanmin shuju, 2011.

Waley, Arthur. “The Poet Li Po.” The Asiatic Review. 15, no. 44 (1919): 584–612.

Wellen, Paul. “Analytic Dictionary of Ezra Pound’s Chinese Characters.” Paideuma. (Winter 1996), 25.3: 59-100. <>

Xiao Difei 萧涤非 et al., eds. Tangshi sanbaishou Jianshang cidian 唐诗三百首鉴赏辞典 [Three Hundred Tang Poems, A Dictionary for Appreciation]. Shanghai: Shanghai cishu chubanshe, 2006.

Xu Yuanzhong 許淵沖. 300 Tang Poems (Classical Chinese Poetry and Prose Series) (English Edition). Beijing: China Intercontinental Press: 2012. 

Xu Zhengzhong 許正中. Tangdai lüshi xi shang 唐代律詩析賞 [Analysing and Appreciating Tang Dynasty Regulated Verse Poetry]. 2nd revised edition. Taibei: Dongda, 2004. 

Zhan Ying 詹锳 et al., eds. Li Bai shixuanyi 李白诗选译 [Poems by Li Bai Selected and Translated]. Nanjing: Fenghuang chubanshe, 2011. 

Zhao Changping 赵昌平, ed. Tangshi sanbaishou quanjie 唐诗三百首全解. Shanghai: Fudan daxue chubanshe, 2006. 



[1] If you read Japanese, you can see for yourself here: If you don’t, grab a friend who does.

[2] The 5th-century Hou Hanshu 後漢書 (Book of the Later Han) recounts a story about a 1st-century official named Guo Ji 郭伋 who on his first inspection tour of a newly assigned prefecture was met at the gates of a one city by hundreds of little boys all straddling bamboo hobbyhorses (騎竹馬) who had come out to greet him, excited by his reputation. They then asked him when he would return so that they could prepare to greet him again in the same way. The point of the story is that Guo Ji was big-hearted enough to go to the trouble of arriving a day early and camping outside the city in order to be sure that he did not disappoint the boys by failing to arrive on the appointed day (後漢書·郭伋傳). I do own one recent annotated collection of Tang poetry with notes so extensive that zhuma 竹馬 is glossed as “using a bamboo pole as a saddle horse” (以竹桿當馬騎), suggesting that such a toy may not be as recognizable to students cramming for exams as it once was (Zhan 45).

[3] I never tire of recommending this word cloud:>.

[4] Google Map Street View for the grue traffic light in Nikkō: https:/,139.6043914,3a,15y,256.43h,106.59t/data=!3m6!1e1!3m4!1sBoKwI2fC5acZlS8gBS72Ig!2e0!7i13312!8i6656&gt</a; Accessed 1 August 2021.

[5] Harry Gilonis slyly alludes to Housman in his own wryly wrought version of this line in Eye-blink, which I highly recommend to anyone with an interest in adaptations of Chinese poetry.