Sino-Platonic Papers, 75 (February 1997)

The Prestige of Writing: 文 , Letter, Picture, Image, Ideography1

Haun Saussy

The disparagement of writing is a motif common, I suppose, to all traditions that have writing.2 Writing is often seen as inadequate to represent speech or thought. But another response to the inadequacy of writing has been to exalt some other kind of writing—occasionally a language reformer’s pet project, but more frequently the writing of the angels, the writing of the citizens of some utopia, of the scholars of some faraway kingdom, or of the forces of nature itself. Imagined writings of this sort telescope critique and critique’s wishful compensation. They attribute wonders—praestigia—to a medium most often noticed in its falterings.

Since Chinese writing became known in Europe, it has often been pressed into service as the model of this perfected writing.3 This enthusiasm must appear outlandish to those whose ‘native’ writing-system is Chinese.4 But it is not enough to show that the indigenous and foreign perceptions of Chinese writing are at variance, or even that the tales told of Chinese script do not stand up to linguistic scrutiny: there is an inventive element to all intercultural interpretation, a fit between its observations and the intellectual needs of its proponents, that expert testimony simply shoves aside. The proper way to analyze an intellectual tangle of this sort, it seems to me, is not to hold it to the standard of specialist univocity, but to situate it ethnographically among the conceptions it echoes or answers. Which aspects of which utopias still beckon, and which have definitely gone on to feed intellectual history, is another question deserving patient consideration.

1. The virtue of wen

The sixth-century literary critic Liu Xie never set out to diagnose what might cause twentieth-century Europeans and Americans to fall in love with Chinese literature, but in the “Yuan Dao” (Tracing the Way) chapter of his Wenxin diaolong he certainly put his finger on it. “The virtue of wen,” says Liu Xie, “is great indeed: for was it not born together with heaven and earth? . . . The mind comes into existence, and language is established; language is established and wen gives forth its brilliance—the very way of nature” 文之為德也 大矣，與天地並生者 何哉!. . .  心生而言立，言立而文明， 自然之道也 .5 The passage and its predecessors and variants throughout the Chinese tradition have inspired eloquence, subtlety, etymological zeal and much sympathetic imagination.6 Liu Xie is certainly no disparager of writing—if he measures the writing of humans against the majestic writing of nature, it is to enlarge the former. Wen 文 , the word I have hesitated to translate, smooths the path of Liu’s argument. It combines in its several meanings many of the attractions of Chinese literature. Wen is (to cite several dictionaries at once) “markings; patterns; stripes, streaks, lines, veins; whorls; bands; writing, graph, expression, composition; ceremony, culture, refinement, education, ornament, elegance, civility; civil as opposed to military; literature (specifically belletristic prose in its distinction from poetry). . .” The coexistence of these various meanings is suggestive; to say “wen is wen” is never just a tautology. Indeed a great deal of writing about the Chinese language and literature (mostly but not exclusively by non-native speakers) has said exactly this, drawing on textual evidence and close reading to amplify the assertion that, for example, “wen in sense 16 is really or ultimately the same thing as wen in sense 7.” Most frequent is the assertion that wen in the sense of literary writing is best illumined by analogy to one of the other senses: literary writing as inherent pattern, writing as civilization, literature as an outgrowth of the Chinese writing system.

 It is this last possibility—surrounded by overtones of some of the others—that will occupy the ethnographer of literary culture. For it is notable that, despite the decay of many simplistic illusions about the singularities of the Chinese written language, the motif of a contrast between alphabetic writing and ideogrammatic writing as modes of representation survives to direct many accounts of the contrasts between ‘Chinese’ literature and ‘Western’ literature. (This level of generality comes with the territory, unfortunately; maybe after a few more decades of mutual research the geographical blocs will melt.) It is as if the initial error about writing-systems, now discredited, has had to migrate into considerations of style, genre, ideology, philosophy, and so forth. Accordingly I shall not insist on a genealogical connection among my exhibits, only a striking homology; and though I may have skeptical words for this or that instance of the contrast of which I speak, not all instances are equally pernicious or unfounded.

I shall first put the matter in the most uncontroversial way I can. Most writings on the concept of wen subscribe to a view of the directly semantic and symbolizing nature of Chinese writing: unlike the alphabetic transcription of a word, which analyzes into letters without intrinsic meaning—cenemes, in Hjelmslev’s (1985 [1938].159) terminology—a Chinese term, already meaningful, can, in the view of those I shall call ideogrammatists, always be split or traced to simple pleremes, or inherently meaning-bearing units. At the level where the pleremes of Chinese writing break down into mere strokes, there is nothing left for the theory of writing and language associated with the word wen to discuss. And yet that is the level at which writing appears as pure syntax, as figures about which it is appropriate to ask how they are put together, not what they stand for. The ‘syntactic’ interrogation of writing holds gratifications which I will maintain are no less great for being delayed. (‘Syntax,’ the reader will have surmised, is here taken in the broadest sense. In Greek the term meant ‘composition’; its use by modern linguists to denote word-order relations is a narrow one that obscures the word’s place in intellectual history.7) I shall first discuss the theory of plerematic writing in some of its more baroque forms, and end by arguing that the eventual fading-out of the pleremes into pure cenematic structure is not the end of wen, or of civilization either for that matter; that is, I will try to argue for the interest of the strokes as strokes, not primarily as signs of anything; and I will work to win some of the authoritative classical texts over to my side.

2. “It passes between two terms”

My first example will be drawn from a book that nobody interested in the American understanding of China can ignore, Ernest Fenollosa’s The Chinese Written Character as a Medium for Poetry.9 Having a professional interest in Chinese literature, I know I expose myself to trouble simply for having mentioned Fenollosa’s name. As we learn very early in our training, Fenollosa was an enthusiast: in his wonderment at the Chinese language he vastly overestimated the number of primary pictograms in the writing system and saw images and parables where a more sober palaeographer would have seen combinations of phonetic clues. The profession has never forgotten it.10 The profession is in this case right, but since the profession’s response to Fenollosa’s ambitious surmising has always been empirical—carried out on the level of examples, and presupposing the interpretation of those examples—it may be useful here to allow Fenollosa to present the system of his ideas so that they may provoke a principled, that is, not merely atomistic, answer. From a different quarter, Jacques Derrida (1967.139-40) has held up Fenollosa/Pound as an example of grammatology on non-phonetic, non-phonologocentric principles — not, perhaps, a wise choice, for reasons that should soon become apparent. Read with sufficient care, Fenollosa’s rejection of phonetic writing reproduces in particularly unequivocal terms the very intuition-grounded epistemology that Derrida’s critique of phonetic theories of writing was meant to undermine. Nonetheless I think the intrinsic interest—let us say the historical interest, just to be safe—of Fenollosa’s essay warrants investigation.

Fenollosa cannot find words enough to praise the Chinese character, which for him is “something much more than arbitrary symbols. It is a vivid shorthand picture of the operations of nature. . . . Chinese poetry. . . speaks at once with the vividness of painting and with the mobility of sounds. . . . In reading Chinese we do not seem to be juggling mental counters, but to be watching things work out their own fate.” (Fenollosa 1964.8-9; italics Fenollosa’s.) In explaining how this is so, Fenollosa with remarkable subtlety and prescience swerves away from the usual definition of the “ideograph,” the definition that makes the character a picture of a thing.

Now the picture has long had its place in Chinese epigraphy. The great etymologist Xu Shen (AD 30-124), in the preface to his dictionary Shuo wen jie zi, put near the top of his list of character-formation techniques that of xiang xing 象形, the “imaging and shaping” of the thing a character referred to.11 Xiang xing was preceded in the list only by the zhi shi 指事, or self-designating, type of character. Following the “Appended Verbalizations” chapters of the classic Yi jing or Book of Changes (from which the label xiang xing probably derives), Xu Shen understood the “image and shape” characters of the Chinese written language as intimately involved with the origins, perhaps even the basic nature, of that writing-system. For Xu’s first readers, indeed, the connotations of “image and shape” may have reached beyond etymology and into cosmogony. “Transformation by light and darkness, imaging, shaping, and sending forth—that is known as ‘generation’ or ‘life’,” says the Da Dai li ji.12  

Later philologists have followed Xu Shen’s lead, seeking out or supposing from comparative evidence a primitive picture-writing as the ancestor of the by now highly conventionalized characters.13 Writing must have begun as drawing—an idea commonly held in areas of alphabetic script as well. But where pictography is most vestigial, the movement away from it provides the principle of the interpretation of writing’s history. So Hegel (in the Encyclopedia) could easily dismiss the Chinese language as a backward medium of thought because, as he sensibly pointed out, most of the things worth thinking about are not representable in pictures.14

Fenollosa, however, as Davie (1955.34) points out, turns the discussion in a new direction. He takes the root meaning of any Chinese character (and thus the root impulse in the elaboration of the writing system) to be the verb, a step in the direction both of greater linguistic sophistication and of a more insistent claim for the character’s iconic virtues. The picture-writing with which Fenollosa’s name is forever associated could not do what Fenollosa wants to see Chinese writing as doing, namely transmitting the intuition of force. All nature, says Fenollosa, is really a vast interplay of verbs, not of nouns; nouns or things are merely “the terminal points, or rather the meeting points, of actions” (Fenollosa 1964.10; see also 19). But the European languages of today, “thin and cold,” have lost their “primitive sap”; they are “cut and dried as a walking-stick” (24).15 One can recover the vital energy of the primitive verbs, says Fenollosa, by going back to the roots of the Aryan proto-language: in them we see what must count for Fenollosa as the perfect linguistic situation, expressions “forced upon primitive men by nature itself.”16 But the Aryan record is patchy, and thus only heightens the importance of the study of Chinese ideograms for “throwing light upon our forgotten mental processes” (21).

By centering his theory of the ideogram on the verb, Fenollosa reintroduces syntax, long absent from those discussions of Chinese which had taken as their starting point the character as picture. Pictures are “put together,” to be sure, but their “syntax” is not an ordering of the parts into a sequence or sentence, as Lessing had long before observed in his influential essay Laocoön.17 Now the sentence is (for Fenollosa) a category that the Chinese language and nature share. “The type of sentence in nature is a flash of lightning. It passes between two terms, a cloud and the earth” (12). So too, the Chinese sentence: subject - verb - object. Fenollosa makes Chinese syntax, on precisely the same principles as the ideogram, an image of nature: it is semantic, or meaningful, through-and-through, and that is what makes its particular uncluttered excellence.18 Ideogrammatic qualities reside not only in the individual Chinese character, but in the Chinese sentence which Fenollosa sees as depicting pictographically a complex of natural actions and relations laid out in a causal line.

Fenollosa’s primitivism may strike us as naïve, misguided, potentially vicious; but to be fair we must admit that it is in the service of a linguistic-critical project, a project of desedimentation. Fenollosa’s essay moves across the same territory as Nietzsche’s essay “On Truth and Lie in the Extra-Moral Sense”—an influential essay if anything in the history of literary theory is. “So what is truth?” asks Nietzsche, and answers his own Pontius-Pilate question with: “A mobile army of metaphors, metonymies, anthropomorphisms, in a word a sum of human relations that have been poetically and rhetorically heightened, translated and ornamented, and after long use come to seem canonical and obligatory to a people: truths are illusions of which one has forgotten that they are illusions. . . .”19 A decade or two later, Fenollosa asked himself the same question, and answered: “a flash of lightning pass[ing] between a cloud and the earth.” By just such a lightning-stroke the sentence-form was forced on primitive men by nature. “All truth has to be expressed in sentences because all truth is the transference of power” (12), Fenollosa goes on to say, glossing himself in strikingly Nietzschean language.20 The ideogram (and I now include the sentence as an expanded ideogram) is true for Fenollosa because it depicts truly, that is, with a transference of power, the operations of nature, which are always transferences of physical force.

Now there is little point in summoning up the thunder and lightning of present-day criticism to attack such naïve epistemic realism. Naïve it is, and self-proclaimedly so: for Fenollosa, as for Wallace Stevens in some of his anti-nominalist moods, the primitive man is the benchmark of perceptual accuracy because he sees naïvely, sees things as they are without the fog of culture, theology, or artifice.21 And the authors of Chinese writing, being, in this nostalgic reconstruction, similarly endowed with the gift of simplicity, should for the primitivist point the way back to the truth of undisguised appearances. As I say, this hardly merits criticism, but those who like to array their enemies in one United Front will be comforted by the close cousinage of Fenollosa’s epistemic standard with the bad habits of ethnographic (and colonial) exoticism. The Chinese—like their written language—are first of all alien, alien because uncivilized, and secondarily praiseworthy; if they were less unlike us, there would be less reason for the primitivist to hold them up as the model of poetic knowledge. In Fenollosa’s enthusiasm for Chinese word order we may even see a double exoticism, a European nostalgia complicated by a Japanese one. He had come to Chinese through Japanese teachers, and word order is the main thing that literate speakers of Japanese must consciously train themselves to relearn when they begin reading Chinese texts (as opposed to Japanese texts with a heavy Chinese-character component). Japanese word order, Fenollosa seems to suggest, is an imposition on or rearrangement of that natural scheme of things which Chinese (he believes) more directly reflects, and so the Japanese medium of his training in Chinese poetry simply drops out of the picture. In this way, however, the Chinese language loses—it would perhaps be better to say it discards—whatever is properly linguistic about it. In this construction, sentences don’t do anything that things that are not sentences don’t do; syntax becomes merely an instance of natural symbolism and disappears as a distinct category of event. Presumably this could not have happened in Japanese, where phonetic writing coexists with ideographic writing and word order is both more complex and more attention-getting.

3. “All Chinese poems are true”

Nowadays The Chinese Written Character as a Medium for Poetry is apt to look quaint: every undergraduate learns to distrust ideas of immediate natural signification, and even a superficial knowledge of Chinese is enough to show that many of Fenollosa’s etymologies, and indeed the drift of his whole enterprise, belong to fantasy. But what Fenollosa thought held for the Chinese written language often still appears, in diluted and distanced form, in the frequent characterization of the Chinese poetic language as “something much more than arbitrary symbols.” In part this must relate to the influence (through Ezra Pound) of Fenollosa’s ideas on modern poetry in English—an influence which has not entirely been a bad thing. (The vigor of art in any period owes a debt to absurd and even wicked thinking.) In part, too, prolonging the ‘ideogrammatic’ theme is a fitting expression of gratitude on the part of the Europeans and Americans who have become acquainted with the Chinese tradition: to utter straightforwardly what Liu Xie said in hyperbolic mode about the wen of heaven, earth and man is very good manners indeed. For the idea that writing, in its most sublime manifestations, speaks with the authority of nature is both inspiring to people in the Chinese tradition and impressive to people out of it. “Nature” is, after all, the one thing that ought to hold constant across differences among “cultures.” But what happens when a pattern of thought—even thought about Nature—crosses into a new context of ideas?

Fenollosa’s attack on classical logic denounced it as bound to the forms of language, not of nature (1964.25-29). His proposal was therefore to cast aside one form of mimesis, seen as artificial and obstructive, in favor of another, supposedly natural and transparent, one. Although the classical Chinese poets and critics never had a chance to be amused or challenged by Greek theories of mimesis, their modern-day expositors invariably open the discussion by rejecting the idea that literature (like artifice in general) imitates a prior reality.22 Stephen Owen made some of the crucial linkages in this rejoinder to Aristotle. The Chinese poem does not imitate or feign reality, but follows from it (Owen 1985.15); while “metaphor is fictional and involves true substitution,” the categorical correspondences one finds in Chinese poetry are “‘strictly true,’ based upon the order of the world” (61). “The traditional Chinese reader had faith that poems were authentic presentations of historical experience,” because the writing of poetry was conceived of as “a virtual transfer of substance” from experience to words and from words to reader, “an organic process of manifestation” (57, 58, 59; compare Jullien 1985.46-47). In a similar vein, Pauline Yu presents the Chinese literary universe as one in which “there are no disjunctures between true reality and concrete reality, nor between concrete reality and literary work.” A Chinese poem is not typically seen as a representation, but as “a literal reaction of the poet to the world around him and of which he is an integral part” (Yu 1987.35, with reference to Liu Xie). For Michael Fuller the Western problematics of subject and object (deeply connected to the theory of mimesis via the correspondence theory of truth and perception23) simply do not hold for the Chinese reader who is able to say: “On a certain day at a certain place a poet encountered a certain situation and wrote the poem. . . . In the classical Chinese tradition, all poems are true by their very existence; the only questions are how they are true and if their truth is of any significance” (Fuller 1993.21). At a time when the top buzzword in other branches of the humanities was “sign” with its predicates “arbitrary,” “differential” and “constructed,” scholars of Chinese poetry, wary of semiotics with its Aristotelian heritage, deliberately turned to an account of Chinese literature that emphasized the opposite qualities, insisting, above all, on the lack of disjunctures among things Western theories had disjoined. The commentatorial language of metaphor gave way to the language of literal meaning and, beyond that, of empirical reference. The shi 詩  poem, with its associations of sincerity, occasional composition, and emotional release, moved to the top of Chinese literary genres to become the “dominant” among them, the genre that guides the interpretation of literature in general.24 What is special about Chinese literature, according to the modern theorists of the classical tradition, is its hyperreality: the claim is that Chinese representation, unlike Greek mimesis, is always true. (Or efficacious, if “truth” seems to involve too Greek a claim of correspondence between word and thing.) Moreover, the truth of the representations is guaranteed by the means of representation. Joseph Roe Allen radically condenses this synecdochical poetic: “One might even say that wen is itself the cosmos, not a sign in place of something else, but a sign that is the very thing to which it refers” (Allen 1992.19). The “not. . . but . . .” structure of the formula summons the student of Chinese literature whose training is European to make a choice: either Aristotelian metaphor and the Saussurean sign, or the Chinese symbol which signifies by unveiling a corner of its real referent. One should remember that the generation to which these critics belong was the first to bring Chinese literature into the ambit of comparative poetics, which meant, at that juncture, taking a typological and categorical look at their materials rather than a historical one. This chronological fact goes some way to explaining the cognitive clash, for it is not certain what the preparatory groundwork for typological study would be, only that it involves a departure from the questions one can ask and solve from within a single literary history.

Now it should be obvious that Ernest Fenollosa and contemporary literary scholars are not often talking about the same things. Rather, they are talking about different things in the same way. The non-mimetic version of Chinese literary theory contrasts with Western attitudes in exactly the same way that the ideogram triumphs, for Fenollosa, over the defective logics and alphabets of Europe. Running through Fenollosa’s praise of the ideogram and the praise of Chinese poetics in present-day literary studies is a common thread of admiring sympathy for a system of representations that can present its claim to truth so powerfully. “Something more than arbitrary symbols” indeed! For the reader lucky enough to endorse all these interpretative programs at once, Chinese writing would be anchored in the vital representation of things, Chinese poetic speech in belief (or culturally sanctioned truth), and the lexical meaning of Chinese poems in inevitable empirical reference. Chinese would be the language that “simply HAD TO STAY POETIC; simply couldn’t help being and staying poetic in a way that a column of English type might very well not stay poetic” (Pound 1931.22).

Admiring sympathy admits of degrees, nonetheless. The critics cited above do not, of course, actually argue that all the statements in Chinese poetry are factual and natural for them. That position would not produce what the critics are preeminently engaged in producing, a contrast between one culture’s way of conceiving meaning and another culture’s. Owen reminds us again and again that the organic hypothesis is part of the equipment of the traditional Chinese reader and to be taken as such—anthropologically, not ontologically. “We speak here of the inclinations of readers and of a poet’s anticipation of those inclinations: the actual truth or untruth of poems, authentic presentation or manipulation for art’s strange motives, is not in question—these, in fact, are usually beyond recovery” (Owen 1985.57). Critics should be, and in normal circumstances are, able to discuss representations that somebody considers to be true without having to take a stand on whether or not they are true. In scholarly discourse, that is, the truth-value of a statement such as “literature expresses the dao of nature” should be subordinate to the syntax of a sentence such as “Liu Xie believed that literature expresses the dao of nature.” The contents of the object-language are relevant to scholarship only as they are bracketed and filtered by the meta-language of quotation and paraphrase. If that were not so, it would be impossible to discuss beliefs without adopting them, and that would cause most work in the humanities—above all the kind of culture-bridging work we are talking about here—to grind to a halt.

And yet in the study of Chinese poetics we find a fair case of incompatibility between the semantics of quoted language and its place in a quoting syntax. The object-language must at least threaten to break through the meta-language’s syntactic frame, for one mission of such cultural comparison is to call into doubt the inclusive, superordinate status of the meta-language these very critics use. If as outsiders we see the ‘organic manifestation’ theory of meaning as a construction, a fabrication, our experience will not reduplicate that attributed to the cultural insiders; far more, the two sorts of experience will certainly clash. “Construction” and its whole family of concepts (fabrication, fiction, persona, imaginary, etc.) represent, for the generation of critics being discussed here, a contemporary Western set of preoccupations that must not be allowed to run the whole show. One might phrase the issue faced by these comparatists thus: acknowledge the constructedness of the concept of ‘construction’ without seeming to fall back into an unreflective concept of nature. This properly insoluble imperative surfaces in symptomatic form. An example is the quotation marks Owen puts around the phrase “strictly true,” evidence of true discomfort with the language of truth. What is strictly true should, strictly, appear in quotation marks only once it has been deemed no longer strictly true—true perhaps for some other, but not for the speaker or his audience. The quoted matter says, “This is the way things are!” and the practice of quotation says, “One way of putting the way things are is, ‘This is the way things are’.” Pauline Yu, too, makes it clear that her subject is a tradition, and surrounds the naturalistic accounts to be found in the tradition with the unmistakable third-person markers “it is believed. . .” and “from the Chinese point of view . . . .” Such devices make it clear that Yu is not falling prey to Fenollosan nostalgia any more than Owen. But the very same devices also make it seem that someone within the tradition would not have seen the tradition as a tradition. Once you are inside the quotation marks, the ‘strictly true’ is just strictly true. That is to say, Chinese literature obliges us to entertain two (at least two) logics. The discipline of Chinese literary studies suggested by these writings is structurally historical: a profession of quotation-mark wielders handling texts from which all quotation marks are deemed to be absent. (Historia in the Herodotean sense: a “report of things heard,” however exotic or incredible.) It is therefore a delicate syntactic situation, one in which relations of sequence, subordination, agreement, modality, enclosure and correlation are paramount. The subtlety and complexity of these relations could hardly be at a further remove from the theory of nature-writing so memorably put forth by Fenollosa. Since, however, the self-definition of the field has so often been made to hinge on the decision to grant Chinese wen immediate representational powers, it seems that the ideogrammatic method has won by a hair—even if at several generations’ remove, and in the realm of literary criticism, not of philology.

4. The lines of force

Mercifully, the word “ideogram” is falling out of our vocabulary: most linguists now prefer “logogram” or some more easily pronounceable regional term such as kanji  漢字 .25 Indeed there is room for clarity in the discussion of Chinese writing, especially among scholars of literature. I shall attempt to distinguish among classes often taken as equivalent. By pictogram is meant a conventional, usually highly schematized, graphic element the interpretation of which is based on a claim of visual resemblance between the graph and its referent. Logograms represent the words, or—to acknowledge the vagueness of the term “word”—the distinct meaning-bearing units of a language through a corresponding vocabulary of distinct signs. (The scholarly consensus at present is to consider Chinese characters a mainly logographic system: De Francis 1984.) In English we spell, or at least we think we do, by matching up alphabetic signs with sounds, but it has long been observed that in reading we usually grasp words and simple sentences as unanalyzed wholes, that is, logogrammatically. One test for logogrammatic qualities might be the persistence of a core of reference despite a wide variety of regional or functional pronunciations (as Chinese characters are not said to change their meaning in being pronounced differently by speakers of Mandarin, Cantonese, etc.; or, for the second case, as a single Chinese character used in Japanese may have widely different readings according to whether it is used as part of a noun, a verb, an idiom, a citation in literary language, etc.). Here too, however, allowance might be made for the role of writing in preserving the continuities of etymology through differences of dialect and period even among European languages: to adopt a panchronic point of view and say that Latin Caesar [kai:sar], Kaiser, tsar, czar, César, English Caesar [see:zer], and so forth are historically the “same” word is indeed to use an etymological root as a kind of translinguistic logogram. Part of the ignominy of ideogram derives from the way it is used loosely as a synonym for both the visual and semantic types of reference, tending to suggest that all true ideograms or logograms are really pictograms—Fenollosa’s thesis, more or less. An antidote would be to proclaim mathematical or logical notation the model for ideography. No one would say that p and q resemble propositions, particularly since we can form reasonings about p and q with no definite propositions p and q in mind. And yet p and q do for logical notation exactly what an ideogram is supposed to do: they mark the places of ideas. As a reference-point for the history of the notion of ideogrammatic writing, I note that the Encyclopédie of Diderot and d’Alembert describes Chinese writing as a hieroglyphic writing that has lost consciousness of (or concern for) its pictographic roots and grown accustomed to treating its characters as conventional “marks” for ideas. “L’écriture chinoise a fait un pas de plus [sur l’écriture égyptienne]: elle a rejetté les images, & n’a conservé que les marques abrégées, qu’elle a multiplié jusqu’à un nombre prodigieux.”26 So ideography is not necessarily pictography: the eighteenth century, at least, was able to see the distinctions. Another case would be the ideographic alphabetic languages of such seventeenth-century projectors as Lodwick, Beck, Dalgarno, Kircher, and Wilkins.27 Those languages assigned to each letter of the alphabet a place on a table of categories: a word, once analyzed into its letter-components, was a set of directions for locating its own meaning. But examples of such idea-writing without pictures may be found closer to hand: punctuation, capitalization, distinctive spellings, quotation marks and so forth modify the import of words and sentences in ways that may have no grammatical or prosodic equivalents in speech.

Ideograms on this pattern are distinct from pictograms: resemblance is not their only enabling medium, nor does an ideogram have to be immediately intelligible to earn its name. Fenollosa’s investment in the “picture” (and in the sentence as picture) is thus responsible for a deeper misunderstanding about the nature of Chinese writing than those to which China scholars often object. It is a matter of the nature of picto-ideograms, not their frequency or etymological importance. The fact that a Chinese graph does not seem to stand for a sound, but for a “thing” or meaning, allows commentators like Fenollosa to believe that it directly represents that thing instead of representing some intermediate object, say a concept, which in turn would take on reality for linguistics only once it had been expressed in vocal or written form. To many linguists the theory of the ideogram chiefly betrays a will to ignore the stages of abstraction, idealization and conventionalization that extract distinct things, words, and concepts from the ruck of sensibilia. Ideogrammatism conveniently obliterates history, returning us to the intellectual purity of the caveman’s awe at the lightning-bolt.

The trouble, however, is that one must become a primitive in order fully to appreciate the pictographic quality of Chinese writing. I find that price a little high, and propose a reformulation that restores to consideration all the things Fenollosa, with his exclusive devotion to lively pictures and juices, had to ignore. If there is a place for the notion of “ideogram”—the writing that posits and makes available for subsequent reference an idea—that place should be everywhere in language, for phonemes, letters, and words too are ideal and only conceivable as idealities.28 Yet “ideas” are knowable not as such, but through their having taken physical (verbal, graphic) form. The bandwidth assigned to the category logogram swells to fill most of the available definitional space, as our linguistic skepticism becomes less hospitable to wordless concepts and self-defining phonemes. Perhaps a braver grammatologist will someday propose a definition of “ideography” that includes alphabetic writing as one of its subsets (and a poor cousin no doubt). That move, however, would do away with the everyday category of “ideogram” in its opposition to ‘phonetic writing.’ All types of writing, and indeed all linguistic sound-forms, will be graphs, or enduringly inscribed forms of ideas. “Marks,” as the Encyclopédie called them. Would we have lost anything valuable?

5. “The images of cogitations”—with and without likeness

Thus the term “ideogram,” which may at one time have seemed a useful typological classification, now becomes a thing needing an explanation. The question is no longer: what is an ideogram? but rather, how did the term “ideogram” come into currency? What was required for the distinction—now part of our linguistic common sense—between phonetic writing and conceptual writing to be made? On what reservoirs of linguistic thought does it draw?

The term “ideogram” enters modern European languages upon the diffusion of Champollion’s decoding of ancient Egyptian script, announced in the Lettre à Monsieur Dacier of 1822 and presented more fully in the Grammaire égyptienne of 1836-41.29 Champollion distributes Egyptian characters into three classes: caractères mimiques, which depict or imitate the objects to which they refer; caractères tropiques ou symboliques, which depict an object having a meaningful relation to the thing or idea to which the character refers; and caractères phonétiques, used merely to represent or cue for sounds. The first two classes Champollion inherited from the ancient tradition of speculation on Egyptian writing. Ever since the classical period of Greece it had been customary to treat the Egyptian writing-system as through-and-through pictographic or symbolic. To decipher an inscription was to unfold the meanings of the pictures it contained: thus an eye was held to represent the knowledge of God, the snake biting its tail represented the universe and so forth. The difference between mimetic and symbolic characters was the same as that between literal and metaphorical speech: one represented immediately, the other represented something through which something else was represented. But both were interpreted without stepping outside the space of meanings: to take an extreme case, ‘Horapollo’ glossed Egyptian writing for Greeks and Romans while scarcely mentioning the Egyptian language. By introducing phonetic characters, that is, characters whose meanings are not identical to their functions in writing, into the decipherment of Egyptian, Champollion disrupted this purely semantic interpretative method. And that is the point at which the ideogram becomes conceivable in all its specificity; the point, also, at which Chinese writing begins to be described as an ideographic system. (The missionaries of the seventeenth and eighteenth centuries referred to Chinese characters as “letters”—an imprecise but at least not wrongly precise usage.)

Champollion knew a fair bit about Chinese, thanks to missionary investigations and scholarly debate; Matteo Ricci knew of Egyptian only through the rebus and picture tradition, so flourishing in the Rome of his day (see Iversen 1993). The document that contributed most to the European understanding of Chinese writing in the years before Champollion assumes that writing is inescapably phonetic. The short treatise De interpretatione attributed to Aristotle sets up a double parallelism between words and things centering, asymmetrically, in the soul: “spoken sounds are symbols of affections in the soul, and written marks symbols of spoken sounds. And just as written marks are not the same for all men, neither are spoken sounds. But what these are in the first place signs [σημει̑α] of—affections of the soul—are the same for all; and what these affections are likenesses [Όμοιωώματα] of—actual things—are also the same” (De interpretatione 16 a 2 - 3; Aristotle 1984.25). Francis Bacon is only the first to draw what will become an obligatory analogy.

For the organ of tradition, it is either Speech or Writing: and Aristotle saith well, Words are the images of cogitations, and letters are the images of words; but yet it is not of necessity that cogitations be expressed by the medium of words. . . . And we understand further that it is the use of China and the kingdoms of the high Levant to write in Characters Real, which express neither letters nor words in gross, but Things or Notions; insomuch as countries and provinces, which understand not one another’s language, can nevertheless read one another’s writings, because the characters are accepted more generally than the languages do extend. . . . (Bacon 1870.3:399-400).

Chinese writing brought the possibility of eliminating one of the levels of mediation through which the De Interpretatione had constructed its picture of mind, language and world. If indeed words symbolized (conventionally) affections in the soul and phonetic writing symbolized words, then a writing that symbolized affections in the soul would thereby symbolize things themselves, since both things and affections were “the same for all.” Such a writing would no longer refer indirectly (by a sign, σημει̑ον) to “things” belonging to an altogether different order of representation, but would be linked to a mental image that in turn linked up directly (as a likeness, ὁμοίωμα) with things. (For further evidence that Aristotle uses “likenesses” to indicate a more powerful form of representation than “signs,” see Politics 1340 a 33.) Given the possibility of making that direct link, it did not so much matter how it was achieved. Bacon divided such “Notes of Cogitations” into “two sorts: the one when the note hath some similitude or congruity with the notion; the other ad placitum, having force only by contract or acceptation,” and put Chinese characters in the latter category—as indeed did most European writers on China before Fenollosa. For Bacon, at least, ideogrammatism does not imply resemblance.

Those who in the seventeenth century could not conceive of a universal conceptual writing without grounding it in “similitude or congruity” read a different canon of books. Etymology, says Claude Duret in 1613, depends entirely “on the resolution of words into syllables and letters, wherein consists the representation of the essence of the things that one wishes to name by their correct names”; his model for etymology is a roughly-understood Qabbala.30 By searching out the word-origins of a modern language one can get back, not only to the full utterance of its original speakers, but to a proto-language forced on men by nature and thus common to all nations. Etymology (like Fenollosa’s “character”) is the model of a pre- or counter-Babelic language.31 Agrippa of Nettesheim cites “the Platonists” (meaning the Neoplatonists) as his authority for the view that “when a noun or a verb is formed the force of the thing [to which it refers] dwells in its articulations, like a kind of life hiding beneath the things it formally signifies.”32 The ordinary speech we use is a degenerated form of “sacred or divine letters, self-identical and permanent in every nation and language.”33 References to such letters in Renaissance texts almost always lead back to the Corpus Hermeticum, a set of late-classical initiatory writings in Greek attributed to the Egyptian god Thoth or “Hermes Trismegistus.” It was generally assumed that Egyptian hieroglyphics were the prototype of these “sacred or divine letters,” vehicles of mysterious wisdom. Thoth predicts in one of the treatises assigned to him, entitled “Definitions,” that his books are doomed to become “entirely unclear when the Greeks eventually desire to translate our language to their own, and thus produce in writing the greatest distortion and unclarity. . . . The very quality of the speech and the [sound] of Egyptian words,” on the other hand, “have in themselves the energy of the objects they speak of.”34 In Egyptian, then, one has the true ideograms of Hermes’ thought; translate them into Greek and they become impoverished markers, “sapless” Fenollosa/Pound would call them, of their original fullness. Note, however, the obtuseness of the Greeks even in their quest for exotica. The thing that strikes us most about Egyptian writing, its hieroglyphic or pictorial quality, is missing from the Hermetic description of its uncanny “powers.” Power resides for ‘Hermes’ in spoken words, not written marks.

The Hermetic praise of hieroglyphic evidently made its mark on the Jesuit proto-scientist and seeker into strange things, Athanasius Kircher, who wrote an early and oft-cited description of the Chinese writing system. Kircher traced Chinese writing back to the originals of the Egyptian Thoth or Hermes himself, known in China — according to Kircher’s typological chronology—as Fu Xi 伏犧. Fu Xi bequeathed to his Chinese descendants a “Serpent Book” in which the letters were formed by “serpents wonderfully intertwined, and made to take different forms representing the diversity of things signified by them.”35 But—precisely as in the passage from the “Definitions”—their removal from Egypt impoverished these characters. The Egyptian sign for the sun, for example, represented “the hidden powers exerted, not merely by the material Sun in this world of sensation, but also those of the archetype of the Sun in the world of ideas. . . . All of this is missing from the construction of Chinese characters, inasmuch as the latter were instituted precisely in order to indicate the simple concepts of words and names, with no other mystery latent in them.”36 Chinese writing for Kircher imitates nature without revealing her secrets. Of course, Kircher knew of Chinese characters only what a dictionary told him, and was utterly misled as to the meaning of Egyptian hieroglyphs.37 

Nature and convention mark the difference between these two ways of conceiving of the universal idiom, of which the Chinese character was, for early-modern Europe, one of the historical attempted realizations. (Duret’s Thrésor is one of the first European books to include a page of Japanese writing.) While Bacon would have been content to see an agreed-upon code serve as the “Characters Real” for a new ideogrammatic language, this second group of writers will be satisfied with nothing less than a participation (via ‘force’) of the symbol in the thing symbolized. Mere mimesis is too weak a word for the relation they wish to establish between sign and referent. Relevant to the distinction between these two visions of writing (because historically indebted to them) is the Romantics’ critical commonplace of a distinction between sign and symbol, or between allegory and symbol. “An allegory simply signifies a general concept or an idea which is distinct from it, while the symbol is the idea itself in sensory, bodily form” (Creuzer 1819.26). Fenollosa relocates this commonplace, moving it from the domain of theology or art history to that of geography: the languages of Europe are cold and thin, the European alphabets inexpressive, medieval logic and that inveterate tribal totem, the verb “to be,” dead weights on the mind of Europe.38 The Chinese character would serve as a “medium,” not only for poetry but for infusing new life into the dry veins of the West.

The examples from seventeenth-century Europe help to make more obvious something that might reasonably have been asserted about the twentieth-century interpretations of Liu Xie’s sixth-century utterances about wen: namely, that “nature” is a variable, rather than a constant. It is hard to imagine Bacon and Duret agreeing on the definition of “representation” or “the correct names of things,” and the reason is that they inhabit differently constituted cosmoses. Yet the basic semantic attitude of Fenollosa—his longing for a lightning-bolt of natural meaning—is that of the neo-Neoplatonics; his dissatisfaction with humanly patched-together idioms is theirs too. Nor should their common imagery of currents, flows, bonds, and magnetic force be treated as merely incidental. But can this European imagination of a Nature endowed with supernatural powers guide the idea of writing in at all the same way as Liu Xie’s cosmos realized itself in written patterns?

6. Syntax versus semantics, or the alphabetic ideal

The visions of natural force that inspire Fenollosa have (or such is my intermediary conclusion) a long European history and an uneasy resemblance with those expressed by Liu Xie; the history tending to cancel out the resemblances, just as the resemblances pressed upon us the history-less repetitions of Nature. To return to the tracing of the European imagination about language, the arguments made for the superiority of ideogram-like characters often invoke a pictorial, even a talismanic logic; and symmetrically enough, those whose linguistic ideal resembles the alphabet have a common bent, a pragmatic, technicist vocabulary and a concentration on certain operations of the mind. If the perfect writing in Fenollosa’s view is one that devotes all its resources to the acting-out of meaning, the anti-Fenollosa sees writing as a kit of parts and rules.

Ideographic-logographic writing and alphabetic writing are usually contrasted in terms of their structure, with the central exhibit being the distinction (found in any articulated language) between semantics and syntax, a distinction in which alphabetic writing systems have a certain constitutive stake. Chinese characters are vehicles of meaning; alphabetical characters, meaningless in themselves, combine to form words which are vehicles of meaning. A frequent analogy in philosophical linguistics holds that syntax, or the capacity of building up new sequences from preordained elements and rules, is the hallmark of life (as opposed to the dead letter of an inscription, in Plato’s Phaedrus [264 b 3 - e 1, 275 d 6]), of the human (as opposed to the animal, in Descartes’ Discourse on Method)39 or of free will, as opposed to biological or political determinism (as in Chomsky 1966). Alphabetic writing carries the syntactic principle from the level of the sentence to that of the individual word.40 But going off in the other direction, the theory of the ideogram carries semantic power to the utmost—the ideogram-as-pictogram is supposed to make visible the very concept to which it refers, without the mediation of sound—and so, symmetrically enough, reduces to a redundancy the grammar of the language in which it occurs. Ideogrammatic theories of Chinese writing thus correlate with the idea that the Chinese language has “no grammar.” (On such views, see Ramsey 1989.49-57.) Whether the ideogram’s power is supposed to compensate for the alleged lack of grammar, or whether the theory of the ideogram depends on a will to ignore what grammar there is in Chinese, is a question deserving special attention elsewhere.

A picto-ideogrammatic language is (or should we say, would be) a rather monstrous case of specialized overdevelopment—a motif running through countless descriptions of the time and tedium required of children who must learn thousands of distinct characters, and the ground for Hegel’s famous pronouncement that “alphabetic writing is in and for itself the more intelligent.”41 Given the longstanding association between syntactic arrangement and intelligence, and the labor-saving ideal of alphabetic spelling systems, it is hard to see how Hegel could have thought otherwise. Cave Beck, the author of an alphabetic yet ideographic universal language, neatly combines all the faults of pictographic writing in his vigorous critique: so-called ideography fails to represent objects with intuitive accuracy, it too often depends on devices of the local languages it is supposed to supersede (puns, metaphors, word-associations), and it requires heroic efforts of memory:

The Egyptians of old, had a Symbolical way of writing by Emblems and Pictures which might be read by other Nations instructed in their Wisdom, but was so hard to learn, and tedious in the practice, that Letters soon jostled them out of the world. Besides, most of their Hieroglyphicks were so catachrestical (the Picture showing one thing to the eye, and a quite different sense imposed upon it) that they justified the Printer who drew a misshapen Cock upon a Sign-board, and wrote under it, ‘This is a bull.’

That the Chinois have a general Character, which serves themselves and their Neighbors, though of different languages, is affirmed by some that have been there, and brought some scrawls of it, which are such for their fashion, that an European with his one Eye (which they afford him) would think they shut both theirs (they so much boast of) when they drew the shapes of those Characters: there being no Proportion or Method observed in their form, which causes them to spend many years, beginning in their childhood, in learning of it, and this may be the Reason none of our travellers have been able to bring away any competent skill in that way of writing. (Beck 1657: “To the Reader,” unpaginated)

The very existence of ideograms in the strong sense (Fenollosa’s sense) of the word is thus challenged by illiteracy, by the fact that people can fail to make sense of the system. That should not happen when a writing-system or artform delivers “the idea itself in sensory, bodily form,” but of course it does. You say niao 鳥  (or even its archaic form ) is a picture of a bird, I say it is a conventional mark that readers are trained to read as meaning the word ‘bird.’ There is something to be said for both accounts, but the first case is harder to make, since it requires agreement on a number of linguistic and vaguely experiential issues (what would constitute a picture of a bird? what would not count as a picture of a bird? what defines ‘being a picture of something’? is the seeing of marks as pictures not a matter of conventional training?—and so on). What would permit us to verify that a representation is a true “likeness” of its content and “the same for all” (whether conventionally or naturally) is a determinate account of the relation between linguistic meaning and reference; and such an account is less to be hoped for, the more reference to extra-linguistic objects is involved. (The situation in mathematics is less desperate because mathematics does not require the existence of its objects.) In any case, the condition of possibility (or impossibility) of ideography is irreducibly semantic.

7. Syntax within Semantics

The problem with the theory of the ideogram and its correlates in literary study is not that they mislead or involve dubious presuppositions (both forgivable flaws, since literary understanding is not a straight deduction from axioms), but rather that they direct our attention to certain objects and make other objects hard to conceive. Indeed, as long as the discussion about wen remains within the province of reference and semantics, it tends to founder on problems of representation—claims that representation can be direct, faithful, adequate, sincere, and so on, versus claims that it can’t. I will therefore try to shift it to new ground: rather than conceive of wen as a writing that differs from alphabetic writing or Aristotelian representation through having a more immediate relation to things or ideas, I shall try to capture its specificity in its expression of part-to-whole relations. That is, the syntax of the sign—not what it means but how it is put together. (Traditional etymological studies, beginning with Xu Shen’s six classes, are not incompatible with this model: note how much less arbitrary it is to rank characters in terms of the complexity of their part-to-whole relations than it is to sort them subjectively on a scale of pictographic qualities.) This involves putting the shoe on a wrong or at least inhabitual foot, since accounts of wen that emphasize semantic power have tended to ignore syntax or to reduce it (as did Fenollosa) to another means of semantic representation.42 Fenollosan signification fits into this scheme as a claim that there are representations that are simply caused, and not at all, or very minimally, put together; and a counter-argument will have only to show how much (and exactly what) every reading of an object as simply caused has to leave out.

What would it mean to read ‘ideograms’ (and their poetic counterparts) not ideographically but syntactically? Would that not compel us to give up the power and majesty of cosmic writing, to abandon the specifically Chinese patterns of thought about writing and meaning? Liu Xie’s praise of wen has often been interpreted in a Fenollosan sense, that is, as proclaiming the unity of Chinese writing with the universe it emulates. My counter-citation (and route of return to Liu Xie) will be the definition and etymology of wen found in the Shuo wen jie zi—a fairly inconspicuous text, but one into which a reading of the very classical texts cited by Liu Xie is compressed. Xu Shen’s original entry for the character reads merely: Wen, cuo hua ye 文，錯畫也: “Wen is crossing lines.”43 Crossing lines? That sounds strangely weak and vague. The great eighteenth-century philologist Duan Yucai adds a commentary drawing on examples from the classical-period texts that Xu would have consulted in making his dictionary:

Cuo 錯 , ‘cross,’ here stands for [its homonym] cuo 逪, ‘meet.’ ‘Meeting lines’ are lines that cross and join. When the Kaogongji 考工記 [Record of Artificers] section of the Zhou li [Rituals of Zhou] says, ‘Green set beside red: this we call wen,’ it gives a case of meeting lines, and such meeting lines are the etymon of wen.

錯當作逪‧逪畫者，交 逪之畫也‧考工記曰‧ 青與赤謂之文‧逪畫之 一耑也‧逪畫者，文之 本義‧

Lines cross or meet most strikingly in the juxtaposition of designs and colors in an artifact. And not just any artifact either: the passage Duan cites gives instructions for the ornamentation of the sovereign’s clothes, an inevitably meaningful and meaning-conferring kind of artifact. On the evidence of this passage, writing is like a mark: a mark of distinction, an elaboration, an addition to the plainness of whatever is without wen.

But then Duan goes on to give a second example from an even more prestigious text—so prestigious that it does not even need to be named, for it is the Book of Changes.

The Yellow Emperor’s scribe, Cang Jie 倉頡 , perceiving the tracks left by the claws of birds and animals, understood that their distinct patterns [li 理 ] could be used as the basis for differentiating them into classes. So he began to make tallies and writings, producing images and shapes according to the kinds [of the things represented] 依類象形 . Thus the name of writing is wen.44

This second example shows us writing as reading, interpretation, tracking of causes, even (incidentally) imitation of something given in the natural world: as a track of a certain shape signifies that animal and no other, so a certain sign represents and singles out a certain meaning. Not composition but causality provides the rationale whereby the ‘writing’ of animal spoors transmits information. Reading, in this version, is none other than perception (and note the similarity between this model of writing and Aristotle’s account of perceptual ‘impressions’).

Taken as semiotic fable, the second example corresponds antithetically to the first: as nature pairs off with artifice, and causation with composition, so the inherently meaningful sign stands in relation to decorative outwardness. Now if the Yi jing passage gives us the original scene of semantics, the Zhou li passage gives us that of syntax (the embroiderers’ wen is made by putting “green together with red,” but the meaning of this wen is to be sought elsewhere). A fuller citation from the Kaogongji will complete the story of how wen comes to inhabit the artifact. “The embroiderers’ task is to combine the Five Colors,” says the text. “The color of the East is green; that of the South is red,” and so on through the four directions plus heaven and earth. “Green and white follow one another or go on top of one another [ci  次]; red and black follow one another; dark and yellow follow one another.” “Following one another” or “overlaying one another” is the name for the relation of polar opposites: north and south, east and west, heaven and earth. These “overlaying” pairs seem to be background colors suitable for robes to be worn in a given season, as the Son of Heaven takes up residence in a certain quarter of his Mingtang (Hall of Lights). But combinations of “neighboring” colors are more unusual, more specialized: each gets its own name and is associated with a definite pictorial shape. “The combination of green and red is called wen. That of red and white is called zhang  章[boundary]. That of white and black is called fu 黼 [axe-head shape]; that of black and green is called fu 黻  [double-bow shape].”45 The Kaogongji gives a “grammar of ornament” (if I may reuse with extra emphasis a fine nineteenth-century title)—a set of rules and forms appropriate to the adornment of a Zhou king. The colors may be meaningful, whether naturally or by convention; so the shapes embroidered on the robe certainly are; but it is the syntax, or combination, of the meanings of the different background colors and the different shapes chosen for the foreground that gives the artifact as a whole its character as an “inscribed object” in the broadest sense, or wen wu 文物 .46 The Kaogongji account tells, unlike the “bird-track” story, of the generation of meanings and meaningful things out of meaningless things. When we are talking about nature as opposed to culture, clothes are a particularly loaded subject; and if one wanted to emphasize the doubleness of the signifier wen one would say that the imperial attire, if truly inhabited by wen, imprints kingliness on the human animal chosen to wear it, rather than the kingly qualities somehow exuding from man to robe. To reason from bird track to bird shows good woodcraft; to reason from the signs of authority to the basis of authority betrays unfamiliarity with the ways of the world. If the two kinds of know-how were identical, perhaps culture would finally become ideogrammatical—which is doubtless what Chinese writers have always wanted it to be, and who can blame them?

Thus the Chinese tradition encompasses both sides of what, in most Western writing on the subject, is presented as an alternative of Chinese and Western conceptions. (Western accounts of the philosophical background of Liu Xie’s description of wen tend to present one side only of the Chinese evidence, the side of the “natural image” or “ideogram,” and moreover to stress that side’s essential Chineseness: see for example Jullien 1985.24-28.) Our closer look at the range of activities whose accomplishment is named wen helps to situate the kind of wen of which Liu Xie spoke: as a cultural ideal, a wish grounded in belief and as such endowed with a different sort of power than that which an iconic reading of his description would give it. It also provides us the outline of a contrapuntal reading of Liu’s praise of wen, a reading that would see it above all as a composition. (A shift towards the latter style of reading is one of the great virtues of Owen’s 1992 discussion over that of 1985.) With this way of reading, Liu Xie’s page begins to swarm with invisible quotation-marks. Wen displays the evidence of itself in the world for the semantic eye to see and read, as difference, analogy, pattern. But to reproduce the actions of wen in writing, Liu Xie must compose: for those whose ears are trained on the vocabulary of the music section of the Records of Ritual, the differences and analogies he notes as relevant are those crucial to the building of the Chinese social order and its artistic monuments (tian, di; gao, bei  天，地，高，卑 ), differences that must be understood as differences rather than as atomic positions (the ritual texts are quite firm on the idea that without one element of a polar pair, the other cannot be thought either).47 “Sounds that form a complete pattern (cheng wen 成文) are called music,” says the “Record of Music,” which further prescribes: perfect musical-dramatic performance is achieved when “the Five Colors [or just possibly: the five elements with their corresponding sounds] form a pattern (cheng wen) and make no havoc, the Eight Timbres follow their modal rules without trespassing” (Li ji, “Yue ji,” in Ruan Yuan 1987 [1815], 37/4a; 38/10b-11a). The pattern of wen may result from nature, but emulating it takes effort. A reading that aimed at preserving both the content of Liu’s writing and its construction could take it as an act of translation—the persistent relay of natural predicates into cultural and textual ones, of the differences that the hands and eyes perceive into the differences that the educated mind sees. Liu allows a place for that translation to occur, namely the mind, that focal creation that reconstitutes the order of nature in perceiving it. If that is Liu’s course, it is quite opposite to that set by Fenollosa, even if the two paths may go by some of the same landmarks.

And getting the direction of travel—the order of operations, the implied syntax of composition—right makes all the difference. If ideology (an ugly term, but one sure to appear wherever literary criticism is done nowadays) denotes the practice of disguising facts of history as facts of nature, then a piece of writing like Liu Xie’s presents pitfalls. Read it as an image of the semantic universe of its writer, and it is purest Chinese imperial ideology; read it as an act of cheng wen —the syntactic alternative I have been developing here—and it becomes a manual for constructing imperial ideology. In the embroiderer’s eye the cloud-rending dragon dissolves into a field of tiny, separate stitches. That may be the price of finding out where dragons come from. But that is also how reading becomes, not merely the reading of what is written, but the reading of the writing-process and the whole cumulative series of writing-processes invoked by every instance of writing. That is why a semantic account of the Chinese character is incomplete (and never so incomplete as in the pure theory of the ideo-pictogram); that is also why the theory of Chinese writing as the immediate perception of cosmic order, while gratifying to contemplate, leaves us holding a thin strand of text.

8. Rhythm, or Sequence without Semantics

An uneasy awareness that sense is made in society, and that the social order most densely packed with meanings is not necessarily the best, surfaces in recent half-disavowals of the theory that the Chinese cultural order is a natural and inevitable reading from experience.48 Are these the alternatives—the celebration of a social order’s belief in itself as the one natural order, and the suspicion that any attempt to ground social order in nature is only the wish-fulfillment dream of the powerful? Let us suppose they are. How will we who are bound to Chinese literature answer those alternatives through the techniques of reading?

The purpose of this essay (up to this point) has been to develop the artificer’s point of view and draw the reader’s sympathies over toward it. A rhetoric whose job is to ground the categories of art in those of nature is by the terms of its own argument committed to not recognizing its handiwork in some of its objects. There has to be a nature with which art can productively contrast, or the grounding will have no ground to stand on. But no such rule of limited responsibility (or limit to self-knowledge) applies to a rhetoric which sees itself as splicing art onto art, (perceived) order onto (presumed) order. To say that the Chinese (or the ancient Chinese) believe, or want to believe, in dragons does not really put the difference between dragons and embroidery to rest. After all, we owe what we know about dragons to those who could both embroider them and tell us about the process—for example, Liu Xie.

But adopting the embroiderer’s point of view—or adopting a social-constructivist platform for interpreting the networks of meaning in Chinese literature—is not necessarily the end of alienation and an instant refutation of all authority. Liu Xie can show us as much, for in China perhaps more than anywhere else, a vigorous social mythology has grown up around the idea that writing (or kingship) is a matter of tempering or weaving together the stuffs that brute nature provides.49 By opting for a poetics of manufacture, one has not automatically put imperial rule and its charters behind one. One may indeed have done no more than express one’s need for a master of works.

Our time is most comfortable with the social-constructivist option, even to the point of admitting the natural semiosis of the Chinese literary universe only as an example of a world constructed by a different set of social imperatives. But before we congratulate ourselves too heartily for being on the side of progress and civilization (postmodern civilization at that!), I should like to point out that the social-constructivist view in its specifically post-Enlightenment form is nothing new in Chinese studies either. Pride in our own contemporaneity comes at the price of obliterating the discipline’s recent past. Who could provide a more subtly historical and social reading of Liu Xie’s praise of heavenly wen than Marcel Granet, the pupil of Durkheim and teacher of Lévi-Strauss? Granet, it is true, has often been taken to task for having presented a synoptic, ahistorical picture of Chinese culture.50 But when we read Granet as an artificer attempting to reproduce an earlier process of artifact-making, perhaps the tenseless, categorical time of his history-telling will fall into place as a made time, a historical “representation” that yet shapes historical experience. Here is one example to show the way—Granet on the genesis of yin-yang 陰陽 and wuxing 五行 cosmologies:

The representation of Time and Space [in early China]. . . derives, not from simple individual sensations or from the observation of nature, but from purely social activities. Its elements are taken from the image of two teams clashing, under particularly moving circumstances, in a ritual joust. . . . A duple rhythm, based on simple opposition and simple alternance, presided over the organization of society. And thus it dictated the twin representations of Time and Space. . . . This rhythmic constitution [of time and space], based on the principle of antithesis between periods of dispersion and concentration, was first expressed through the correlated ideas of simple opposition and alternation, and the representation of Time and Space was from the first linked to the feeling of a difference of value between two sorts of distance and duration. . . . The idea that distances and durations were not all of equal value led to the idea that intervals of time, like intervals of space, possessed differing natures. This step forward was accomplished once the leading representation of Space was no longer the spectacle of two groups of warrior bands standing face to face, but rather a square formation in which the axial line separating the two groups had transformed itself into a center occupied by a Chief. . . . (Granet 1994.93-99)

Granet’s reconstruction of the history of Chinese time (an ambitious project, and we should give it the full measure of its ambitions) puts social organization in the position of an agency determining the representation of nature. This much a quick and opportunistic reading of Granet’s historical fable will give us. But a closer look finds rhythm—syntax in a pure form, a syntax prior to all content—“presiding over” the social reality.51 “A duple rhythm, based on simple opposition and simple alternance, presided over the organization of society [and] dictated the twin representations of Time and Space. . . .” Granet’s seeming ahistoricality represents an attempt to derive history from something that is not yet history. “Alternance” and “opposition” are the logical operators whereby the linguistics of Granet’s time tried to account for the cenematic order, the order of the constitution of sign-forms (cenemes) prior to their acquisition of meanings (see Saussure 1974.164-165, Hjelmslev 1985). Now there are certainly people who will contend that even alternance and opposition are culture-bound entities, and that by invoking them as his operators Granet has not in the least penetrated to the backstage workshops of meaning. But that objection does not affect the structural suggestiveness of Granet’s story, which now appears as a three-layered one: the story of the generation of a natural order out of a social order, which in turn is a specification, a semantic investment, of an order of merely potential signifiers. In Granet’s story of the constitution of imperial order, the moment of alternating opposites gives way to a period in which the axis of difference, that line without area, “transforms itself” into a space that a chief can inhabit. Liu Xie put writing into that spatialized center. To continue the passage we began by citing at the outset of this article:

Now Azure and Yellow intermingle their colors,
the square and the round differentiate their forms;
sun and moon in their recurrent orbs display the symmetrical images of heaven,
mountain and river in brilliant patternwork spread forth the ordered shapes of earth;
—all this the wen of dao.

夫玄黃色  雜，方圓 體分；
日月疊壁，以垂麗天 之象；
山川煥綺，以鋪理地 之形；
此蓋道之 文也‧

Syntactic reading can find the wen that interests it in similar processes—orderings and intervals that are not to be found in the red or the green, but only in what happens when they are put together. Not to see the sun and moon any more, but only the distinction between them: this sounds like a hard discipline for readers of Chinese poetry, but it is worth straining a bit to be a participant in the making of wen.

This essay was originally published in February 1997 as issue no. 75 of Sino-Platonic Papers and is licensed under the Creative Commons Attribution NonCommercial NoDerivs 2.5 License.

This issue is also available in its original appearance as a 3.2 MB PDF file.