Journal of Chinese Literature and Culture

Suppose I. A. Richards, famously known for his experiment of teaching poetry "by isolating the text from history and context,"¹ gives his students the following poem:

The phoenix cover and the lovebird curtain are nearbyWhere I would go if I could get there.Shrimp whisker brushes the floor and double doors are still.I recognize the shuffle of embroidered slippersInvisible in the bedroom,Her forced laugh, her voiceLight and lovely, like a woodwind.

Her makeup done,She idly holds a lute.Love songs she likes to relish. [End Page 47] Into every note she seems to put her fragrant heart.Listening outside the curtainGets me so much heartbreak!Misery such as thisOnly she could share.²

What kind of interpretation would Richards expect from his students?

The first thing students would try to do is to "mak[e] out the plain sense of the poetry," as this is the "chief difficult[y]" their teacher told them to tackle when reading a poem.³ Students will find without difficulty that the poem is about a male persona overhearing a female character's singing. The situation, however, is not so plain. Even though the singer is said to "relish" the love songs in a "light and lovely" voice, and even though her intermittent laughter can be heard, there is something wrong. The word used to describe her laughter—"forced"—arouses students' suspicion about the sincerity of the words she is singing. The genuineness of her tone is instantly called into question.

This should alert students to an instruction from their teacher. "The tone of [a character's] utterance," says Richards, "reflects … his sense of how he stands toward those he is addressing."⁴ For the singer to have a clear target toward whom she could reflect her unwillingness to sing, there has to be an unwanted listener in her presence. This pitiable listener cannot be the persona, since he is "listening outside the curtain." Obviously, a third character, not mentioned in the poem, sits squarely or reclines languidly on the other side of the "lovebird curtain."

And three's a crowd. The situation becomes interesting. While the invisible listener—whom the singer is "forced" to entertain—will take the singer's melodious words as genuine, the eavesdropper persona is not so sure about this. Since the singer's "fragrant heart" only seems to be true, it can be untrue. Excited by their new finding, Richards's students might be able to gather enough internal evidence (from the meticulously detailed description of the singer's boudoir) to conclude that, in all likelihood, the persona himself has been there before, on the other side of the curtain ("where I would go if I could get there"). Precisely for this reason, the ambiguity in the singer's voice both encourages and devastates him. The misery and desperation the persona expresses at the end carries more than a ring of truth. It seems to be genuine to his audience, the readers of this poem, a love song of his own.

Irony, paradox, ambiguity, and tension, crisscrossed, interloped, or overlapped with one another at different levels of the meaning of words—every ingredient for an intense close reading is there. For students from Richards's class, the mission of interpretation seems to be accomplished. [End Page 48]

For Stephen Owen, however, whose shrewd and close reading of the poem we went over above,⁵ the examination of the "internal evidence" found in the poetic text itself is far from enough. Reading the poem in its original Chinese certainly helps with a more intimate understanding of the formal features of the poem:

咫尺鳳衾鴛帳，欲去無因到。鰕鬚窣地重門悄。認繡履頻移，洞房杳杳。強笑語。逞如簧、再三輕巧。

梳妝早。琵琶閒抱。愛品相思調。聲聲似把芳心告。隔簾聽，贏得斷腸多少。恁煩惱。除非共伊知道。⁶

But it is the examination of the poem in a rich context, instead of focusing on the text as such, that enables Owen to make the best of close reading as a powerful interpretive tool. By reading the poem as an integral piece in a poetic tradition, he retraces the trajectory of Chinese literati's adoption, reformation, and eventual appropriation of a poetic form with a special folk origin and reveals the mechanism of the generic features of this poetic subgenre to which this poem belongs.

A fundamentalist close reader from the New Criticism camp, who insists in excluding any outside evidence from the interpretive probing, will not know that the poem under discussion is not just another poem but a song lyric, or ci 詞, originally meant to be sung by an entertainer at banquets or pleasure quarters. Its title, "Listening outside the Curtain," is the title of a song tune, the musical prosody of which shapes the formal structure of the lyric. The author of the piece, Liu Yong 柳永 (ca. 987–ca. 1053), a member of the Northern Song literati well known for his morally dubious experiences on both sides of many "curtains," might just try to enliven an old tune title by creating a mini poetic drama. Like numerous versions of the same song before this one, Liu's "Listening outside the Curtain" had been, and would be, performed repeatedly in [End Page 49] many other occasions by different "fragrant hearts." The genuineness of the feeling it expresses is further complicated.

The love song is both the stylized imitation of love and at the same time the words in which a truth of love can be spoken. The singer is both a professional, paid to enact passion, and a human being, to whom love, longing, and loss can actually happen. We would be overly credulous to believe every statement of love-longing is indeed love; we would be foolishly cynical to believe that every statement of love-longing is purely professional or part of a hollow game. And we can't tell the difference.

(38)

Instantly, many questions and doubts arising from the preliminary reading of this song lyric can be answered. More important, Owen's close reading serves as a brief yet clear explanation of the ci genre's musical origin and the drive behind the development of the genre.

Owen does not stop here. Whereas he calls into serious question the genuineness in the singer's voice, on one hand, he pays even more attention, on the other hand, to the expressive potential he detects in the seemingly genuine tone in the eavesdropper's yearning at the end of the lyric. "As he persuasively dramatizes his own 'genuine' concern for the genuineness of the beloved's song words," observes Owen, "he drives the reading of song lyric toward being more like that of shih [shi 詩]" (45). Inevitably, Owen finds himself bringing shi into consideration in his search for "truth."

A time-honored poetic form for personal expression, shi "could make the assumption of genuineness. … In contrast, genuineness was a problem in the song lyric" (45–46). When one hears a singing girl performing "her" love song created by a male ci writer, one is listening to a formulaic duet "in the voice of others" (or daiyan 代言). This voice of others, however, was so "light and lovely" (zaisan qingqiao 再三輕巧). Its expressive melodious tone attracted the attention of the literati, who used to express their heart's intent, or zhi 志, in the comparatively more straightforward shi form. Unable to resist the artistic appeal of this novel poetic voice, they wanted to make it their own. To achieve that, they just needed to try every means to turn what was formulaic in the ci into something specific and concrete, to turn the unreliable reliable and the categorical particular. Seen in this light, what literary history describes as the evolution of the ci genre between the beginning of the Northern Song period and the time of Su Shi 蘇軾 (1037–1101) is "a transformation from a normative and typological song form to a highly circumstantial form" (45).

It is the clues hidden in Liu's ditty that leads Owen to his conclusion. However, had he not expanded the scope of his close reading, he would not have been able to elicit the generic formal features of the ci in its maturity, which allow [End Page 50] the literati poets "a genuineness of voice almost impossible in shi" (69). Unlike early practitioners of close reading, Owen does not take the text under scrutiny as a self-contained and self-referential enclosure. His suspicion about the genuineness of the poetic expression of the ci, aroused by "Listening outside the Curtain," urges him to reach out to other song lyrics by the same author, then to similar works in the same tradition by other poets, and further, to works in the genre forms other than that of ci for a comprehensive comparative study.

Of course, we cannot pretend that Owen, as a critic of Chinese poetry, does not know about the transformational history of the ci genre before he picks up this poem. What is really going on is that, besides being the learned narrator in this act of close reading, he also needs to play the role of an innocent reader. The most challenging job for this double role is to defamiliarize—not exactly in the Russian formalist sense of the word—what his learned narrator knows so as to provide an exciting unfamiliar world of words for his innocent reader to make new discoveries. Actually, even the learned narrator himself can be surprised by the insight yielded by the new look of the all too familiar texts under the exacting pressure of this defamiliarization. A good example of this is seen in Owen's discovery of the workings of poetic clichés in the song lyric. Like most discussions of this poetic form, known as the long and short lines (chang duan ju 長短句), Owen's probing of its generic features also involves an examination of the genre's most distinctive formal features: its asymmetries (59). Unlike other studies of this subject, however, Owen does not limit his attention merely on the expressive sound effects created by the asymmetric line formation. Instead, he tries to find out why a cluster of stylized poetic expressions, or clichés, once embedded in a set of irregular lines, gain refreshing vividness.

His finding is thought provoking. "This is a question of taxis ('arrangement,' the sequencing of words and periods)" (58). "Song lyric works with clichés, normative responses, and commonplace categories of feeling" (62) by arranging these highly stylized and hackneyed—yet recognized as "classical"—poetic gestures in "more discursive, often vernacular" syntactic units (58). When a poetic cliché is thus qualified by the vernacular context of the specific and particular, it gains new life. It could stand as an isolated phrase representing a unique mind state, could add to another utterance "as if an after-thought," or "formally enact a sudden shift, an odd association, a flashback, an image left hanging" (59), among other things. If poetic clichés are none other than poetic emotions stylized in categorical or normative words, then "the verbal embodiment of subjectivity was achieved not 'in' words but 'in between' words" (58).

In Owen's vocabulary, cliché is not a pejorative. In another place, he declares that "poetry will always try to speak the difficult truths of the heart, and to break free of the tribe's clichés that involuntarily rise to the lips to take the [End Page 51] place of everything that is hard to say. But a successful poetry recognizes that this process is a struggle, that such words do not come easily. As a culture acquires more history, credibly simple words seem more and more difficult to achieve."⁷ This is tantamount to saying that cliché is a necessity in Chinese poetry. Specifically, in the case of song lyric, "to speak the difficult truths of the heart" is to embed clichés in particular situations with detail and nuance. Suddenly, Owen's understanding of the interplay between cliché and the ci context in which it is embedded sounds like T. S. Eliot's well-known theory on the relationship between feeling (condensed in the poetic tradition) and emotion (issuing from personal experience).⁸ Yet since Owen is not bent on, as Eliot is, belittling the role of the individual talent in the preservation and development of poetic traditions, his assessment of the interactive relations between the old and the new seems more balanced and hence more helpful for a practical explanation of one of the important elements of the mechanism of the ci as a poetic subgenre distinctively different from the shi.

As an interpretive approach, close reading seems to be extremely natural and congenial to sinologists working in the field of classical poetry. When Hans Frankel used this method in 1964 to make new discoveries in Cao Zhi's 曹植 (192–232) poems, he had to carefully title his article "An Attempt at a New Approach."⁹ Today, close reading is still one of the most convenient and necessary tools for researchers and academicians in the field of Chinese poetry. One just needs to look at Owen's example to see to what extent close reading has advanced from an analytical tool, unconditionally committed to tenets of New Criticism (seen in Frankel's case), to a sophisticated exegetical device in the study of classical Chinese poetry. It empowered its practitioners to perform their duties, that is, to enlighten and delight generations of readers and students of Chinese literature and to dig out hidden meanings in scholarly pursuits.

Ironically, Owen's success also reveals a limitation inherent in the use of close reading as an interpretive tool. So much depends on the critical sensitivity of an ideal close-reader and on the salient readable features of an individual ideal text to be close-read. Imagine a reader, lacking Owen's caliber of critique, facing a song lyric that does not present a suggestive curtain between the persona and the singing girl. There must be a reason that close reading both makes New Criticism and also breaks, in a way, New Criticism. One might well wonder whether it is possible to develop this effective interpretive tool into a stable, reliable, always accountable, or even predictable critical approach that can be applied indiscriminately to the study of poetic texts in general (as opposed to selected texts only). To use an example close at hand, is there a way to examine the interplay between poetic clichés and its "vernacular" ci context—the observation by Owen mentioned above—in light of the theoretical investigation of the [End Page 52] coexistence of concreteness in sensory impression and the abstractness in reference in the Chinese poetic imagery?

In fact, this was precisely the kind of question posed by the pioneers in their early experimentation with close reading in the study of Chinese literature. Not long after Frankel's trial of this critical method on Cao Zhi, Yu-kung Kao and Tsu-lin Mei started their collaboration in the application of close reading to the study of Tang poetry. "By choice and by habit," the two young scholars declared with emphasis, their critical orientation was Empsonian "linguistic criticism." They also took note of Northrop Frye's Anatomy of Criticism and Roman Jakobson's 1958 call for linguistic approach to literature also ringed in their ears.¹⁰ No matter what -ism caught their attention, they never wavered in their determination to "indicate how specific linguistic features are multiply effective in a poem."¹¹ In other words, they took it as their mission to reveal the secret of the "underlying aesthetics" of Chinese poetry.¹² The vision is systematic, the scale all-encompassing, and the methodology linguistic, hence "scientific"—nothing short of an ambitious critical scheme informed with, among other things, the principles and ideals of close reading.

Specifically, what is the "underlying aesthetics" this grand scheme of linguistics-based criticism aims to elucidate? In a retrospective explanation Kao defines it as follows:

This aesthetics is basically an interpretative code, through which a poet can go beyond the textual meaning and the reader can understand its contextual significance. Through this code the poet and reader can communicate and exclude the uninitiated. This aesthetic code cannot be acquired as a mere set of rules, prescriptions, and proscriptions; it is learned only by internalizing models, with or without the assistance of explicit interpretation and prescription. Precisely because it always presents itself indirectly, it is difficult to articulate this aesthetics as a code, but the very fact that it never becomes fully explicit protects its power to suggest, to change, and to develop. The fact that I attempt to outline this implicit code in the following pages indicates that I do believe the code can be made explicit to a certain degree. Nevertheless, we should never forget the level on which this code always presents itself—submerged in and integrated with particular texts.¹³

The passage can be read as the manifesto of an "investigator," instead of a "critic," of the secret of Chinese poetics. Although the self-admonition at the end promises an emphasis on the close reading of "particular texts," the passage is much more than a simple pledge to the doctrines of New Criticism. By defining his coinage "aesthetics" (meidian 美典)¹⁴ as an "interpretive code," Kao echoes Jakobson. For a message to get across in any verbal act, says Jakobson, it requires "a CODE fully, or at least partially, common to the addresser and addressee (or in [End Page 53] other words, to the encoder and decoder of the message)."¹⁵ However, before Kao commits himself to Jakobson's belief that this code can function as a "metalanguage," or "a scientific tool" in the study of literature,¹⁶ he finds it imperative to acknowledge the complex and nearly "indescribable" nature of this code. Like Jakobson, Kao also believes that one should not confuse "literary studies" with "literary criticism": the former, in Jakobson's words, is "the description of the intrinsic values of a literary work,"¹⁷ while the latter is characteristic of "a subjective, censorious verdict" based on "a critic's own tastes and opinions on creative literature."¹⁸ Yet unlike Jakobson, he does not believe that embracing the former necessarily implies the rejection of the latter. Being reflective and introvert, the intuitive "knowing" of a literary critic is prone to be imagistic or even metaphorical, refusing to be pinpointed in analytical language. Nevertheless, this empirical knowing might have followed a logic or certain "objective" criteria of its own. There is a reason that most traditional commentators and not a few modern critics of Chinese poetry like to emphasize the "comprehendible yet ineffable" (zhike yihui buke yanchuan 只可意會不可言傳) property of the esoteric "aesthetics."¹⁹ After all, "the very fact that it never becomes fully explicit protects its power to suggest, to change, and to develop."²⁰ With this understanding, we will be in a better position to know what Kao means when he declares that he would make the implicit code explicit "to a certain degree" by "outlining" it. Outlining the implicit in analytical language does not need to be an oxymoron.

Kao's first serious effort in eliciting this interpretive code is seen in his collaboration with Mei Tsu-lin in a close reading of Tang dynasty regulated style poetry. The material for the study, Du Fu's "Autumn Meditations" series, is carefully chosen to fit their purpose. The verse form Du Fu uses is the most regulated and exacting form, a "poetry consciously written and meant to be read or heard with constant attention."²¹ In addition, thanks to Du Fu's obsession with the "striking effects" of poetic craftsmanship,²² his works offer themselves as ideal artifacts for "a New Critical 'reading,'" "an exercise in reverse engineering: the examination of an artifact to see how it was made and how it worked."²³

Possibly influenced by Jakobson's emphasis on the phonetic features in poetry,²⁴ Kao and Mei spend much effort discussing the "figure of sounds." The results are mixed. For example, it is quite convincing to show how Du Fu's design of overly reduplicative sound patterns "betrays a weariness" from facing beautiful scenery for too long.²⁵ However, based on the sound effects, the reading of the following example,

… A disdained K'uang Heng, as a critic of policy;As a promoter of learning, a Liu Hsiang who failed.

匡衡抗疏功名薄，劉向傳經心事違。 [End Page 54]

may be farfetched: "The punctuated repetition of kong-like sounds and the jostling of vela nasals convey some of the agitation" resulting from "the poet's failure in his moral and official career" (48). Then, in another example,

The clouds roll back, the pheasant-tail screens open  before the throne;Scales ringed by the sun on dragon robes! I have  seen His Majesty's face.

雲移雉尾開宮扇，日繞龍鱗識聖顏。

it is true that a careful modern reader, who has some knowledge of "alliteration," might notice that "there we have two rhyming syllables followed by two alliterative syllables in the leading line, and then, in the matching line, three alliterative pairs in a row." However, one might find it hard to believe that "the exuberant display of phonic patterns and especially in the cloying concentration in the third couplet is the kind of hubris portending decline [of the Tang empire]" (50). Here Kao and Mei might be overzealous in their effort to emulate Cleanth Brooks and Robert Penn Warren's meticulous analysis of the relation between sound and sense in Understanding Poetry. Since Chinese does not have consonant concatenation, to say the least, and the prosody of recent style poetry is highly regulated and therefore forbids much license in rhythm, the formulas provided by Brooks and Warren will not work well. To examine how sound effects serve meaning in recent style verse, one needs a different strategy.

When Kao and Mei base their investigation of poetic functions on exploiting the intrinsic features of Chinese language, the result is remarkable. This can be seen in their detailed analysis of a special type of productive ambiguity, which is possible only in Chinese language. The first example they examine is a couplet from the first poem in the "Autumn Meditations," which can be read "as it is" at first glance—

Clustered chrysanthemums have twice opened  tears of other days;The forlorn boat, once and for all, tethers my  homeward thoughts.

叢菊兩開他日淚孤舟一繫故園心

A different reading, which requires both the poet and the reader to pause and dwell longer on their meditation of the images in front of them, is no less arresting:

Clustered chrysanthemums have twice opened, and  tears of other days are shed; [End Page 55] The forlorn boat is tied up for good, and my  thoughts go home.

叢菊兩開，他日淚孤舟一繫，, ，故園心

We have here an example of "ambiguous parallelism," in which "a couplet whose two lines each have two grammatical structures, and the structures pair off two by two" (54). Then, in a couplet taken from the second poem in the same series, we have a different kind of poetic configuration, more commonly seen than the ambiguous parallelism above:

It is true then that tears start when we hear the  gibbon cry thrice;Useless my mission adrift on the raft which came by  this eighth month.

聽猿實下三聲淚奉使虛隨八月槎

This is a typical example of "pseudo-couplet," pseudo because—in the original, but lost in the translation—what is "paralleled" is the form, not the content. In other words, this is "a couplet whose two lines are grammatically parallel at the level of words, compounds, and phrases, but not at the level of deep structure" (55). This will become clear once we set the "meaning" right:

聽猿三聲實下淚奉使虛隨八月槎

The parallelism is lost, and the reader is allured to turn the "correct" word order in the initial line back to its "incorrect" form so as to enjoy the aesthetic effect of the desirable parallel structure. The desirable communications on different planes are thus realized between the poet's choice and the reader's appreciation, between message and grammar, and between superficial structure and deep structure, amid the welcome noises of "variety, dissonance, contrast."²⁶

Kao and Mei are not satisfied with merely revealing how the ambiguous and the pseudoparallel work but also want to explain in plain language many whys behind the functions of these two devices of poetic rhetoric. First, "it is the couplet as a structural unit that provides the natural environment in which ambiguity and pseudo-parallelism flourish." The evolution of the couplet structure, in turn, is affected by an increasing tendency in the development of the recent style poetry "to dispense with grammatical particles in exchange for economy of expression" (55). The driving factors behind this tendency can trace to the unique properties of Chinese as a noninflectional analytical language.

On the concluding page of their exercise, Kao and Mei acknowledge unapologetically that nothing they say is likely to alter the generally accepted [End Page 56] opinion about Du Fu's art. However, they believe that the criteria and values presupposed on the study of Du Fu before their thesis are, "without exception, peripheral to the central concern of poetry." "This exercise in linguistic criticism," the two authors say with confidence, "has provided some evidence" for the achievement of Du Fu's verbal artifacts (73).

If the first product of Kao and Mei's research is a modest "exercise," then the ensuing projects of their joint effort are meant to be methodological steps toward a much more ambitious goal. For example, although their 1971 thesis states by its title, "Syntax, Diction, and Imagery in T'ang Poetry," that it is an investigation of the constituent elements of recent style poetry, it has a grand plan, inspired by Northrop Frye's insight: the study of the larger structural principles, the recurrent general stylistic features, or the "aesthetics," of recent style poetry. The roadmap for the study looks like this: "We will begin with the simplest linguistic unit capable of assuming a poetic function in itself, namely, nouns or noun phrases. Next come the attributive sentence (noun followed by stative verb), the intransitive sentence, the transitive sentence, etc."²⁷ The approach is decisively linguistic and can be carried out only through close reading. The seemingly meticulous—almost to the point of being trivial—examination of the "minimal components of a poem" (62) is conducted in a well-controlled manner in the context of the antithesis and interaction between "texture" and syntax. By texture Kao and Mei mean "the local interaction of words—once the eminent domain of Empsonian criticism" (61). The language used in recent style poetry is already weak in syntax. In addition, the linguistic and prosodic features such as the natural pause after every disyllabic phrase at the beginning of a five-syllable or seven-syllable line, the independence of each verse line, and the autonomy of the parallel couplet unit in recent style poetry also conspire to impede the forward-moving syntactic drive (63–64). "When syntax is weak, textural relations abound"; hence, "ambiguity is the norm instead of the exception" in recent style poetry (91). Keeping this in mind, the authors subject to scrutiny a series of poetic functions of Chinese language in recent style poetry at the level of words and sentences. At one end of this series, the convenient juxtaposition of nouns, the "simplest linguistic unit capable of assuming a poetic function in itself," is explained as the result of the "discontinuity" of verse lines caused by their "having too little grammar" (64). At the other end, the reason for the following fact becomes self-evident: the unifying syntax, typically absent in the middle section of a poem, is often expected in the concluding couplet, where—and when—a discursive conclusion is needed. The discussion of the poetic functions thus proceeds along the related axes in the highly regulated discourse of recent style poetry: the imagistic language versus the propositional language, the discontinuous versus the continuous, the objective [End Page 57] versus the egocentric, the sensory awareness versus intellectual understanding, the spatial versus the temporal, and so forth (59).

The approach Kao and Mei adopt is not in everyone's favor. One of the faults found by its detractors is its alleged confirmation bias. As Stanley Fish puts it, "I found that in the practice of stylisticians of whatever school that relationship [between description and interpretation] was always arbitrary, less a matter of something demonstrated than of something assumed before the fact or imposed after it."²⁸ Kao and Mei's practice is free from the blame. Although the two authors customarily declare that they always have their interpretive intuition verified by the views of traditional commentators, and that the task of their linguistic inquiry "is simply that of pinpointing" the consensus of those commentators,²⁹ they have no need to make their analysis fit conclusions preconceived or already known. Two examples can be used to illustrate this. The first is Kao and Mei's revelation of the "characteristic copresence of concreteness in sensory impression and abstractness in reference in recent style poetry" (94), due to the well-known facts that simple images in Tang poetry have a strong orientation toward qualities instead of specific things, and that recent style poetry as a whole might impress one as being "pervaded by the dreamy abstractness, the suffused vagueness" (83), which W. K. Wimsatt discusses in his Verbal Icon.³⁰ Notwithstanding, by combing through numerous examples, Kao and Mei notice that simple images in recent style poetry has a way—"through its plenum of quality-evoking words" (94)—to create a different kind of concreteness. Concreteness in "sensory impression ('vivid')" is no less desirable than concreteness in "centrifugal reference ('specific')" (94). The two authors are not dismayed to find that their finding "may be regarded as a significant counter example to the Imagist theory of image making" (94).

The second example comes from their study of the "dynamic image" in Tang poetry. Thanks to the Chinese worldview, Chinese art criticism always holds in high regard the keen awareness of the dynamic vital process in nature. As a result, a dynamic image can be understood as a static image "coming alive," and not merely one of active transference of power through the agent-action-object causative relationship. Because of this, certain rules regulating the "universal" grammar are challenged by the uniqueness of Chinese language. To discuss the "dynamic images," Kao and Mei found that they had to do something not required of a student of European language and literature: come up with "a typology of Chinese verbs and their varying degrees of dynamism" (98). The first thorny issue they address is to clarify the meaning of a misnomer static verb, which refers not only to the copula shi 是 but also to adjectives playing the role of predicate. A better way to solve the problem, they say, might be to replace the label static verbs with "static verbs and the static aspect of other verbs" (98). [End Page 58] Then there are "verbs of perception and cognition" (99) or "superfluous verbs" (100), such as see (jian 見) or know (zhi 知), which "do not present an additional fact but serve to emphasize the facts already presented" (100).

The significance of Kao and Mei's practice cannot be overemphasized. What they do—without claiming so—is to upset the system of classification of parts of speech, a more or less "foreign" system imposed on Chinese language. On the surface, they seem to be trying to change the size label on a procrustean bed, but actually they are putting the old label on a new bed they create. In their study, the poetic function of "dynamism words" are examined more in the "vital processes happening in nature and dynamic interconnection between individual agents" (96) than merely in the process of action. This way, words denoting posture, location, connection, simile making, changes in time and place, and so on, all have active roles to play in the making of dynamic images (100–101, 108–9). So do various stylistic choices and rhetorical devices like the manipulation of "word classes," novel observation, the creation of similarity and contrast, inversion, resultative complements, and personification (103–14, 116–18). Kao and Mei demonstrate that dynamism is not only created through analytical syntax but very much also generated from the texture of interword relationships.

Having finished their investigation of the basic constituent elements at the level of words and sentences, which serve as artistic materials of recent style poetry, Kao and Mei go a step further and focus their attention on how these constituent parts work together to generate meaning at a higher level. The result of this study is found in their 1978 thesis "Meaning, Metaphor, and Allusion in T'ang Poetry."

The critical theory they adopt for this study is the principle of equivalence put forward by Roman Jakobson. The origin of the theory traces to Ferdinand de Saussure's thoughts on the syntagmatic and the paradigmatic relations in his semiotic system, but it is Jakobson who raises the possibility of applying the theory to the study of literature:

What is the indispensable feature inherent in any piece of poetry? To answer this question we must recall the two basic modes of arrangement used in verbal behavior, selection and combination. … The selection is produced on the base of equivalence, similarity and dissimilarity, synonymity and antonymity, while the combination, the build up of the sequence, is based on contiguity. The poetic function projects the principle of equivalence from the axis of selection into the axis of combination. Equivalence is promoted to the constitutive device of the sequence.³¹

The poetic functions Jakobson finds suitable for his approach are formal elements, especially those represented in prosody. Kao and Mei, however, believe [End Page 59] that they can apply the theory to the study of meaning. What comes to mind is the equivalent relationship between the vehicle and tenor in a metaphor and the pairing of a contemporary topic and a past event in a typical case of allusion. For example, in the following couplet,

Floating cloud, wanderer's mind;Setting sun, old friend's feeling,

浮雲遊子意落日故人情

the wanderer's mind is compared to the floating cloud, and an old friend's feeling to the setting sun. The metaphors conjure up in the poet's mind paradigms of equivalence along a vertical axis, the axis of selection. When the poetic thoughts are realized in language, it necessarily has to express itself in the sequence of a time flow, along the horizontal axis of syntagms, but—very important—without grammatical "connectors." No poetry in any other language can "materialize" Jakobson's theoretical module in such a perfect manner! Thanks to the lack of restrictive syntax in Chinese, the "vertical" equivalence relation can be kept almost intact in the "horizontal" line of combination, in which the component parts are "loosely," if not often ungrammatically, connected.

The similar projection of Jakobson's "principle of equivalence from the axis of selection into the axis of combination" can also be seen in allusion of various types, abundant in recent style poetry. Take the following example:

If Winged General of Dragon City were present,He would not let the Hunnish cavalry cross Mount Yin.

但使龍城飛將在不教胡馬度陰山

"The implication is that the present dynasty, lacking a general of Li Kuang's stature, has a border defense that is altogether too porous"; conversely, say Kao and Mei, "the further implication is that although the present is unlike the past in military prowess, it would be comforting if the two were more alike."³² Being a linguistic principle, "equivalence" places within its jurisdiction both the similarity and its opposite, the dissimilarity.

Again, Kao and Mei do not just describe how but also try to explain why Western theories work in the study of Chinese poetry. Through a close reading of numerous examples, they find that two unique features of the "refined" language used in recent style poetry that account for the characteristics of its texture and other poetic elements in local organization, and also contribute to the flourish of the equivalence relation in Tang poetry. The first of these is the quality-oriented nature of nouns and noun images (295, 298, and 346), which favors the equivalence relationship between members in the same or similar [End Page 60] "quality categories" (316). The other feature has to do with syntax: "Since Chinese is a language weak in syntax to begin with, and syntax is further weakened by various conventions in Recent Style poetry, the result is that the metaphoric relation dominates over its complement, the analytic relation" (287). And "metaphoric relation" is nothing but another name for equivalence. It is not surprising that "Jakobson's theory can account for the facts of Recent Style poetry with greater ease than for those of Western poetry" (287). At the same time Kao and Mei also conclude that, though remarkably powerful in accounting for the phonological aspects, the principle of equivalence would not work well, or at all, in the study of grammatical and referential meaning in Western poetry, almost for the same type of reasons that it works in recent style poetry (347).

Taking an overview of Kao and Mei's linguistic investigation of the aesthetics of recent style poetry, one will notice an organizing methodological pattern that allows examinations of various poetic functions at different levels and from different angles to inform and interpret one another. The mainstay of this pattern is a cross-reference system of the paradigmatic versus the syntagmatic. We have seen how the local organization of words and sentences are affected by the tension between texture and syntax ("texture is the inverse of syntax!"),³³ how new meanings are generated through the projection of metaphorical relation from the axis of selection into the axis of combination, and how recent style poetry as a tightly regulated verbal structure become the site for the "war of words"³⁴ between imagistic language and propositional language, between the searching for "the centripetal relations among words"³⁵ and the drive for centrifugal reference to the specific time, place, and experience. Indeed, even the methodological choice of Kao and Mei's linguistic probe itself should be understood in the framework of the complementary relationship between the metaphorical and the analytical. The two authors are fully aware that their critical mode "is alien to the Chinese tradition," and yet they can take the hint dropped by traditional critics—such as Liu Xie's 劉勰 (ca. 465–ca. 532) poetic "awareness" of certain "premises rooted in the Chinese tradition"—and "provide the structural analysis of the working of those premises" (323–24).

This organizing methodological pattern later allows Kao a vantage holistic point of view, from which he can examine the poetic functions in Chinese poetry in the light of his reconsideration of the most distinctive features of Chinese language. The convenient binary division between spoken and written languages, Kao observes, is inadequate for a true understanding of Chinese language. He therefore proposes an audacious and probably controversial concept: the antithesis between a "character language" (wenzi yuyan 文字語言 or ziyu 字語) and a "voice language" (shengchuan yuyan 聲傳語言 or shengyu 聲語).³⁶ [End Page 61] Neither depends on the other to exist. As an ontological entity, character language is not to be confused with a written language, defined as the written form of a language. The relationship between the character language and the voice language reminds one of the antithesis between the paradigmatic and the syntagmatic.

With this insight, Kao turns his attention to the ci, a poetic form originating from the voice language tradition, "voiced" in the form of singing and performing by folk musicians and entertainers in their outward expression of experiences with references to specific times and places.³⁷ Perhaps not coincidentally, the blossoming of this melodiously "fluid" (xuanlü 旋律 in Kao's term) (8) poetic form coincided with the heyday of recent style poetry, whose highly regulated physical structure (tuwei 圖位) (8) was created—in writing, by literati members—to accommodate their inward-looking reflection of personal sentiments at a focused lyrical moment (10). Just as happened several centuries before, when the Han dynasty literati appropriated the yuefu 樂府 folk songs, literati poets from Tang through Song, who became fascinated by the expressive power of the ci form, tried their hand at it. Eventually they transformed the originally musical art form into a refined verbal structure strongly informed with the aesthetics of the character language. On the basis of his comparison between the formal features of the folk ci and the literati ci, and the study of the connections between the poetic functions of the languages of literati ci and that of the literati recent style poetry, Kao concludes without hesitation that the formation of the literati ci is not the result of "natural" historical evolution but the product of the joint efforts of literati poets in their conscious pursuit for a new aesthetics.³⁸ What is most noteworthy in the long process of this genre transformation, Kao says, is the evolution of the xiaoling 小令 (small ci) form, a "transitional" genre between the early song lyrics of the voice language folk tradition on one end and the character language literati ci on the other. Through a close reading of the structure of xiaoling in various types, Kao is able to trace how the rhythmic temporal flow of sound (jielü 節律) and the spatial structure of character (tushi 圖式) compete with and intermingle into each other, and how the "horizontal forward drive of voice language" and the "vertical juxtaposition of character language" merge and join forces.³⁹

Comparing the structure of ci against that of recent style poetry, Kao points out the most fundamental difference between the two. The important role played by couplets in recent style poetry is taken over in ci by "concentricity" (tongxin jiegou 同心結構), a term Kao creates to refer to the basic structural unit that connects neighboring lines (16). Kao finds that while Jakobson's model of coordination versus equivalence can well describe the relationship between the two lines forming a couplet, it is not sufficient to [End Page 62] account for the interline relations in a concentricity (15–16). What connects the lines in a concentricity is a common thematic center or focus (zhongxin 中心 or jiaodian 焦點), which can be a word, an image, or just an idea, around which each line in the unit can describe or narrate "from a different angle or at a different point in time, involving various kinds of mental activities in addition to sense-impressions" (16). The integral structure of a ci poem consisting of several units of concentricity can be called one of "stratification" (16). This structure of concentricity/stratification works at more than one level. While each unit has its own center, all the units within a ci poem share a common center at a higher level. In this way, the whole poem is sustained by an "incremental structure" (18–19).

The "common center," therefore, functions as a "topic" (19), and the poetic acts performed by the lines in a concentricity unit serve as comments on the topic. What Kao is doing here is explicating the most distinctive formal feature of ci, using the "topic + comment" construction, one of the most distinctive linguistic features of Chinese language. The poetic function of this linguistic structure has attracted the attention of Kao and Mei in their study of the language of Tang poetry at the level of local organization, especially the noun + noun configuration.⁴⁰ Now Kao begins to see the potential of its application beyond the examination of syntax. Characteristic of the "character language," topic + comment construction is actually a mode of thinking in which a sense impression generated in the mind is perceived and appreciated repeatedly from different angles, receiving attention that keeps dwelling on, lingering on, and turning back to it (yichang santan 一唱三嘆).⁴¹ It is a reflective mode in the literal sense of the word.

What we have here is a tell-tale example of the linguistics-based literary study at its best. First, the function of topic + comment construction (in opposition to the "normal" subject + predicate construction) as a unique feature of Chinese language catches the attention of linguists.⁴² When Kao and Mei notice in recent style poetry the unusually abundant presence of this construction, they found it apt to use the paratactic feature of this construction to account for the three characteristic stylistic features of recent style poetry at the syntactic level: discontinuity, dislocation, and ambiguity.⁴³ Then, in his investigation of the development of the ci genre, Kao becomes aware of the analogous relationship between this linguistic construction and the poetic structure of a song lyric.

The tapping of the rich potential of the topic + comment construction does not stop here. More exciting results are seen in the use of this originally syntactic construction as a theoretical framework in the study of classical Chinese poetry. It is here that Zong-qi Cai's holistic structural investigation of the aesthetics of [End Page 63] Chinese poetry claims our attention. In Cai's practice, the topic + comment construction not only serves as an effective tool but also functions as the framework of a self-contained critical approach or, so to speak, becomes the structure of his critical vision.

Focusing on its poetic function, Cai defines the topic + comment construction as follows:

Instead of an agent responsible for some action or condition, topic refers to an object, scene, or event "passively" observed. Comment refers to an implied observer's response to the topic. As a rule, this response tells us more about the observer's state of mind than about the topic. The absence of a predicative verb between the topic and the comment aptly underscores their relationship as noncontiguous, nontemporal, and noncausal. The topic and comment are yoked together by the implied observer through analogy or association, in a moment of intense observation. The result is quite different from that of a temporal cognitive process. Topic + comment tends to reactivate the vortex of images and feelings, previously experienced by the observer, in the mind of the reader.⁴⁴

The definition itself deserves a close reading. The first thing to take note is the reciprocal relationship between the topic and the comment. The topic should be attractive enough to generate meaning by inviting observation. The comment, touched off by the topic, should react toward the topic. According to Jakobson's linguistic module thinking, the two are metaphorically "equivalent" to each other. The fact that the two are copresent "in a moment of intense observation" implies that the relationship between them is simultaneous, spontaneous, and synecdochical (as the observed's and the observer's feelings and attitudes intermingle and become parts of each other). Their relationship along the paradigmatic axis is also confirmed in negative terms. There is no predicative verb to yoke the two together, and so their relationship is "noncontiguous, nontemporal, and noncausal"—in short, not syntagmatic. All the above instantly calls forth a spatial coordinate system of what Jakobson describes as the projection of "the principle of equivalence from the axis of selection into the axis of combination" mentioned earlier in this essay.⁴⁵ The poetic effect which the topic + comment construction is capable of evoking is captured in the well-chosen image "vortex of images and feelings," which vivifies the force of suction from the direction of the topic to be commented, and the rapid centripetal inward-looking force of the comment toward the topic.

Verse lines in the topic + comment formation are abundant in the Shijing 詩經 (The Book of Poetry). Cai uses the first two lines from "The Peach Tree Tender" to illustrate this construction: [End Page 64]

Tender but sturdy, the peach tree,Bright and lustrous, its flowers.

桃之夭夭灼灼其華

[Mao no. 6, Mao shi zhengyi 1:279]

The peach tree and its flowers are the topics, and the two reduplicatives yaoyao 夭夭 and zhuozhuo 灼灼, with their radiating thriving liveliness, are pleasantly sensuous responses, or comments, touched off in the mind of the implied observer. The external objects and the inward responses are juxtaposed without any grammatical connective.

Thanks to its "extraordinary evocative power,"⁴⁶ the topic + comment construction not only works with shijing verse lines but also continuously plays its special role in the evolution of verse line patterns as various poetic genres transform and develop in later ages. Its operative machinery, so primordially detectable in the shijing verses, becomes more and more complicated and sophisticated.

The predominant line pattern in Shijing is a tetrasyllabic verse consisting of two disyllabic segments. Later, when the pentasyllabic pattern comes onstage with an additional syllable, the rules of the poetic game are fundamentally changed. Loathing its loneliness, the extra syllable tends to form alliances with existing syllabic combinations in the line, thus destabilizing the intraline balance of power. The unstable loyalty of the monosyllable to other members in the line and the multiple possibilities of permutation of syllable segments further complicate the situation. One of the consequences is the emanation of new sentence constructions capable of generating a plenum of new meanings.

Fascinated by the aesthetic effects thus generated, Cai focuses his attention on the formation of the new verse line patterns. What he discovers in the "unsurpassable" (dengfeng zaoji 登峰造極) artifice in an example like the following is "simply shocking" (rangren zhenhan buyi 讓人震撼不已):⁴⁷

[In] the morning wind I cherish my bitter heart;[Amid] the sound of crickets I lament the shortness  of time.

晨風懷苦心蟋蟀傷局促

The translation reads the couplet in the surface context of the poem. Since the two prepositions ("in" and "amid") are understood in the original, the "morning wind" and the "crickets" are not qualified. That is to say, the role they play in the two sentences becomes uncertain. Because of this, the morning wind and the crickets seem to presume the role of the subject and vicariously feel what the persona feels in his imagination: [End Page 65]

The morning wind cherishes a bitter heart;The sound of crickets laments the shortness of time.

晨風懷苦心蟋蟀傷局促

An example of the pathetic fallacy, this interpretation is possible only when we read "morning wind" (chenfeng 晨风) and "crickets" (xishuai 蟋蟀) metaphorically (for, literally, the wind does not cherish feelings and crickets cannot lament). Cai therefore calls it "ambiguous reading," or xuyi 虛義. Just as we marvel at the poetic effect of this somewhat equivocal reading, we are surprised by the revelation that "morning wind" and "crickets" happen to be titles of two poems in the Book of Poetry.⁴⁸ One of them does "cherish a bitter heart," and the other "laments the shortness of time." The syntax of the verbatim reading—which Cai dubs shiyi 實義—of this shijing-related subtext,

"Morning Wind" cherishes a bitter heart;"Crickets" laments the shortness of time

《晨風》懷苦心《蟋蟀》傷局促

matches the grammatical structure of the xuyi reading above. As the wind and the crickets from the earlier texts urge a more xuyi reading of the present text, the "exact meaning" of these two seemingly simple lines becomes indeterminate.⁴⁹ Indeed, if we find it difficult to fit this example into one or more of William Empson's seven types of ambiguity,⁵⁰ he probably would create an eighth type to accommodate it. For uninitiated readers, the xuyi reading is good enough to enjoy the poetic beauty of the couplet, but for those who have ideas about its shijing connection, they can bring their knowledge into their more intense appreciation of poetic art.

Cai might not be the first to notice the Shijing origin of the "morning wind" and the "crickets," but he is the first to detect, and to explain in plain language, the subtle interaction between the shiyi and the xuyi illustrated above. In the shiyi reading, one sees a logical and affirmative statement in the subject + predicate format, whereas in the xuyi reading the relationship between the imagined subjects and their predicates is ambiguous. The morning wind and crickets function not really as agents of actions but, rather, as outside stimuli that induce the persona's feelings.

Cai's analysis of the overlapping of xuyi and shiyi here is about the intense interaction between metaphorical language and analytical language, between the analogical-associative relation and the temporal-logical relation, that is, between the topic + comment and subject + predicate constructions. Without using the terms, Cai demonstrates in his illustration that, in the best examples from the "Nineteen Ancient Style poems," even in sentences apparently belonging to the [End Page 66] "simple subject + predicate" category,⁵¹ one can still sense the verve of the topic + comment syntactic structure so prevalent in the Shijing verse lines.

An even more illuminating discovery by Cai, however, is the poetic function of the topic + comment versus subject + predicate framework at the level of stanza and the overall organization of a poem. To understand how this works, we can take another look at the two Shijing lines examined earlier, this time in the stanza where they are embedded:

Tender but sturdy, the peach tree,Bright and lustrous, its flowers.This girl is going to marry,Good for her house and family.

桃之夭夭灼灼其華之子于歸宜其室家

In contrast with the first two lines, the next two are in the subject + predicate form, with line 3 a declarative announcement and line 4 a subjective judgment. While the topic + comment at the top of the stanza displays a contemplative representation of scenery, the subject + predicate at the bottom expresses subjective thoughts and feelings.⁵² There seems no logical connection between the two. The flourishing peach tree has nothing to do with the coming marriage of the young lady; they are just "yoked together." The relationship between the two parts in the stanza is "noncontiguous, nontemporal, and noncausal."

If we take a closer look, however, we can see that the first two lines present a mini drama, in which the implied observer is excited by a peach tree in full blossom. For the persona, who is lost in a moment of observation, some analogy between a productive tree and a young lady going to be married does not seem totally unlikely. In this way, the topic + comment construction on top serves as the topic, or an outside stimulus, which prompts an associative response, or a comment, represented in the subject + predicate construction at the bottom. Suddenly this sounds like a discussion of bixing 比興 (literally "compare and evoke"), a perennial topic of debate in the study of Chinese poetry.

And not just sounds like—it is: "We may contend that a topic + comment constitutes an analogical-associative framework" (563). Through a close syntactic analysis of the analogical-associative mode of representation, Cai pins down the elusive mechanism of bixing. The replacement of the translation of the term "comparing and evoking" by that of "analogical-associative mode" is not a small change (563). It marks a shift from the appreciative to the descriptive, from the metaphorical to the analytical.

The word mode in the new translation of bixing also tells us that Cai does not limit the use of the analogical-associative framework to the examination of syntactic or stanzaic choice—he takes it more for a mode of thinking. For [End Page 67] example, when we see a poem consisting of a juxtaposition of a topic in the form of a stanza (or stanzas) and a comment in stanzaic form, we have a poem in the binary topic + comment structure. A good example Cai uses to illustrate this structure is "The Gourd Has Bitter Leaves" (Bao you kuye 匏有苦葉; Mao no. 34, Mao shi zhengyi 1:302–3), a poem "marked by neatly balanced external depiction and inner reflection, with a transitional couplet (lines 9–10) placed in between" (565). The scaffold of the form mirrors the structure of the lyrical act, with the topic + comment construction extended to the compositional organization of the poem.

In his further exploration, Cai finds that, because of its effectiveness in the organization of depictions of external scenes and expressions of inner feelings, the binary poetic structure based on the topic + comment configuration "became the dominant structure in pentasyllabic poetry during the Han and the Six dynasties."⁵³ The structure is found in fourteen of the "Nineteen Ancient-Style poems."⁵⁴ Later, the structure was codified in recent style poetry, the most typical structure of which "features nature description in the first two couplets and emotional expression in the other two couplets."⁵⁵ The capacity of the binary formation can be expanded in various ways, such as arranging a "parallel, yet progressive clusters of the scene-emotion combination," resulting in an "aggregated" formation. Together with the linear structure, the binary and the aggregated formations "have become archetypal structures of Chinese poetry, with numerous variants developed in different genres after the Book of Poetry."⁵⁶

In this way, Cai develops a self-consistent theory on poetic structure, which germinates from, and take as its base, the topic + comment principle. The theory will bring fundamental changes to the methodology in the study of generic structural features of poetic genres, the interconnections among different genre structures, the intragenre transformation and evolution of poetic forms, and so forth.

Productive results brought about by the application of Cai's theory have already been seen in his own research. We can see this in two examples. The first shows how the creative use of the topic + comment principle in the close reading of an individual song lyric contributes to the study of the organizational functions of a key generic element of ci poetry. This is a case of moving from the small to the big, and from the local to the whole, or, to use a fancy phrase for a particular reason that will come up later, from the concrete to the universal.

The variety of archetypal structures mentioned in Cai's theory are not the products of certain manipulation of the topic + comment and the subject + predicate modes by means of mechanical stacking and attaching. The persistent search for the most appropriate poetic forms for variegated lyrical expressions always involves a thoughtful matching, coordination, and adjustment of [End Page 68] structural components of different shapes and tones. To exemplify the working of this intense interaction between the thematic demand and the formal necessity from converse directions, Cai dissects as follows the structure of Liu Yong's song lyric "Facing Swishing and Splashing Evening Shower Sprinkling from the Sky over River" in the tune of "Basheng ganzhou" 八聲甘州: Horizontally, each lead word at the beginning of a line and the lyrical acts it leads form a topic + comment relationship, representing one step in the persona's reflective experience at a specific lyrical moment. Vertically, the series of lead words shows a progression of the persona's thoughts and feelings. What Cai calls "the interlocking aggregate structure"⁵⁷ is almost graphically mapped—almost because, while the paradigmatic relationship between different layers of experience is clearly visible, the difficulties in typesetting (on the part of this essay) do not allow the imaging of the syntagmatic relationship between the lead words—a step-by-step linear poetic process along the axis of time, which should look like this:

對⇒漸⇒是處⇒不忍⇒望⇒嘆⇒想佳人⇒誤幾回⇒爭知我

The concentricity/stratification structure Yu-kung Kao conceives of, mentioned above, finds a perfect match in Cai's elucidation and analysis here.⁵⁸ The ingenuity of this poetic form, Cai explains further, does not stop here. This "interlocking aggregate" structural design allows the lyrical act to occur on two planes. On one plane the persona loses himself in the contemplation of a poetic "inscape." On the other plane, he observes and traces the twists and turns of the lyrical experience from the viewpoint of an "other," following the hints dropped by the series of leading words, hence the "double subjectivity" structure not seen before or later in the Chinese poetic tradition.⁵⁹

It is close reading that enables Cai to make his insightful discoveries of this kind. Close reading is not just reading closely. As John Crowe Ransom says, close [End Page 69] reading is the methodology of a "systematic" study.⁶⁰ Cai's close reading of Liu Yong's art is not meant to be an exegesis of a particular poem. By applying the topic + comment versus subject + predicate framework to his examination of lead words, Cai advances onto a higher level the study of the poetic functions of a key element in the development of the ci genre. The specimen to dissect is small, but the holistic vision of his linguistic investigation is always at the back of mind.

The second example shows how the holistic perspective provided by the same framework helps explain the cause and effect behind the intragenre transformation of poetic forms. So, this is a case of using the whole and the overall to reveal the small and the local, or going from the total to the individuals. "If we trace the development from Han yuefu to Late Tang regulated verse, or from the early short ci to the late long ci poems on objects," Cai observes, "we can perceive a clear intra-generic trajectory from orality to literacy."⁶¹ Then, history can repeat itself time and again, and sometimes in a reverse manner: "Interestingly, an obsessive pursuit of textuality (diction) and intertextuality (allusion) often marks the last great glory of a thoroughly 'literatified' [wenren hua 文人化] genre and heralds the rapid ascendancy of a new genre of oral folk origin."⁶² The drive behind these two trends from different directions can be explained in light of the interaction epitomized in the topic + comment versus subject + predicate framework. As the stylistic features of the oral tradition are characteristic of the subject + predicate relationship, and those of the written tend to be topic + comment oriented, it is only fitting to think of the "process of imitating, and eventually transforming an oral tradition into a purely literary one by the literati"⁶³ as the aesthetics of the paradigmatic winning over that of the syntagmatic. Cai's earlier study of the evolution of the pentasyllabic poetry, together with his "afterthoughts" in later works along the same line, is a good example of this.⁶⁴ His "aesthetics-conscious" perspective is also notable in the idea of shibian 詩變 (the change of poetic form), which defines his study of the xiaoling's evolution from recent style poetry as the literati's quest for a new poetic form.⁶⁵ In fact, the overall editorial scheme for Cai's landmark anthology How to Read Chinese Poetry is informed with his reflective considerations on intragenre development.

Cai's approach to his research target, therefore, is no less enlightening than his findings. Enthralled by the poetic effects produced by the intricate play between the topic + comment and the subject + predicate modes, traditional critics and commentators have made numerous comments on it. At their best, these comments and observations seem to come close to a certain understanding of the workings of its poetic function yet still fell short of pinning that down in specific language. Most of the commentators can only be content with their ineffable impressionistic responses, or expressed amazement. The following [End Page 70] comment on the syntax, and perhaps also on its impact on the structural organization, of the Nineteen Ancient Style poetry is an interesting example: "There are no syntactic rules in the 'Nineteen Ancient-Style Poems,' some people say. That is not true. Those poems surely follow syntactic rules of their own, but leave no signs of their distinctions for us to trace."⁶⁶ The rules are there and yet are not there. The comment does not seem to be helpful. The next comment, by Fang Dongshu 方東樹 (1772–1851), seems to have noticed some "signs of their distinctions," which the first commentator fails to see:

When these ancient people wrote, if there was a forward movement there must have been a backward movement; if there was a thrust downward there must be a thrust back upward. To soar like a startled wild goose or to wind along like a swimming dragon: this is the way we follow their rules of composition and the way we seek to understand their meaning. Having grasped this point, we will understand why these poems are thought to be "seamless like clothes made by heaven."⁶⁷

Cai obviously likes the comment, as he offers this generous compliment: "In this single passage Fang sums up all I have said about the two aesthetic movements."⁶⁸ But does Fang? He certainly describes in imagistic language the impression made on him by the poetic beauty in question. And yet, "unexplained beauty arouses an irritation in me"⁶⁹—we share this typical New Criticism complaint from William Empson and would like to have that unexplained beauty explained. A comparison is in order between Fang's remarks and what Cai says about the two aesthetic movements.

Credit should go to Fang for his intuitive feeling of the interaction between two forces in the organization of a poem, the "forward movement" and the "backward movement." Unable to grasp the intangible movement of the forces represented in language, however, he has to resort to comparison. What he can see in the nearly visible kinesthetic—hence more or less tangible—movement of a calligrapher's brush seems comparable to the intangible feeling he experiences in his psyche when reading a poem. Exactly what the shape of that feeling is depends totally on Fang's readers' acumen in their understanding of the movements of some "startled wild goose" or that "swimming dragon."

Reading Cai's comment on the same subject side by side with Fang's, one wonders whether the two critics use the same language:

In the "Nineteen Ancient-Style Poems" both the binary structure and the multilateral texture are born of the constant movement back and forth between the outer and inner world in the poets' process of imagining. In turn, they activate similar movements of the temporal and spatial imagination of the silent reader. In the mind of poet [End Page 71] and reader alike, the intensification of these two aesthetic movements will lead to a point where the boundary between the outward and the inward dissolves and a timeless and spaceless poetic vision emerges.⁷⁰

The language is unadorned, with words trimmed down to the very essence. The only word that looks beautiful is aesthetic, but that is exactly what the passage is about. Of course, to fully understand what Cai means, it takes a good understanding of the binary structure, together with its underlining framework of topic + comment versus subject + predicate construction discussed above. All this, in turn, is sustained by Cai's close reading of various linguistic elements of poetic language, exemplified in his meticulous categorization and tabulation of various types of syllables and syllabic units, their combinations, syntactic types, prosodic patterns, organizational structures, and so forth.

As Cai continues, he does need some help from the figurative language—a "common phrase"—to describe what Fang refers to as the seamless "clothes made by heaven": "While this aesthetic principle was established with the 'Nineteen Ancient-Style Poems,' it became the ultimate matrix for all the intricate rules of temporal progression and spatial correspondence—in rhythm, meter, grammatical category, and semantic meaning—in T'ang regulated verse. Later this aesthetic principle is often spoken of with this common phrase 'moving in a circle; going and returning' [hsünhuan wang-fu 循環往復], and is observed as a golden rule for writing and reading Chinese poetry."⁷¹ Recalling Yu-kung Kao's dialectic attitude toward what Jakobson labels as the "subjective, censorious verdict" based on "a critic's own tastes and opinions on creative literature,"⁷² we can see that Fang and other traditional commentators belong to the camp of the critic. According to Kao, appreciative criticism by these literary critics is creative in nature.⁷³ Endowed with a sharp insight and equipped with a logic of its own, the traditional metaphorical criticism serves as a rich resource anyway. To draw on this resource, one needs "to analyze objectively the subjective critical experiences."⁷⁴ Cai does just that: going over his research, one cannot but marvel at the sincere respect he pays to the subjective traditional criticism. However, in his negotiation with the wisdom of traditional critics, Cai never satisfies with repeating or paraphrasing them, or simply translating their language from the classical into the vernacular, or from the Chinese into the foreign, passing the ineffability of their intuitive ideas to his own readers. His dialogue with Fang Dongshu seen above is but one of many good examples. When he explains how "Fang sums up all I have said," he is using his plain analytical language to lay bare what Fang can feel only metaphorically. Actually, his innovative interpretation of bixing, discussed earlier, is another product of an intense dialogue with some of the best critical minds in Chinese literary history, [End Page 72] such as Liu Xie, Zhong Rong 鍾嶸 (ca. 469–518), Kong Yingda 孔穎達 (574–648), and Zhu Xi 朱熹 (1130–1200).⁷⁵

If the metaphorical understanding of traditional critics smacks of—to stretch the use of now the most familiar module used in this survey—the topic + comment principle, then Cai's analytical discourse is a cool-minded demonstration of the subject + predicate, which is specific, down to earth, and nearly "scientific." The complementary relationship between the two types of mindset and the two modes of discourse reminds us of the clever remarks on one—or more than one—key concept left for us by Wimsatt, one of the most outspoken New Critics, while he ponders over the meaning of the paradox of "the concrete universal":

A modern literary critic, John Crowe Ransom, speaks of the argument of a poem (the universal) and a local texture or tissue of concrete irrelevance. Another literary critic, Allen Tate, manipulating the logical terms "extension" and "intension," has arrived at the concept of "tension" in poetry. "Extension," as logicians use the word, is the range of individuals denoted by a term (denotation); "intension" is the total of qualities connoted (connotation). In the ordinary or logical use of the terms, extension and intension are of inverse relationship—the wider the one, the shallower the other. A poem, says Tate, as I interpret him, is a verbal structure which in some peculiar way has both a wide extension and a deep intension.⁷⁶

Judging from the "figure of sounds" in Wimsatt's voice, we can tell that, no matter how "peculiar" the way the poetic functions function, the ideal poetry he envisages can live only in a poet's wishful dream—unless he composes in Chinese, and unless he uses not only voice language but also character language, a verbal structure in the physical-material sense of the term. As to the antithesis between denotation and connotation, or between the individuals and the total, Yu-kung Kao has already proved that the relationship between the two does not always need to be inverse. An image in Chinese poetry can at the same time be vividly specific and abstract, or concrete and universal (a good time to recall the function of poetic clichés in the song lyric). And now, Zong-qi Cai has just demonstrated that the theoretical framework of the topic + comment paradigm versus the subject + predicate syntagm can be projected from the poetic texts to the critics who close-read the poems—just as Stephen Owen shows at the beginning of this article how a learned narrator and an innocent yet curious novice reader can inform each other in an intense act of close reading. Maybe we should not feel surprised to see that, in the semantic field of Chinese, poets, texts and close readers are a perfect match, and the framework of the topic + comment plus the subject + predicate is omnipresent. [End Page 73]