The linguistic turn in literary study opened a new theoretical horizon for the discipline of literature, but it also inevitably brought limitations, and was faulted by many scholars such as Abrams. Its most fundamental defect is this: applying the theoretical paradigm of linguistics to the study of literature, and thinking the problems of literature solely from the angle of language, in effect also expelled the “human being” from the field of vision—author and reader alike were told to roll up their sleeves and stand aside; and once the human being was estranged, the aesthetic and affective character of literature, and indeed its complexity as a spiritual phenomenon, could no longer be addressed. Whether Saussure's linguistics, Bréal's semantics, the New Criticism, Roland Barthes's structuralism, Derrida's deconstruction, or Foucault's discourse theory—all isolated language from the human being and pondered language, and even literature, in isolation. Even Bakhtin's metalinguistics, which closely relates language to social-historical context and ponders discourse and literary problems in language's historical use, especially in the process of dialogue, still treats the human being, within its theoretical horizon, as no more than a “mask” equivalent to a social role, rather than the rich and protean “human being” itself. To escape the malady of Saussure's synchronic study of language, Jakobson tried to draw on communication theory to reform it; but his “six-element theory” of linguistic communication did not truly solve the problem. Although it added new elements such as “sender” and “receiver” to the theoretical model and examined the dynamic changes of linguistic structure and function, in fact he again falsified the diachronic process of communication into the synchronic relation of structure and elements. His “sender” and “receiver” are at most a “transmitter” and a “receiver,” far from a living “human being,” so that the exceedingly complex spiritual activity of the human being in literature is reduced to a mere technical process; as for “information” and “code,” their connotation seems even thinner than that of “linguistic sign.”
That literature is the art of language, and takes language as the core of study, is of course not wrong. The problem is that the language in literature is living language, the concrete, vivid operation of language, closely related to the human being's spiritual activity; to interpret literature merely with an abstract linguistic model is bound to simplify and distort literature and the true state of literary activity. The linguistic study of literature must grasp “speech” (parole) rather than “language” (langue). As Saussure explained of “speech,” he said: “The individual is always its master; we call it speech” (1). When we turn our gaze toward speech, we necessarily grasp at the same time that exceedingly important factor, the “human being.” From this angle, Austin's speech-act theory provides the most apt new paradigm for literary study. It not only avoids the limitations brought by the linguistic turn, but, by virtue of its concrete, detailed description of the speech-act, makes it possible to disclose in depth the inner mechanism of literary activity and to explore the mystery of literature.
Austin's speech-act theory is precisely an interpretation of the concrete operation of speech. Austin's main view is that to say is to do, and that this is a behavioral system: it contains a series of actions at three levels—the locutionary act (the phonetic act, the phatic act, the rhetic act), the illocutionary act, and the perlocutionary act—together with illocutionary force. The performance of a speech-act must be combined with a specific social context and the corresponding specific social convention (2), and requires the sincere participation of the participants (the addresser and the addressee). This takes language together with the human being and the various factors of the social-historical context as an integral whole for examination, thereby displaying more truly the actual state of speech activity. Only that Austin regards literary language as a non-serious, insincere act—hence hollow or void—and excludes it from his field of vision. Since what Austin examines is the actual operation and effect of everyday discourse, with particular attention to a specific domain among the many kinds of discourse, the theoretical paradigm he proposes is strongly targeted; yet, confined to the particular form of everyday discourse, it cannot but limit its flexibility and range of application. But his methods of thought and basic views nonetheless provide us an important reference. Thereafter Searle, Fish, Derrida, de Man, and Miller respectively developed and extended speech-act theory from various angles. Yet this theory still requires us to rethink it and to remold it as a whole.
I
We hold that the discourse-act should be regarded as a behavioral system jointly constituted by three dimensions—the expressive act, the constructive act, and the referential act. These differing dimensions of action coexist within the structure of the discourse-act system, and changes in their mutual relations lead to changes in discourse function. As an act, discourse possesses force and intentionality: the force of discourse derives mainly from the illocutionary act, while discursive intentionality coincides with the referential act of discourse and is concretely manifested in different referential and intentional modes—external reference, internal reference, self-reference, intertextual reference—and in different referential relations such as denotation or exemplification. All three of these dimensions interweave and adapt to one another, jointly constituting the most basic systemic structure of the discourse-act. Within this behavioral system, on the whole only one dimension of the discourse-act holds the dominant position, and the modes of performance of the other discourse-acts must adapt to it; therefore, as the dominant discourse-act is replaced, the whole discourse-act system and functional structure are bound to be adjusted.
The factors determining the structure and function of the discourse-act system are manifold. Discourse, as a dynamic act rather than a static combination of words, has a systemic structure and function that are protean and indeterminate, related both to the features of discourse itself and to the social-cultural context and to the human being's attitudes and intentions. But from the angle of discourse itself, the three features—situationality (the real context), rhetoricity, and logicality—may roughly distinguish the various discourse types. These three features often coexist within each discourse domain, intertwined with one another, and generally only one feature plays the dominant role. Within the discourse-act system, the primary-and-secondary relations among the referential act, the constructive act, and the expressive act change accordingly. For everyday discourse, situationality often holds the advantage, and the referential act of discourse holds the dominant position, manifested mainly as external reference. Only through the “vertical principle” by which discourse is linked to reality is external reference strengthened; only when the discourse-act interacts with reality—through the exchange of information and the interaction of action with the real world—and is effectively performed does discourse possess practical force. To be sure, the referential process of discourse is at the same time a constructive process; it is an act in which reference and construction interweave. But the constructive act can only occupy a subordinate position: it is conditioned by the referential act and is necessarily corrected by the referred “object,” so that its construction is in fact a “reconstruction” within the referential process. Everyday discourse activity is continually corrected in the interaction of referential act and constructive act, constructing an “apt” “objective world.” It is precisely external reference that maintains the close relation between discourse and the real context. Apart from the referential act of discourse and the real context, the interaction cannot proceed smoothly, and the illocutionary and perlocutionary acts could not be performed. The real context is one of the most important factors. Only under the above conditions are the locutionary, illocutionary, and perlocutionary acts performed; they are therefore merely discourse-acts of the second level.
Logicality is like a chain that binds the discourse-act tightly, intending to make the discourse-act transcend the fetters of the real context and to reject the intervention of the human being (the receiver, the interlocutor). The chain of logic deliberately throttles the discourse's capacity to construct and to refer to reality, leaving behind almost only the discourse's expressive act. From this angle, scientific discourse is fully self-sufficient; it casts off the entanglement of the real context and of the human being and strives to become an “oracle,” an eternal truth “valid everywhere under heaven.” Its expressive utterance is not, as Miller says, not an “act”; it is likewise an expressive act, only one that has nearly smothered the constructive and referential acts so as to possess full self-sufficiency. It perches solemnly and aloofly above the human being and the world, either decking itself out as an “authority” for people to worship and gaze up at, or looking down on ordinary mortals and ossifying into the “monologue” Bakhtin speaks of. When everyday discourse stresses external reference, the referred object becomes the focus of the human being's conscious attention, while discourse itself recedes into the background of the referential act—this is what Zhuangzi calls “forgetting the fish-trap once the fish is caught,” and precisely the “self-forgetfulness” of language that Gadamer indicates. In such referential activity the human being establishes an objective and cognitive relation with the real world, and language, as a mere “tool,” is overlooked (one sees through it in the “foreshortening” of perspective, in Ingarden's sense; it sinks into what Heidegger calls the “ready-to-hand”).
Literary discourse, by contrast, is characterized by the constructive act. To highlight its rhetoricity, so as to attract and influence the reader to participate in the construction of discourse and to foreground the constructive act of discourse and jointly create a discourse-world—this requires that the referential act be set aside and converted into internal reference. The author or reader is immersed in the joy of creation and lives within the world constructed by discourse, while temporarily forgetting the real world. In this condition, the discourse's own “horizontal convention” begins to play the principal role, and the “vertical principle” by which discourse is linked to reality is temporarily suspended (Searle). That is to say, when the constructive act of discourse occupies the dominant position, the referential act is degraded to a subordinate act, bound to change with changes in intentional mode; at this point any reference is necessarily confined within the fictional discourse-world. It can only degenerate into reference internal to this constructed discourse-world—internal reference—and cannot transcend this discourse-world to enter the real world directly. The link between the literary discourse-world and the real world is temporarily severed, and the two worlds become other to each other. Literary discourse creates a discourse-world of its own; it floats above the real world, gaining relative independence and self-sufficiency, and the fictionality, literariness, poeticity, and aesthetic character of literature are founded on this basis.
Within the fictional world constructed by literary discourse there still exists a second level of discourse-act, namely the various connections internal to this fictional world. The locutionary, illocutionary, and perlocutionary acts are likewise performed within the literary fictional world according to the rules Austin proposed, and produce effects—only that all this occurs within the literary fictional world and cannot enter the real world directly. This is precisely why Austin excluded literary discourse from his own field of vision. The illocutionary force of discourse here strengthens itself, breaking free of the real context, repelling the intervention of the human being, so that words and the discourse-world refer internally rather than externally. The referential object of everyday discourse becomes a concrete detail and figure within the work-world; the everyday, actually existing object is transformed into a virtual object within the fictional world.
Because literary discourse takes the constructive act as its feature, its illocutionary force is strengthened, words and the discourse-world it constructs become the focal center of attention, and so the exceedingly complex combinations and metamorphoses of discourse, the “geological” richness of the word's implication, are all displayed. The “many-layered structure” of literary language expounded by Ingarden, the “polyphonic harmony,” is founded on this basis. The internal reference brought by the constructive act of discourse, like a whirlpool, draws the human being in and attracts him to participate jointly—the “gaps” and “blanks” form what Iser calls the “appeal structure,” strengthening and summoning the full cooperation of imagination. Hence the construction of the fictional world by discourse is at the same time the process of the human being's unfolding of imagination; the two interfuse and are inseparable. In this process the identity of discourse also undergoes a transformation: it is no longer a “tool” or an “object” but becomes a “subject” cooperating with the human being on an equal footing. A new relation is born: human being and discourse, human being and word, generate and interfuse with each other—precisely a “subject-subject” relation, a relation of “no-thing and no-self” or “the interfusion of thing and self.” It provides the human being with aesthetic experience, an ontological horizon, and is also the most creative state of the human being and language. The reason literature possesses sensuous richness, the reason it can give the human being a livable, roamable fictional world, is that in the constructive activity of discourse the implications of language are fully excavated—its sound, rhythm, color, imagery, savor, and metaphoricity are all activated by internal reference and participate in the construction of the fictional world. At the same time, all the human being's capacities—his perception and experience, imagination and consciousness, the conscious and the unconscious, thought and emotion, spirit and body—are all, under the summons of the force of discourse, opened up, and his creative power is given its fullest display. Thereupon a sensuous kingdom suffused with life, a fictional aesthetic world, unfolds before our eyes.
The discourse-act, as a special act of the human being, embodies the distinctiveness and richness of the human being. One may say that the discourse-act harbors all the human being's relations with the world, harbors all the human being's modes of existence and activity; the mode of the discourse-act is the human being's mode of existence. Because literary discourse, by its rhetoricity, foregrounds the constructive act of discourse and strengthens illocutionary force, thereby breaking free of the real context, its meaning is bound to be the least fixed and the most growth-laden.
II
In literary activity the mode of action of discourse is not single but possesses at once the constructive act, the expressive act, and the referential act; they influence one another, interweave, and shift unceasingly—only that, on the whole, the constructive act of discourse holds the dominant position and influences the modes of performance and function of the other acts. The writer continually tries to vary the dominant mode of the discourse-act, transforming the structure of the discourse-act system and thereby leading the reader to pass through the various discourse-acts. Literary discourse has the most complex and changeable modes of action, and therefore provides the human being with the most varied modes of existence and the richest experience: the constructive act and internal reference jointly create an independent fictional world, letting the human being gain a wholly new space of existence and fully enjoy the abundance and wonder of this sensuous world; the referential act (external reference) reestablishes the cognitive relation between the human being and the real world, making the cognitive and critical functions an important component of literary activity, whereupon ethical-moral practice too intervenes; and the self-reference of discourse foregrounds the act of narration itself, constituting “metanarrative” (as in Gide's The Counterfeiters); intertextual reference connects and communicates different texts, making many texts converge into a polyphonic symphony, effectively expanding the space of the literary work (as in Coetzee's Foe and the allusions of classical Chinese poetry), or deliberately creating another effect, parody (as in Eileen Aird... and Calvino's Invisible Cities). The multifunctionality of literature derives precisely from the multidimensionality and variability of the literary discourse-act, from changes in the primary-and-secondary relations of the discourse-acts of different dimensions; author and reader, moving continually among the various discourse-acts and crossing their borders, realize and reciprocally augment the different functions—the aesthetic, the cognitive, the ethical-moral, and so on.
In literary creation the writer often, by means of rhetorical strategy, endows discourse with illocutionary force so as to influence the reader's relation to discourse, regulating the structure and function of the discourse-act system in order to carry out his own distinctive creative intention—for instance, Shklovsky's “defamiliarization” and Brecht's “alienation effect,” both effective means of regulation. When discourse and word are defamiliarized, they are bound to become the center of human attention; the suspended external reference is forced to convert into internal reference, the relation of human being to discourse and word is changed, and discourse and word are revived; the constructive act of discourse seizes the opportunity to take the dominant role. Only when the constructive act of discourse is in the ascendant can a sensuous aesthetic world unfold before our eyes. Defamiliarization is not, as Shklovsky says, for the sake of prolonging the duration of perception; its essence lies in changing the everyday habitual relation between reader and discourse, changing the “tool” identity of discourse, and thereby changing the structure of the discourse-act system and foregrounding the constructive act of discourse. Brecht's alienation effect is the reverse: it, by means of interference, refreshes and foregrounds the referential act (external reference), pulling the reader out of the activity and the constructed fictional discourse-world, pushing the social phenomena involved before the reader's eyes, so that with a renewed mode of cognition and a critical eye he regards the social phenomena the work concerns—thereby strengthening literature's practical capacity and educative function.
The distinctive rhetorical strategy employed in Yu Hua's To Live constitutes an exceedingly complex relation of discursive force. In the novel the protagonist Fugui recounts his own life in a calm, detached, aloof tone. In the course of his life one blow follows another: he first squanders the family fortune through gambling, then a succession of natural and human calamities are visited upon him—his son dies, his daughter dies, his wife dies, his son-in-law dies, and at last even his only grandson dies, leaving only a scrawny old ox and the “two old codgers who refuse to die” depending on each other. Tragedy and comedy themselves possess utterly different “forces.” Comedy lays stress on recounting a character's “action,” its “meaning”; and because these actions violate the social conventions the reader is accustomed to, they separate the character and his action from the reader, push them to the foreground position, lay them bare, and force the reader to play the role of “onlooker” and “judge.” Tragedy, by contrast, recounts a character's fate, which cannot but touch the reader deeply, arouse his concern for fate, and invite him to undergo suffering together with the character, thereby producing resonance and sympathy. And the first-person narrative perspective in a certain measure filters and relieves that sense of guilt, providing the conditions for the reader's sympathy. At the same time, in To Live the protagonist-and-narrator Fugui is not pouring out his sorrows, still less weeping, but recounting his own tragic past in the discourse of a calm onlooker; and this is one who has already lost the keen pain of “living,” a “veteran” who has seen through life, even one who seems no longer truly “living” narrating his own “living.” This discursive strategy forcefully dispels the closeness inherent in first-person narration, and, by the repelling force the discursive register possesses, creates a psychic distance between reader and work. Moreover, this narrator “I” is one who makes a living roaming about collecting folk songs and ballads; Fugui's story is what he chanced to hear ten years earlier, in the mood of one gathering anecdotes—that is, it is retold in the first person by another “I” of ten years before. This not only creates the work's “intertextuality”; the use of the narrator “I” and this casually-heard recollection of an old man's bygone past, already grown indifferent to the years gone by, constitutes multiple perspectives, deepening the irreparable distance between reader and tragic story. In To Live there are interwoven throughout multiple discursive forces of different modes and different directions, which force the tragic story—that would otherwise make the reader unwittingly become a “kindred” “experiencer”—to alter its effect, inducing him to turn into an “onlooker,” “judge,” and “thinker.” After the old man Fugui has told his own story, the novel turns back to the “I” who is the ballad-collector and narrator, and through this “I” describes: “The old man and the ox gradually receded into the distance, and I heard the old man's coarse, moving voice drift over from afar; his song floated in the empty evening like the wind, the old man sang: ‘In youth I roamed about, in middle age I longed to dig for treasure, in old age I became a monk’—cooking-smoke rose curling from the rooftops of the farmhouses, dispersing and vanishing in the air filled with sunset glow.” Here Yu Hua, by means of the drifting, fading song, constructs multiple intertextual relations between the ballad and Fugui's tragic life, further broadening the work's domain, its sense of time-space, and its space of thought; the vivid sense of presence in the description pulls the gradually-receding bygones back before the reader's eyes, letting an atmosphere true yet faint, like evening breeze and sunset glow, linger about the reader's mood. Yu Hua uses various rhetorical strategies to strengthen the constructive force and function of discourse: it not only recounts for the reader a story of “living” and “dying,” but pushes a series of questions before the reader, forcing him to ponder “what after all is living?” “How is one to live?” “Is even a living as good as death still a kind of living?” “How should we face the various modes of living, the ways of life?”
III
The reading activity is a process of mutual action and mutual adaptation between reader and discourse. In this process, on the one hand, the features and intentionality of discourse induce the reader's reading intention; the reader unfolds his psychic process in step with the discourse-act, and, following the structure of the discourse-act system, its dominant mode of action and intentional mode, continually adjusts his own reading attitude to follow and adapt to the discourse activity, so as fully to realize the function of discourse. On the other hand, the reader's reading-psychic attitude and intention also influence the structure of the discourse-act system, leading to a switching of the primary-and-secondary relations among the various discourse-acts, making a certain discourse-act stand out while the others recede into the background, at the same time leading to changes in the referential and intentional modes of discourse, and finally causing the meaning and function of the work to vary. Discursive features and the reader's reading psychology, these two poles, jointly determine the structure and function of the discourse-act system. But because conventions are sedimented within discourse, and convention is originally an “agreement” reached between discourse and writer-reader, and has already become a kind of social unconscious, it permeates literary discourse activity and becomes an exceedingly important latent structural factor organizing the narrative and the constructive act of discourse, and is therefore bound to grant discourse priority. The distinction and study of genre would lose its basis and meaning were it otherwise. Only, conventions are after all only a “requirement,” an “ought,” not a coercive “law” or “must”; that is, they possess only “negotiability,” not “decisiveness,” and the reader may still violate these agreements and go his own way. For example, with respect to the “basic annals” and “biographies” in Sima Qian's Records of the Grand Historian, the historian values the corresponding social conventions and cultural conventions and follows them to pursue what is referred, either striving to verify historical facts or questioning what is stated, so as to seek out historical truth; whereas for the general modern reader, the original conventions have become so remote, even nearly losing their actual efficacy in regulating reading, that he reads merely out of interest, wanting to gain the pleasure of reading—whereupon the constructive act of discourse seizes the dominant position by force of circumstance, and narrative, by the vivid figures and winding stories it constructs, wins the reader's favor and “literature” takes on another aspect. As for deconstructionist interpretation, which disregards conventions, or deliberately stresses the historicity of conventions and ingeniously regards and reads the work otherwise, wantonly altering the structure of the discourse-act system so as to propose unexpected “novel views”—these modes of reading escape the author's intention by “unconventional” routes, dismantling the textual structure and opening new horizons of reading. Literary reading is the joint result of discursive features, social conventions and cultural conventions, and the reader's reading attitude; this is precisely why one and the same literary work can both become a common topic for mutual discussion and exchange, and yet produce wholly different understandings and effects, provoking endless divergence and dispute.
Since in literary activity the dominant act of discourse is often the constructive act, how then does this act operate? Is it a wholly rule-less, arbitrary act, or one with laws to follow? In “The Logical Status of Fictional Discourse,” John Searle analyzes fictional discourse in depth. He holds that truthful discourse must depend on the “vertical rules” that establish a connection between language and reality, whereas fictional discourse, by means of “horizontal conventions,” suspends the “vertical rules”; therefore “the pretended performance of the illocutionary act constitutes the work of fiction, and this pretended performance is in fact made to succeed through the performance of a discourse-act intended to invoke the horizontal conventions; these horizontal conventions suspend the illocutionary obligations that normally exist in all utterance.” According to Searle, the distinctive mode of literary discourse, or the patterned opening of a story such as “Once upon a time…,” all hint to us that the “horizontal conventions” replace the “vertical rules,” so that literary reading is regarded as a pretended discourse-act, or as the construction of a fictional world. What, then, besides the function Searle indicates of changing discursive intention, is the essence of these “horizontal conventions”? How do they organize the constructive act of literary discourse? This still awaits our further, deeper inquiry.
We hold that in literary discourse activity the “horizontal conventions,” as a historical existence, contain an important aspect, namely the content of “social conventions.” In analyzing the various domains of language, Morris pointed out that although the novel is fictional, “the domain of the novel is still connected with the actual world and actual conduct—although this connection is often not very close; it is, as it were, an imaginative exploration, providing the possible circumstances suited to actual conduct” (8). In fact, the most fundamental connection between the novel and the actual world, actual conduct, lies in social conventions. Social conventions are the most basic structural element of a society; they regulate human action (including speech, expression, dress and adornment, and interpersonal intercourse) and organize the various social activities. The real social order is built on this basis. Literature is the study of the human being; human action, intercourse, and conflict are its important objects of description. Even the human being's psychic activity is closely related to his action, mutually influencing, interweaving, and often expressed through action. Therefore literary discourse must construct the work with reference to social conventions; apart from social conventions, the writer cannot describe characters, recount stories, or structure the world of the work, and the reader cannot interpret the work, cannot fathom the deep root behind the eruption of conflict and the development of plot, and therefore cannot truly understand the work. From another angle, the writer's distinctive attitude toward, and handling of, social conventions determines the features of his work. For instance, a realist work must relatively strictly follow social conventions in describing characters' actions and organizing the plot; deviate from social conventions and the work loses its truthfulness and credibility. A comedy recounts a character's series of convention-violating actions, while using social conventions as the background of the character's activity and the standard for judging the action; in a fantasy work the character's capacity for action is exaggerated to the extreme, yet the social mode in which action is expressed must still conform to social conventions, for only thus can it achieve communication with the other characters and with the reader. If a work wholly violates and deconstructs social conventions, then the work is absurd; yet writer and reader must still understand social conventions, otherwise they cannot recount or understand this absurd world. From this angle, whatever the type of narrative literature, none can do without social conventions—they either follow social conventions from the positive side, or refer to them from the reverse side, or deliberately challenge and subvert them; social conventions always latently condition the narrative activity, condition the constructive act of discourse. Lose social conventions, and the action, speech, expression, and even psychology of the characters all become unintelligible, society too becomes unintelligible, and correspondingly the constructive act of literary discourse can no longer be performed.
Social conventions have already been deeply embedded in the bodies of writer and reader, becoming a kind of social unconscious; they permeate literary discourse activity and become an exceedingly important latent structural factor organizing narrative activity and the constructive act of discourse. Because social conventions are the product of history, formed by unequal social groups and living communities and undergoing differentiation, they are stamped with the imprint of different nations, classes, and strata, and with different ideologies, harboring power relations; politics, economics, morality, ethics, race, and the relations of the sexes are all contained within social conventions, and these have already become a kind of collective unconscious, operating without a trace, silently making unequal relations work. For instance, Su Tong's Raise the Red Lantern (Wives and Concubines) makes full use of the conflict among different social conventions to structure the work. Songlian was once a university student; the campus life and the relatively simple mode of handling interpersonal relations had formed in her a new-style way of life, so that even her psychic activity, her every move, could not shed its student air. Meishan, by contrast, was an opera singer; as a member of a special group at the margin of society, of low status and unsettled life, she had developed a personality at once lonely and self-abasing yet sharp-tongued. The specific social conventions these social subcultural groups enjoyed could only be subject to the social conventions occupying the mainstream position in feudal society. When they successively became the third and fourth concubines of Chen Zuoqian and entered this great feudal household, they could not but observe the strict conventions of the feudal clan. Yet the behavioral habits bred by their original conventions could not be changed at once; and conflict is born precisely from this. If at first Chen Zuoqian, finding Songlian's youthful immaturity and student manner novel, took a liking to her, it is precisely this mode of behavior of Songlian's that, again and again in the household's life, angered him. As the patriarch of a great feudal household, Chen Zuoqian is a ruler who brooks no challenge: he may, with grand pretensions, take a third, fourth, and fifth concubine, while the women must keep to their station and strictly observe the way of the wife, not transgressing in the slightest. At Chen Zuoqian's fiftieth-birthday banquet, Songlian, to show affection and ingratiate herself with Chen Zuoqian and to mend their relation, kissed Chen Zuoqian before the assembled crowd—only to make Chen Zuoqian feel doubly embarrassed and chagrined, so that he sternly upbraided and rebuffed her. Under these hypocritical feudal conventions, Chen Zuoqian may privately do all manner of base and filthy things, yet before the crowd must don the solemn mask of a gentleman of honor—not he alone, for even the servants are accustomed to this unequal convention. The inequality contained in feudal social convention brewed the tragedy of Meishan's being thrown into the dry well and Songlian's going mad, while Chen Zuoqian, just as merrily, went on to marry a fifth concubine.
Of course, social convention, as a structural element, is chiefly relevant to narrative literature; for lyric literature such as the lyric poem it is otherwise. Although the poet's creation cannot do without social conventions, such works often do not narrate “events” and so do not directly involve human action; social conventions therefore lose their importance. The most important factor structuring lyric literature is the law of life and emotion. The writer must not only follow the law of nature but must also follow the law of life and emotion, remaking nature according to the law of life and emotion so as to kindle the resonance of life and the resonance of emotion. Even narrative literature is likewise conditioned by the law of life and emotion, which has already been sedimented into literary form, latently regulating the rhythm of the discourse-act and the course of the plot's development, and setting a distinctive life-and-emotion keynote for the whole work. It is only that, because the actions of characters and the plot directly bear on human fate—and it is precisely out of concern for fate that the importance of characters' actions and plot in narrative literature is foregrounded—the presence of the law of life and emotion is concealed. Looked at one layer deeper, even the conflict between characters' actions and social conventions is a distinctive mode by which the law of life and emotion and social conventions mutually adapt. From this angle, the law of life and emotion is the conditioning factor sedimented in the deep layer of the “horizontal conventions”; the constructive act of literary discourse likewise cannot but be powerfully regulated by it, and thereby becomes the inner source that arouses emotional resonance and conveys the sense of life. Though social conventions originally have a complex link with the law of life and emotion, and have, in the process of human activity, been internalized and made unconscious, becoming the human being's behavioral habit, for life itself they are after all an external bondage. The law of life and emotion, on the contrary, belongs to life itself; it endows the human being with freedom and lets life be expanded. Therefore social conventions and the law of life and emotion, though coexisting within the horizontal conventions, often constitute, through their mutual opposition, the tension peculiar to literature. If in the realist novel social conventions occupy the relatively advantageous position, and the lyric poem fully exercises the law of life and emotion, then the modern stream-of-consciousness novel often wavers within the tension-field of these two regulating forces. And it is also for this reason that only when reality has become a bygone, departing far from us, and has grown relatively blurred and faint, does the breath of life and emotion latent at the bottom of the work lay itself bare, lending the work a faint, ineffaceable savor.
Different from the above two, cultural convention is the most surface factor among the “horizontal conventions.” Cultural convention is acquired in the course of the use of literary discourse; it refers not to the grammar and logic that ordinary language should follow, but to the distinctive “procedures” that literary discourse possesses, the distinctive discursive forms a specific literary genre enjoys. This is at once what the writer must rely on in writing and the reader's guide in reading. As one whose aim is innovation, the writer always intends to challenge old procedures and seek a mode of expression apt to his own life-craving; yet the old procedures and cultural habits become his heavy burden, and because this challenge faces a powerful cultural tradition and social community, he, beset by the “anxiety of influence,” can only, through “misreading” and “distortion,” remake his own distinctive mode of expression. The root of cultural convention is originally planted in human life; yet once drawn into a traditional “procedure,” it turns into something external that fetters the human being. So Wen Yiduo declared that writing poetry is “dancing in shackles.” On the one hand, cultural convention always has a thousand threads of connection with the law of life and emotion: with its variations of linguistic sound and rhythm, with imagery and the combination of images, it manifests the colors of life and emotion. In this sense the inner law of life and emotion is directly embodied in the external form of literary discourse, embodied in cultural convention. On the other hand, once literary discourse form congeals into the various “cultural conventions,” it is bound to depart from life and emotion themselves and to regulate the manifold modes of life-activity with “ossified” conventions; and as a kind of convention bred under a hierarchical culture, it cannot avoid being tainted, sedimented with hierarchical consciousness and unequal relations that distort the expression of life and emotion. Thereupon cultural convention again deviates from the law of life and emotion, and cultural convention parts ways with the law of life and the law of emotion, thereby constituting an inner conflict. Therefore literary activity often seeks the path of life's breakthrough in the crevice between following cultural convention and resisting it.
We may further concretize the “horizontal conventions” enjoyed by the constructive activity of literary discourse: the horizontal conventions contain at once social conventions, cultural conventions, and the law of life and emotion; they either coordinate or conflict, interweave and interfuse, jointly regulating literary activity and the constructive act of discourse. Among them, cultural convention is chiefly the factor used to regulate the organizational form of discourse; it, according to convention, guides the constructive act of discourse, endows discursive form with function, leads the reader's reading and adjusts the reader's relation to discourse, and thereby influences the structure and function of the discourse-act system. Social convention leads the action, intercourse, and conflict of the work's characters and the development of the plot, thereby shaping character and constructing the world of the work; that is, social convention chiefly regulates the events and features of the work. And the law of life and emotion operates at a still deeper level: setting out from the human being himself to correct the whole literary activity, it endows literature with an ineffaceable breath of life and an emotional ground-color. The law of life and emotion is also a most active regulating force in literary discourse activity; it is precisely its conflict with the various conventions that becomes the inner motive force driving literary innovation. The reason literary discourse activity is exceedingly complex is that it is conditioned at the same time by many conventions and laws, lodged within the tension-field constituted by different conventions and laws; and these conventions and laws change with changes in the social-historical context, thereby leading to the historical transformation of literary discourse activity. This complexity far exceeds that of a game: in a game the rules are single, simple, fixed, rigid, fair, and equal, whereas in discourse activity the conventions and laws are plural, complex, vague, and flexible, and themselves embody different power relations—so that even the writer cannot improvise on the spot at will, which makes this contest protean and unfathomable. In speaking of the law of literary evolution, Eagleton says that literary “form is at least a complex unity of three elements: it is partly formed by the history of ‘relatively independent’ literary forms; it is the crystallization of the dominant ideological structure…it embodies a series of special relations between author and reader.” If we examine this closely, it is not hard to find that all of this operates through the inner tension among conventions.
IV
In the constructive activity of literary discourse, the constructive act, as the dominant act, always forces the other discourse-acts to undergo variation and to cooperate with it; external reference is temporarily suspended—at this point all discourse-acts cannot but shrink within this fictional world, and reference can only be internal reference. Internal reference becomes an organic component of the fictional world. Isenberg once gave an example: if in a novel we read a sentence such as “On the 31st of December 1820, a man stood for some time at the door of Notre-Dame in Paris,” then it merely leads us into a fictional world; as for the real-world Notre-Dame outside the novel, it tells us nothing. That is to say, the constructive act of literary discourse possesses an immense assimilating force: with the fictional world it constructs, it replaces and supersedes the real context of the above everyday discourse, depriving the referential act of external reference and prompting it to follow the change of intentional mode and convert into internal reference, so that the concrete, actually existing Notre-Dame and the whole event are transformed into a concrete detail within the fictional world, what Roland Barthes calls a touch of verisimilitude in the rhetorical arrangement of the work-world. The constructive act of discourse, as the dominant act, always forces the other discourse-acts to vary and cooperate, suspending discourse's direct link with the real world. The literary fictional world that literary discourse constructs also has the appetite to devour, digest, and assimilate scientific discourse, transforming it into material and nourishment for the fictional world—for instance, the scientific discourse involved in Elizabeth's lectures in Coetzee's Elizabeth Costello: Eight Lessons, all of which has been transformed into an internal element of this fictional world, or become a special touch in shaping the character.
Within the fictional world constructed by literary discourse there still exists a second level of discourse-act, namely the various internal connections of this fictional world: the commands, requests, promises, exhortations, and conversations among the work's characters—these discourse-acts likewise possess performative capacity within the fictional world, forcefully driving the characters of the work-world to organize all manner of intercourse, bringing about the intensification of conflict and the advance of the plot, and jointly participating in the first-level constructive activity of discourse, thereby interweaving the discourse-acts into an integral whole and weaving the literary work into a vivid, many-colored world. Yet the referential act, after all, possesses a powerful inertial force; once the constructive activity of discourse ends or is interrupted by interference, the seamless cooperation of human being and discourse dissolves, human being and discourse split, and discourse once more sinks into the “tool” by which the human being deals with the world. Thereupon the referential act of words covertly seizes the dominant position of discourse, and often manifests itself as external reference—only that this reference has already been changed in time, and its referential object and referential mode have also been altered. Because the fictional character of the discourse's constructive activity makes the constructed thing hard to find a wholly matching, direct counterpart in the real world, reference can only be performed, through the “social conventions” shared by the literary work and the real society, by the “exemplification relation” Goodman speaks of rather than by “denotation.” “Social conventions” exist both in real society and in the literary work; as a structural element shared by the two, they build a bridge between the literary fictional world and the real world, so that one can refer to reality through the “exemplification relation.”
Frye once proposed that literary language is both “centripetal” and “centrifugal.” He held that the “centripetal” is the larger verbal pattern formed by words, forming a “self-sufficient verbal structure”; the “centrifugal” is words moving toward the things they refer to. Both tendencies exist at the same time, but in the literary verbal structure the final direction of meaning is inward; the criterion of outward meaning is secondary. In fact, the “centripetal”—the constructive act of discourse—holds the dominant position, while the “centrifugal” is the referential capacity of discourse (external reference). Frye acutely noticed the multiplicity of literary discourse; only he did not go on to discover that the “centripetal” and the “centrifugal” do not simply coexist, did not discover the switching of the dominant act of discourse between them. Ricoeur further pointed out that literary reference is a “deferred” reference, and refers by way of “metaphor.” He said: “The literary work, only under the condition of suspending the reference of descriptive activity, can through its peculiar structure display a world; in other words, in the literary work, by suspending the first-order reference, the discourse manifests its reference as second-order reference”—namely “metaphorical reference.” Ricoeur likewise spoke of the distinctive way in which literary discourse refers to reality and gave a detailed exposition; only he had not yet employed speech-act theory and so did not clarify the inner mechanism and connecting mechanism therein.
This distinctive mode of action of literary discourse endows it with a distinctive way and force of participating in social practice. The constructive act of discourse draws the reader into the constructive activity, separating him from the real world, deeply experiencing the creative process of discourse-construction and the existential savor within this constructed sensuous world; the exemplification relation refers to social phenomena, and because it discloses the specific social conventions behind these phenomena, discloses the structural principles of a specific society, it possesses at once breadth and universality and depth. The literary type (the typical) makes full use of precisely this distinctive behavioral mechanism of literary discourse. Take, for example, the depiction of Ah Q's character and his “tragedy of love.” Ah Q, who understands “the great defense between man and woman,” naturally knows that marriage must follow “the command of parents and the word of the matchmaker”; yet he, without instruction, spontaneously imitates the Western manner of the knightly gentleman, kneeling before Wu Ma to declare his love—and this kneeling, moreover, is not purely Western but Chinese-style, kneeling on both knees. The gentleman's wooing should be both romantic and courteous, yet Ah Q can only blurt out crudely: “Sleep with me, sleep with me!” This Sino-Western blend, this mode of action mixing the refined and the coarse, neither fish nor fowl, wholly deviates from the conventions of the society Ah Q lives in and even more from the rules of action prescribed by his own status and position. Ah Q's comic character lies precisely in his repeatedly violating social conventions. The Ah Q shaped by these behavioral features is a person with no home, dwelling in the Tutelary God's Temple and doing odd jobs for others, his origins obscure, his surname and name vague, his conduct shadowy—a person at the very bottom of society. In a society of strict hierarchy, he is destined only to be exploited, bullied, humiliated, and unable to resist. The very conventions of this specific society have naturalized this unequal relation and prescribed that Ah Q must endure humiliation. Yet human nature and human wish are unwilling to submit, and so he can only turn to bully the weaker—Little D and the little nun—to gain psychic equilibrium, or simply adopt the “spiritual victory” method, attaining in imagination the position of the strong. Whether bullied by others or bullying others, Ah Q must obey the power relations within social convention. Even in his dream of “revolution,” he only thinks of moving the silver ingots, foreign money, and foreign cotton-gauze shirts of Squire Zhao's house, and of commanding Little D to carry them—“if he doesn't carry fast, slap his face”—and of seizing other people's women. Ah Q cannot conceive of changing the deep-rooted unequal convention; he merely reverses his own status and position, swapping places with Squire Zhao. That is to say, his violation of social convention is performed not by destroying the convention but by inverting roles; and the reader, too, only by understanding social convention can perceive the absurd ludicrousness of Ah Q's action, can interpret the distinctive psychology behind it. What is referred is not an Ah Q who exists in real life, but the social conventions embodied in Ah Q's action, and the conflict between social convention and human life, human desire—a distortion universally present in social reality and repeatedly undergone and experienced by people. Hence the figure of Ah Q exemplifies all similar social distortions, so that everyone feels Ah Q's shadow clinging to himself. Literature, precisely by constructing this figure of Ah Q and borrowing social convention to establish a link with real society, refers to and thereby criticizes reality through the exemplification relation.
The distinctive way of the constructive act of literary discourse, and of the referential act of discourse, together with its distinctive critique of reality, is the reason works of literature win a broader and more profound critical force than works of other genres. Compared with history, which attends only to “narrative,” literature lays greater stress on portraying “human action,” and therefore discloses more deeply the social conventions of a specific society, discloses the structural principles of that society. It is precisely in this sense that we can correctly understand Aristotle's view that poetry is more universal than history. The constructive act of literary discourse draws the reader into jointly creating the literary fictional world; and because the social conventions this fictional world refers to and discloses are at the same time the structural principles of real society, it gains a profound truthfulness and critical force. The fictionality and the truthfulness of literature are therefore not mutually opposed but cooperate in a distinctive way; and the literary type is built precisely upon this distinctive force of the discourse-act.