Narrative Cultures in the Mirror

This new journal, Narrative Culture, widens the scope of narrative studies to embrace traditional narratives across media and in various forms of cultural expression. Following this trend, I use the plural form "narrative cultures" to address narrative studies across cultures. Almost thirty years ago, Alan Dundes recalled that anthropologists and folklorists need to engage in bold and imaginative forms of the comparative method if they are to "get beyond the data-gathering or data amassing stage" (139), and his remark remains true today. A comparative approach to stories, and their cultural contexts, is crucial in order to understand particular themes and to gain insight into the workings of human imagination. At its best, intercultural comparison offers a chance to gain insights into vernacular cultures and to overstep the boundaries of common sense so as to think out of the box, ask fresh questions, imagine new problems.And yet the comparative approach is not without its dangers. All too often, folklorists and anthropologists turn narratives brimming with conceptual chal- lenges into reflections of their own common sense. In Through the Looking Glass and What Alice Found There, Lewis Carroll nicely explains how this principle works. As Alice stands in front of her drawing-room mirror, she speculates that the image in the mirror is a glimpse of an alien place she calls Looking-Glass House, where "there's the room you can see through the glass-that's just the same as our drawing-room, only the things go the other way" (133). There, "the books are something like our books, only the words go the wrong way." It follows (although Carroll leaves the inference to us) that to decipher those books you simply have to put their words back in the proper order so as to find your own script. According to this analogy, I call "looking-glass scholarship" the academic effort to find in alien settings the reiteration of things already known.Instances of looking-glass scholarship are rather pervasive in disciplines that rely heavily on typologies, such as folkloristics (where the drive to fit protean narrative data into standard genres, types, and motifs is ongoing) and anthropology (where different cultures often illustrate predefined social types). In this essay I address an exemplary case of looking-glass scholarship. I examine how a short article by Claude Levi-Strauss, called "Goodbye to the Cross-Cousin," deals with a complex narrative from a remote culture-The Tale of Genji, an eleventh-century Japanese literary masterpiece by Murasaki Shikibu.1 My subject is the interaction (or lack thereof) between a narrative hailing from the classic tradition of Western anthropology and a literary work from Heian Japan.2 Sir Edmund Leach has put on record that "it is to Levi-Strauss's lasting credit that he has made it once again intellectually respectable to indulge in broad cross-cultural comparisons" in an- thropology (xvi), and Levi-Strauss himself explained that his method is about using differences to think about commonalities (L'Homme 32-33). And yet, arguably, what he says about the Genji resonates with a leitmotif of Western thought rather than with anything Japanese.Looking-Glass AnthropologySo my tale of two narrative cultures deals with refracted meanings. In a sense this is inevitable because anthropology is an exercise in mirroring identities (and the same goes for folkloristics). As Clifford Geertz memorably put it,Know what [the anthropologist] thinks a savage is and you have the key to his work. You know what he thinks he himself is and, knowing what he thinks he himself is, you know in general what sort of thing he is going to say about whatever tribe he happens to be studying. (346)It is true that early anthropologists have set up the stage for their speculations by means of a stable contrast between two essences, "them" and "us." Take three short examples. Lewis Henry Morgan, who aggregated humankind into "ancient" and "modern" societies, depicted the lowest reaches of ancient society in terms of scant notions, promiscuous sexual mingling, and collective property (41, 384, 535). …

(where the drive to fit protean narrative data into standard genres, types, and motifs is ongoing) and anthropology (where different cultures often illustrate predefined social types). In this essay I address an exemplary case of looking-glass scholarship. I examine how a short article by Claude Lévi-Strauss, called "Goodbye to the Cross-Cousin," deals with a complex narrative from a remote culture-The Tale of Genji, an 11th-century Japanese literary masterpiece by Murasaki Shikibu. 1 My subject is the interaction (or lack thereof) between a narrative hailing from the classic tradition of Western anthropology and a literary work from Heian Japan. 2 Sir Edmund Leach has put on record that "it is to Lévi-Strauss's lasting credit that he has made it once again intellectually respectable to indulge in broad cross-cultural comparisons" in anthropology (xvi), and Lévi-Strauss himself explained that his method is about using differences to think about commonalities . And yet, arguably, what he says about the Genji resonates with a leitmotif of Western thought rather than with anything Japanese.

Looking-glass anthropology
So my tale of two narrative cultures deals with refracted meanings. In a sense this is inevitable because anthropology is an exercise in mirroring identities (and the same goes for folkloristics). As Clifford Geertz memorably put it, Know what [the anthropologist] thinks a savage is and you have the key to his work. You know what he thinks he himself is and, knowing what he thinks he himself is, you 1. /102/ Lévi-Strauss merged this short essay with another one in a bigger text, "Regards croisés" (in Lévi-Strauss, Regard). An English translation, "Cross-Readings," is available (in Lévi-Strauss, View). Unless otherwise stated, all quotations from Lévi-Strauss's paper are my own translations. All Genji quotes are from Royall Tyler's translation (Shikibu, Tale), and all are crosschecked for semantic accuracy with René Seiffert's equally imposing translation (Shikibu, Dit). 2. /102/ The Heian period in Japan covers four centuries or so, conventionally starting in 782 and ending in 1167 CE. During this time, the center of political activity was based in (or near) the capital city of Heian Kyō (the "City of Peace and Tranquility," present-day Kyōto). In a nutshell, this long period witnessed the decline of the emperor's power and the rising monopoly of political power by the Fujiwara clan, the withdrawal of Japan from the outside world, and the apogee of a refined court culture, engrossed with pomp and circumstance and yet diffident of worldly goals /103/ in agreement with the pervasive Buddhist ethos. Ivan Morris's classic study (1-40) offers a good description of this period in relation with the Tale of Genji. know in general what sort of thing he is going to say about whatever tribe he happens to be studying. (346) /87/ It is true that early anthropologists have set up the stage for their speculations by means of a stable contrast between two essences, "them" and "us." Take three short examples. Henry Morgan, who aggregated humankind into "ancient" and "modern" societies, depicted the lowest reaches of ancient society in terms of scant notions, promiscuous sexual mingling, and collective property (41,384,535). Sir James Frazer, for his part, associated savages with a tenacious belief in magic as part of a "solid stratum of intellectual agreement among the dull, the weak, the ignorant, and the superstitious, who constitute, unfortunately, the vast majority of mankind" (55-56). In a similar vein, Emile Durkheim contrasted the mechanical social solidarity typical of "lower" societies with the organic solitarily underpinning Western societies, and he declared that among Naturvölker "general mental confusion" was the rule (3). 3 Overall, these are variations on a single theme-the Western privilege of social and mental distinctions, in contrast to a state of aboriginal (and persistent) confusion.
Durkheim develops this shared axiom with remarkable clarity. He declares that the achievement of classification based on "the idea of a circumscription with fixed and definite outlines" is altogether recent. Among the lowest peoples, who dwell in undifferentiated quasi-hordes, there is also "complete indifferentiation between sign and thing, name and person, places and inhabitants." This failure to differentiate persists in the lowest segments of developed societies, he declares, for even "a considerable part of our popular literature, our myths, and our religions is based on a fundamental confusion of all images and ideas." And Durkheim points out that "metamorphoses, the transmission of qualities, the substitution of persons, souls, and bodies, beliefs about the materialization of spirits and the spiritualization of material objects, are the elements of religious thought or of folklore" (3-4).
As Durkheim states that the learned classes of civilized societies have the privilege of thinking along clear-cut categories, he takes his own social type as the apex of a hierarchical order. From this lofty vantage point, Durkheim and his contemporaries gaze at the lesser types scattered along the chain of human progress. But, of course, anthropologists did not invent this way of thinking. Their marked tendency to reduce the rich variety of human societies to a hierarchy of stable essences belongs to a hoary tradition. You may be aware of Alfred Whitehead's famous remark that "the safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato" (39)-to, that is, Plato's assumption that true reality lies in a world of immutable and eternal /88/ ideas. 4 As Arthur Lovejoy memorably showed, Plato's essences yielded the notion of a great chain of being in which all creatures have a stable essence, and all are arranged in an immutable scale of dignity preordained by the Creator. This view was once so pervasive that without an acquaintance with it, Lovejoy warns, "no understanding of the movement of thought in the Occident ... is possible" (23). Its fundamental postulate of incarnate essences reflecting a supernatural design clashes, of course, with Charles Darwin's naturalistic paradigm of a chain of species yielding one another by means of random mutations and adaptations. But the proponents of evolutionist anthropology had no use for Darwin's evolutionary notion of purposeless evolution; instead, they plodded ahead with the teleological view of Western Man as the pinnacle of the human evolution-as if the entire human past strove towards becoming the anthropologists ' present. 5 In other words, the founders of anthropology plotted their discipline within the chain-of-being framework. This is perhaps understandable, for the essential contrast between "us" and "them" is a convenient positional device for anthropological storytelling. It empowers anthropologists as the rightful subjects of scientific inquiries while determining that subalterns everywhere are the objects of such inquiries. But the underlying presumption of a hierarchical scale of different essences raises an epistemological issue. Durkheim's argument about the confusions of primitive thought is, of course, a downright 4. /103/ Whitehead adds for further emphasis, "Thus in one sense by stating my belief that the train of thought in these lectures is Platonic, I am doing no more than expressing the hope that it falls within the European tradition." In the Platonic heritage, Whitehead specifies, "[t]he things which are temporal arise by their participation in the things which are eternal" (39-40).
5. /103/ I am taking this point from Beer (14). phantasy. 6 And yet, he is factually correct in saying that folklore and mythology feature metamorphoses, and display (as I prefer to phrase it) the art of thinking across categorical lines. Which raises the following question: how can anthropologists (and folklorists) presume to grasp a fluid way of thinking by means of models that take static and discreet categories as the norm? Or, as Wendy Doniger (O'Flaherty 10) put in another context, "why is it that people have attempted to apply hard scientific criteria to phenomena that they themselves have defined as soft?" This background information makes it easier to appreciate Lévi-Strauss. A self-professed "inconstant disciple" of Durkheim,7 Lévi-Strauss transposed his predecessor's dichotomy of contrasted modes of thought into a polarity of "domesticated" versus "savage" modes of thought. In this transformed dichotomy, "savage mind" does not refer to the mind of savages; rather, it denotes a universal state of the human mind-its spontaneous analogical mode, as opposed to its cultivated analytical mode. 8 Lévi-Strauss's "savage" is "cerebral" (Geertz 358) in the sense, I reckon, that Lévi-Strauss presumes that people everywhere strive to explain the world according to metaphoric patterns and symbolic properties, which are /89/ the operative means of the pensée sauvage. Lévi-Strauss, like Durkheim, posits that scholars work in terms of discreet categories whereas myths and folktales display a transformative pattern. But, unlike Durkheim, Lévi-Strauss grants that scholars may revert to transformative (analogical, or symbolic) thinking. In fact, he professes they must do so. In an unusually candid discussion, he puts the matter thus. Even though it is a "very dangerous game" to place one's own "intellectual mechanisms at the disposal of the traditional pattern, 6. /103/ As Needham points out, beliefs in metamorphosis do not entail that classificatory concepts are lacking; indeed, the very idea of ontological change could not arise without definite categories, "for to believe that a man may change into a parrot one must first have ideas of 'man' and 'parrot' so distinct that a change from one into the other may be conceived at all" (xi). so as to let it live and perform that mysterious alchemy that gave it solidity and permanence throughout continents and millen[n]ia," a measure of "special intuition and subjectivity"-and even a "grain of slightly mad recklessness," he adds-is necessary to understand mythic thought ("Art" 107-08). And Lévi-Strauss claims to have braved this danger (and reaped the inherent insight) first hand, for he depicts his four-volume Mythologiques as the result of the interplay between American-Indian thought processes and his own-and he describes his myth analysis as "the myth of mythology," of which he was not so much the author, as he was the unwitting executant . So while Lévi-Strauss upkeeps the typological and binary mode of thought he shares with Durkheim, he is aware of deeper strains of thoughtinvolving ambiguity, and accommodating paradox-that threaten the neat borderlines between categorical essences.

Adieu to the cousine?
Having briefly considered Lévi-Strauss in light of the specular worldview of anthropology, let us examine his interest in the romance-novel by Murasaki Shikibu. The Tale of Genji, written around 1005 CE, is a fictional blend of court-life description, historical chronicle, mythic allegory, and folklore repertoire that covers a period of about seventy years ahead of Lady Murasaki's own day. Here is how Lévi-Strauss presents the Japanese masterpiece: Written in the eleventh century, the Genji Monogatari by Murasaki Shikibu is not only one of the purest masterpieces of world literature in its poetic inspiration, in the poignant melancholy of beings and things it exudes, and in psychological analyses so deep and subtle that it took the West seven or eight centuries to achieve their equal. In this dense, slow narrative attentive to the finest details of Japanese court life during the Heian period, one finds a drove of precious ethnological data, /90/ especially about a social change that certainly took place elsewhere, too, but about which we have little information outside of this invaluable source. 9 (Regard 107) The title of Lévi-Strauss's original paper on the Genji, "L'adieu à la cousine croisée," conveys the idea of a final parting with one's 9. /103/ In this translation I follow the extant English rendering (Lévi-Strauss, View 73) with a few changes. female cross cousin. 10 This essay takes the Genji as a precious witness to subjective attitudes regarding cousin marriages when such marriages were going out of fashion at the courtly society of Heian Japan. Lévi-Strauss argues that male characters in the Genji deem cousin marriages boring and vulgar, lacking in excitement as well as in distinction, whereas marriages at a greater distance are an exciting means for creating unprecedented alliances.
This reading of the Genji is ancillary to a theoretical point. Supposedly, Lady Murasaki's story testifies to a social change that took place in Japan as well as elsewhere. The "disaffection" for cousin marriages in the Genji reflects Heian Japan's opening "to history"-an "evolution" bound to happen whenever "a society faced with history consciously accepts to enter it" (Regard 107-08, 113). This reading applies to Japan a preexisting model of the social evolution of human societies from endogamy to exogamy, from biologically close relations to ever-wider social alliances. Morgan's grand evolutionist scheme assumes that the breaking up of the intermarriage of brothers and sisters, and then of cousin marriages-"a conjugal system nearly as objectionable," Morgan says (58)-was a condition for the human "rise in the scale" (60) towards civilization. Lévi-Strauss, in a book dedicated "to the memory of Lewis H. Morgan," devised his own variation on his predecessor's scheme. In this variation (Structures xxi, xxiii, 48, 73, 533-4. Cf. Regard 79-92), the breaking up of the consanguineal family (a state of incest) yielded elementary structures of kinship (based on cousin marriages) that eventually gave way to complex structures of kinship (allowing for openended matrimonial strategies).
Lévi-Strauss is certainly justified in taking the Genji as a valuable source regarding the ethos and the praxis of the courtly culture of Heian Japan. But his reading of this story according to a preset scheme of evolution from elementary to complex structures of kinship is unpromising. It hinges on the claim that the Genji testifies to a historical shift in attitudes regarding cousin marriages-to a drift away from cousin marriages-whereas Lady Murasaki's novel describes a social setting revolving on a "marriage politics" that hinges on perpetuating cousin marriages.
10. /103/ Cross cousins are the children of a brother and a sister (whereas parallel cousins are the children of two brothers or two sisters). Thus, from a man's standpoint (the one Lévi-Strauss takes), the female cross cousins are the father's sister's daughter and the mother's brother's daughter.
As Ivan Morris explains the so-called sekkan system, /91/ This was the system whereby the Fujiwara leaders made sure that imperial consorts would be chosen exclusively among Fujiwara girls. As a result the head of their family was almost invariably the father-in-law or grandfather (or sometimes both) of the reigning sovereign. … [The emperor] came to the throne as a callow youth and was promptly married to a Fujiwara girl, their son would be appointed crown prince, and when his father was obliged to abdicate, usually at the age of about thirty, he would succeed him and the cycle would start again. ... [During the entire period with which the Genji monogatari is concerned], the country was governed by the emperor's father-in-law or grandfather who ruled as regent [sesshō] while the sovereign was a minor and as chancellor [kanpaku] as he came of age. And the incumbents of these crucial posts were the heads of the Fujiwara clan. (48, 50) So Lévi-Strauss bases his reading of the Genji on an unpromising premise. Fortunately, he makes an empirical point as he quotes three male characters who state, or imply, that marrying your cousine is a boring prospect. This is something we can check.
First, Lévi-Strauss considers a character known as Tō-no-Chūjō, who is Genji's lifelong friend and rival. Whereas Genji is the son of an emperor, Tō-no-Chūjō hails from the Fujiwara clan-and his sister, predictably, is Genji's wife. In the relevant scene, Tō-no-Chūjō muses that in the public eye a possible marriage between his younger daughter and Genji's (and his sister's) son, Yūgiri, "even if no disaster, still offered nothing brilliant." Tō-no-Chūjō then tells his mother: "they are relatives, and people will find all that a bit tedious and dull, which will be a pity for him, too. It would look better for him to be given a warm welcome somewhere strikingly desirable and quite unrelated." But Tō-no-Chūjō also tells his mother (as Lévi-Strauss does not tell us) the real, bitter reason why he objects to this "curious match between cousins" (Shikibu, Tale 387-88. Henceforth, Tale). 11 Some years previously, Tō-no-Chūjō had presented his elder daughter to the emperor in the hope that she would eventually become the empress; but his high ambition 11. /103/ René Seiffert's translation makes it clearer that because the intimate relation that these cousins have struck is unremarkable, everyone will take it for granted, which does not suit Tō-no-Chūjō's project of giving this daughter to the heir apparent (Shikibu, had been ruined by a rival candidate presented by Genji. 12 Tō-no-Chūjō, refusing to give up, had then entertained the project of giving his younger daughter as a consort to the next emperor-but in the meantime this daughter has started sleeping with Yūgiri. So Tō-no-Chūjō's prospect of marrying his daughter to Yūgiri amounts to recognizing his political defeat to Genji. Even so, as he realizes that Yūgiri is bound to become the most eminent commoner in the land, Tō-no-Chūjō soon /92/ starts pressing the young man to marry his daughter. Thus, he pursues in practice the very marriage he had bashed in words. And, come the marriage ceremony, Tō-no-Chūjō acknowledges that the bride's situation is more satisfying than that of her elder sister who is an imperial consort (Tale 567, cf. Dit 1:620)-a ringing endorsement for the marriage he had reviled out of bitterness.
The second character Lévi-Strauss mentions is Yūgiri himself, who in the next generation takes to enacting the Fujiwara marriage politics for his own advantage. He considers giving his prized sixth daughter to one of two eminent young men. Regarding his own (supposed) half-brother Kaoru, Yūgiri acknowledges that "so close a marriage might well be dull," but even so the idea of letting Kaoru "go to a complete outsider struck him as a great shame" (Tale 925, cf. Dit 2:431). So, the notion of keeping matrimonial arrangements in the family is clearly attractive. But Yūgiri decides to bestow his daughter on his uterine nephew Niou (who is the heir apparent) with the thought that it would be an embarrassment and a disappointment to have that prized daughter "stoop too tediously low" (Tale 931, cf. Dit 2:438). This rationale (marriage into the imperial line as a means to avoid a tediously low alliance) brings to mind the fact that Tō-no-Chūjō had deemed tedious a marriage that kept him from dominating the imperial line. The two cases, taken together, suggest something entirely predictable-in the Fujiwara sphere, the "tediousness" of a marriage relates to its perceived lack of political advantage, not to the connubial experience per se. 13 12. /103/ The reason for this political rivalry is that Genji is a prince demoted to commoner status in the Minamoto clan. Therefore, his descendants are not in the imperial line, and Genji competes with his friend in placing consorts with the emperor /104/ according to the Fujiwara marriage politics.
13. /104/ Considerations of love and happiness were ancillary to the political game of alliances, as Yūgiri's tirade about not allowing his favorite daughter The third character Lévi-Strauss mentions is prince Niou, who brazenly declares that he sees "nothing attractive" in the prospect of marrying Yūgiri's daughter (his own maternal cousin) when pressed to do so. But, again, one has to consider the context. Niou's problem is that allowing his maternal uncle to become his fatherin-law will curb his youthful freedom-his "roving fancy," as the narrator puts it-to pursue amorous adventures (Tale 866,932,cf. Dit 2:360,439). And yet, although Niou resents the prospect of a curb to his youthful freedom, he duly marries his maternal cousin.
Lévi-Strauss also mentions a reigning emperor who endorses cousin marriages apropos of an exceptional quandary. In a nutshell, the reigning emperor Kinjō is worried about his third daughter, whose mother-a favored concubine of modest background-had died precociously. While wondering about what husband might protect and honor this princess, Kinjō recalls how his father, the emperor Suzaku, had solved a similar problem. Suzaku had decided to give his cherished daughter by a favored concubine of humble background to his younger brother Genji. This /93/ had been an unusual step. "Genji" is actually a classificatory name given to blood princes turned into commoners in the Minamoto clan, which means that Suzaku had decided to give this princess to a commoner. But this marriage had worked well. It had provided the princess a son, Kaoru, who supported her in her ripe age. And so the emperor Kinjō decides to follow this precedent by marrying his own princess to her Minamoto cousin Kaoru (Tale 577, 929-30). Lévi-Strauss reads in this decision a craving for security. In his mind, cousin marriages are a means to neutralize imbalances in the social order, and thereby to reinforce the cohesion of the descent group, which explains the use of such marriages in precarious situations.
But the professed need to reinforce the cohesion of the descent group in precarious circumstances sounds like a Durkheiminspired mantra. While applying anywhere, anytime, it fails to explain why, in the stable social settings of Heian Japan, all the characters quoted by Lévi-Strauss engage in cousin marriages.
Nor are distant alliances the paradigm of amorous frisson in the (and his own reputation) to stoop too low shows. He says: "It would be all very well to choose for the sake of her happiness, but in the end it would be an embarrassment, and a great disappointment as well, to have her stoop too tediously low" (Tale 931, cf. Dit 2:438).
Genji, as Lévi-Strauss presumes. When Yūgiri, having married a cousin, eventually seeks romance and adventure, he woos another first cousin. Prince Niou's most exciting amorous adventure (with a woman he finds living in a secluded mountain village) turns out to be with a second-degree cousin. Moreover, both cases involve a streak of bonding between close male friends (who are also cousins), for Yūgiri actually woos the widow of his recently deceased best friend Kashiwagi, and Niou actually replaces his best friend Kaoru with the woman who replaces the woman Kaoru loves (I will return to this shortly). Such details suggest that amorous excitement in the Genji has to do with bundling close ties in tight knots, rather than with seeking distant alliances. Lévi-Strauss (speaking about Yūgiri, Kaoru, and Niou) acknowledges that "truth be told, the kinship relationships here become so complicated that these ties do not exclude others" (Regard 108). Precisely.

Symbolic patterns in the Genji
I have argued that Lévi-Strauss' reading of the Genji illuminates some enduring axioms of classical anthropology rather than anything Japanese. Presently I wish to draw attention to the interpretive costs of this approach. Because Lévi-Strauss reads the Genji in the mirror of a Western tale of binary oppositions, he misses the crucial point that the Japanese narrative relies on (as one critic put it) "the /94/ interplay of repetition and substitution. Like actions are repeated in a somewhat different guise with different actors and somewhat different results" (Bowring 24), which muddles categorical distinctions in accordance with a cultural "tendency toward semiotically blurred articulation" (Ikegami 16). So let us heed the pattern of repetitions and substitution in the Genji. As we break free of the mirror spell, what can we expect to see?
The story starts like a stepmother fairy tale. Once upon a time, in a certain reign, there was an imperial consort not of the first rank whom the emperor loved more than the others. As she gave birth to a beautiful son, a jewel beyond compare, it became yet clearer that she was the emperor's favorite. But this lady soon died, and the mother of the emperor's eldest son (the future emperor Suzaku) resented the shining boy who threatened her own son's expectation of being nominated heir apparent. Moreover, a Korean physiognomist found that the young boy "has the signs of one destined to become the father of his people and to achieve the Sovereign's supreme eminence," although this would entail disorder and strife. The wary emperor, rather than set the unprotected boy "adrift" as an unranked prince (Tale 13, cf. Dit 1:14-5), demotes him to commoner status in the Minamoto clan (a "Genji" is a bearer of the Minamoto surname).
So, the tale of Genji is about the riddle of a prince demoted to commoner status and yet fated to achieve the emperor's eminence. It is a tale of wondrous glory and (after Genji's death) of karmic reckoning. Throughout the plot, different characters shade into one another, so that what cannot be accomplished in the lifetime of one character is often achieved by other figures that become weakened extensions of the first. In particular, as Haruo Shirane notes, "the deprivation or death of a lover leads to an overwhelming desire to compensate for that loss by obtaining a substitutive figure..." ("Uji" 124). An overall obsession with tracing new forms of the same essences, with devising new variants for the same patterns, yields a leitmotif of incestuous longing that pervades the story in various ways.
Let us consider a major thread in the novel. After Genji's mother dies, the emperor marries the daughter of a former emperor who resembles the dead lady to an astonishing degree. The emperor transfers his affection for the dead lady to this princess, known as Fujitsubo. Likewise, Genji-upon learning how much Fujitsubo resembles his mother-becomes obsessed with her beauty. A few years later Genji sleeps with Fujitsubo. She bears an adulterine son (the future emperor Reizei) who passes for Genji's half brother. Then Genji repeats, in a weakened form, his violation of an imperial consort by sleeping with the young aunt-and /95/ promised consort-of his older brother, the emperor Suzaku. Caught in flagrancy, Genji lapses into exile.
Although Genji's exile has its proximate cause in the dalliance with Suzaku's concubine, it is the secret with Fujitsubo that gives weight to his trial. Conspicuously, Genji's exile follows the classic pattern of stories about a noble youth compelled to leave his homeland and to wander abroad, to be confronted by a severe trial that enables him to come of age and become a leader at home (Shirane,. Such exiles often involve purification and the expiation of sins, and they hinge on a symbolic representation of death and rebirth . Genji himself, at the lowest point in his exile, performs a purificatory rite that consists in casting adrift a life-size human figure laden with his sins and impurities. Still, he proclaims his innocence to the gods. Then a powerful storm almost kills him, but the same force leads him over the ocean to a wealthy estate further down the coast, in Akashi, where Genji seduces the local lady and begets a daughter. A wealth of details hint that the Akashi estate reflects the mythical palace of the Sea Dragon King, and this scene mirrors an ancient story in which an ancestor of the imperial lineage went down to the bottom of the sea, married the daughter of the Dragon King, and became the grandfather of the first emperor, Jimmu (see Shirane,. Then Genji is recalled to court, and the fate predicted by the Korean physiognomist starts to come true. His older brother Suzaku abdicates, and Genji's adulterine son accedes to the throne as emperor Reizei. The new emperor eventually bestows on his father a position equivalent to that of a retired emperor, which means that Genji acquires an emperor-like status. 14 Moreover, he starts competing with the Fujiwara in the game of marriage politics (which is why he thwarts Tō-no-Chūjō's hopes of becoming a chancellor.) First, Genji manages to have an adopted daughter proclaimed empress to the emperor Reizei; then he marries to the heir apparent (Kinjō) the daughter he begat in Akashi. Therefore, Genji becomes the (secret) father of the present emperor, the fosterfather of the present empress, the father of the empress-to-be, and (consequently) the maternal grandfather of the next heir apparent, prince Niou.
So the riddle set forth by the Korean physiognomist is now solved. Genji has achieved, as Shirane puts it, "what both the Heian emperors and the Fujiwara regents dreamed of but never attained" (28). Crucially, Genji achieves this unique position because he usurped his father's prerogative as he slept with his mother's simile. Arguably, a sexual transgression against the father by means of a mother /96/ proxy qualifies as incest by substitution. My point here is that the Genji hinges on a foundational act of metaphoric incest, which reverberates down a string of feminine substitutes (see . It is noteworthy that the traditional names for both Genji's mother (Kiritsubo) and her replacement (Fujitsubo) stem from the dwellings each of them occupies at the palace-respectively, the paulownia pavilion (kiri 14. /104/ Jun daijō tennō, "equivalent" to retired emperor, which, as Shirane (Bridge 27) points out, is the only fictional rank in the Genji. tsubo) and the wisteria pavilion (fuji tsubo). Both the paulownia and the wisteria flowers are lavender colored; and, after Genji sleeps with Fujitsubo, he finds a young girl, Murasaki, who remarkably resembles Fujitsubo (being her brother's daughter). After Genji kidnaps the girl and raises her as his foster daughter, Murasaki becomes his life-long companion. Relevantly, murasaki is the name of a gromwell from the roots of which a lavender dye is extracted. Shirane effectively sums up this metaphoric thread: Lavender, the color of affinity and erotic linkage, joins Murasaki to the Fujitsubo lady, just as it earlier associates the Wisteria Court lady with her predecessor, Kiritsubo, or the Paulownia Court lady. Like the blossoms of the wisteria, the flowers of the kiri, or paulownia, are lavender. (Bridge 47) The incestuous longing in this lavender thread has mythic roots, to which Lady Murasaki alludes discreetly. When Genji meets his older brother after retuning from exile, in an exchange of poems he refers to himself, "languishing in disgrace beside the sea," as "the forlorn Leech Child [that] for year after endless year could not stand on his own feet." In response, Suzaku enjoins Genji to forget the past bitterness now that they "at last have circled to meet again around the sacred pole" (Tale 276, cf. Dit 1:305). The brothers' exchange is based on allusions to the mythic story of the celestial gods Izanagi and (his younger sister) Izanami, who met around a sacred pole and had intercourse to procreate the islands and the deities of Japan. The first time round they proceeded the wrong way and produced the leech child (who could not stand upright and was sent adrift in a boat), and only at the second attempt was their act of creation successful (Nihongi 10-21; cf. Kojiki 50-51). At the most obvious level, Genji identifies himself with the leech child to rebuke Suzaku for having allowed his exile by the sea, whereas Suzaku alludes to the new meeting by the sacred pole to suggest that Genji's banishment is a bygone misstep now put to rights.
But the mythic template has subtler, and more momentous, implications. The ideographs for "leech child" also mean "'sun-lad' or 'sun-child'" (Kojiki 399), /97/ and folkloric contexts associate the leech child with Amaterasu, the sun-goddess from whom the imperial lineage claims descent . In this light, Genji's identification with the leech child suggests his imperial destiny; and his self-identification with a life-sized mannequin sent adrift (like the leech child was abandoned "to the winds"-Nihongi 19) gains new shades of meaning. After Genji sent adrift that lustral doll, eventually the winds pushed Genji himself over the sea in a small boat. His landing in Akashi at sunrise evokes "the light of sun and moon"; and, in Akashi, another significant scene presents Genji playing his seven-stringed koto "with the island of Awaji looming in the distance" (Tale 261-63, cf. Dit 1:285, 287). This confirms we are still in a mythical scenario because the island of Awaji was produced out of the same primordial coupling as the leech child-and, just as the leech child foreshadows the sun goddess, so Awaji was the "placenta" for the creation of other islands (Nihongi 13,15,17). All in all, the leech-child analogy suggests that the exile prepares Genji's rise to glorious brilliance. And the scene of Genji looking at the placenta of creation while playing his seven-stringed koto-the instrument of royalty (Field 70)-intimates an impending renewal of kingship.
Actually, even before Genji came of age, people at court used to call him "the Shining Lord" (Tale 15, cf. Dit 1:17), which expresses the "association of radiance and royal mystique" (Shirane, Bridge 11) also hinted in the name of the leech child. Crucially, Genji is not alone in this association. Lady Murasaki adds that since "Fujitsubo made a pair with him, and His Majesty loved them both, they called her the Sunlight Princess" (Tale 15, cf. Dit 1:17). In those early days, the luminous affinity between Genji and Fujitsubo (who are almost the same age) foreshadows the crucial event in their relationship. It implies, as Shirane notes, "that imperial power belongs to them and their child, the future emperor Reizei" (Bridge 12). Royall Tyler remarks that Reizei's conception and birth set Genji "on the path toward a hidden sovereignty of his own," and he speculates that Genji may have made love to Fujitsubo with this outcome in mind (257). However this may be, Norma Field remarks that after Genji's return from exile "there is an aura of archaic brother-sister rule" (30) in the way Fujitsubo and Genji control the court. (You recall, of course, that initially Fujitsubo takes the position of Genji's mother. The motherlike connotation of Fujitsubo has mythic roots, too. In the story previously mentioned, the father of emperor Jimmu married his own mother's sister, who was also his foster mother [Nihongi 103-04, 107-08]. So the mythical pedigree of imperial ancestors includes a marriage with a mother substitute as well as instances /98/ of brother-sister union. Fujitsubo's relationship with Genji effectively echoes both patterns.) Hence, the message Genji imparts to his brother also tells Suzaku that the prince once set adrift has emerged from obscurity to reestablish the power of the imperial lineage. Alas, Suzaku's allusion to a new meeting around the sacred pole is a piece of wishful thinking. It falls short of recognizing that this image really applies to Genji and Fujitsubo-who echo the incestuous primordial pair as they tighten their power on kingship and oust the Fujiwara faction 15 -rather than to Genji and Suzaku. Indeed, Suzaku soon abdicates in favor of Reizei, which fulfills the project of Fujitsubo and Genji. Henceforth, Genji's glory is unmatched.
But after a time of supreme glory, decline eventually settles in. The lavender axis of feminine replacements turns somber as the retired emperor Suzaku convinces Genji to marry his prized Third Princess, who is the daughter of a former Fujitsubo consort (a younger sister of Genji's beloved Fujitsubo). Genji's obsessive pursuit of "the murasaki women" (Field 215) goads him to comply with this dangerous request. Soon, a son of Tō-no-Chūjō, Kashiwagi, rapes this Third Princess and she gives birth to an illegitimate son, Kaoru. This rape of the daughter of a Fujitsubo consort reads like a weakened replica of Genji's own sin of yore; 16 but now the tide has turned. As previously the birth of Reizei had propitiated Genji's rise, presently the birth of Kaoru hastens his downfall. In quick succession, Murasaki dies and Genji himself passes away. Eventually the narrative topos shifts from the capital to a dismal village by a roaring river, Uji, the name of which invokes "gloom," "dreary," and "inside" (Okada 197;Tyler 283).
It is tempting to look at Akashi and Uji as the topological markers for, respectively, a narrative movement of ascent toward worldly magnificence and the matching descent toward dissolution and otherworldly renunciation. Whereas in Akashi Genji meets a woman and so becomes the ancestor of a future emperor, the mythical resonances of Uji include the themes of a woman hopelessly waiting for her lover, and of "kingship almost gained but lost" . Whereas Akashi hatches Genji's 15. /104/ Of course, at this point Genji's sexual connection is with Murasaki, the substitute for Fujitsubo down the lavender line.
16. /104/ Kashiwagi fancies that his sexual misdemeanor is a weakened variety of the crime of violating an emperor's wife (Tale 652); and the same comparison occurs to Genji, who broods on the parallels between this misdeed and his own sin of yore (660-61, 678). magnificence, Uji displays the dejected debris of that splendor. Notably, it harbors a younger brother of Genji, the Eighth Prince, who (supported by the Fujiwara faction) had competed with Reizei for being nominated crown prince, and-having lost-withdrew from court with his two daughters. Uji also receives the assiduous visits of Genji's pseudo-son Kaoru, whose fundamental doubt about his own origin leads him to study the Buddhist scriptures with the fallen prince. Out of the fateful meeting between Genji's defeated brother and Genji's /99/ pseudo-son, a new cascade of feminine substitutions arises. But this reiteration of the "search for a yukari (literally, 'affinity,' 'link'), or substitutive figure" (Shirane,"Uji" 123) is bleakly unproductive. It is led by Kaoru, "literally a child of deception," and systematic misunderstandings now prompt "a general sense of isolation and desolation" . A general sense of delusion thickens as the story draws to its last dregs.
In this last chain of feminine substitutions, after the Eighth Prince retires to a Buddhist monastery to die, Kaoru sets his sights on the elder daughter, Ōigimi. She, however, is bent on following her father's advice not to marry and on joining him in death sooner rather than later. Hence, Ōigimi entreats Kaoru to take her younger sister, Nakanokimi, as a replacement for herself. In a countermove, Kaoru encourages his best friend, prince Niou, to become Nakanokimi's wooer as a replacement for himself. (Niou does marry Nakanokimi, but then he neglects her after being pressed to marry Yūgiri's daughter, as described above. And Kaoru himself accepts to marry the emperor Kinjō's second princess, after the precedent of Genji marrying the emperor Suzaku's Third Princess, as described above.) Eventually, Ōigimi manages to starve herself to death; and now Kaoru regrets his decision to, as Okada put is cogently, "enlist Niou to be his surrogate for the sister Ōkimi had tried to make into a surrogate for herself..." (198). So Kaoru presses Nakanokimi to yield to his desire for Ōigimi (of whom Nakanokimi is, of course, the "likeness"). But Nakanokimi deflects Kaoru's wooing by mentioning a hitherto unknown half sister, Ukifune, who is the true "likeness," the astounding "memento," of Ōigimi (Okada 198). So Kaoru finds Ukifune and becomes her lover, but then Niou also seduces her. Ukifune, trapped between the two rivals in love, decides to throw herself into the Uji river.
An important semantic thread stands out in in this last chain of surrogates. Ukifune means a drifting vessel, 17 and the term hitogata ("likeness," literally "human form") designates Ukifune as a surrogate for Ōigimi. Remarkably, hitogata also designates the lustral mannequin Genji had sent adrift in his own likeness, laden with sins, at the lowest point of his exile. Indeed, Nakanokimi had been thinking about that sort of ritual hitogata in a lustral stream when she mentions Ukifune as a "doll image" for Ōigimi (Tale 954-55). Otherwise put, Ukifune comes up as a hitogata in the ritual sense of "a disposable scapegoat" (Field 259).
So Ukifune, the ultimate surrogate, personifies a ritual scapegoat as the story draws to a close. 18 This purificatory image suggests that a cycle ends along with the story itself. After a tale of dizzying success (followed, in the Uji chapters, by /100/ a deepening sense of alienation) readers ultimately face the stark Buddhist view that reality is an illusion and life itself is a dream-like bridge over which people cross from one state of existence to another (see Shirane,. The last chapter, called precisely "The Floating Bridge of Dreams," portrays Ukifune (in seclusion with a Buddhist nun) supposedly unable to remember what she "dreamed" previously, and staunchly refusing to look back on her past life and lovers (Tale 1119, cf. Dit 2:676). At story's end, the ultimate surrogate engages in purification and reckoning.
And yet, the lustral theme arguably hints an impending renewal. Genji sends out a lustral doll just before he enacts the mythic role of grandfather of the first emperor, and then Ukifune enacts her purificatory-doll role at precisely the time when Genji's grandson is the heir apparent at court. So the hitogata parallel again brings to the foreground the underlying mythical thread. The tacit suggestion seems to be that the scion of Hikaru ("Shining") Genji, prince Niou-whose traditional name expresses nioi, a luminous quality "related to hikari, 'light,' and to kakayako, 'to shine' or 'to be 17. /104/ In the Japanese text, this character (like all the others) bears no first name. But in a poem she refers to herself as "a boat loose in the current," ukifune.
This self-designation is apt, and the term invokes a host of semantic resonances in the novel, which lead Earl  to assert that "the evocative nature of the heroine's name-or identity-can scarcely be doubted." 18. /104/ Shirane offers a good discussion of Ukifune as a surrogate, including the notion that "this 'floating boat' (ukifune) will, like the ceremonial hitogata, be eventually sacrificed and washed away" (Bridge 156). In the same vein, Field asserts, "Ukifune is predetermined to serve as a sacrificial figure" (262). radiant'" (Field 228)-will start the new cycle of kingship arduously prepared by Genji. Shirane has shown that Lady Murasaki's novel sets the marriage-politics machinations of the plot in stark contrast to a golden age of "direct and benevolent imperial rule" still represented by Genji's father at the very beginning of the story (Bridge 8-10). And (if I am right) at the end of story a new golden age beckons in accordance with the cyclic warp of mythic time (see Eliade).

Through the looking glass
This reading of the main thread of Lady Murasaki's novel is but a modest exercise. Whatever the value of its propositions may be, the basic aim of the exercise is to provide an idea of how utterly Lévi-Strauss's discourse on the Genji fails to engage with its object. Whereas Lady Murasaki embeds her story in mythic time, Lévi-Strauss perceives the chronicle of a society entering history; and whereas the Genji revolves on sexual entanglements redolent of incest, 19 Lévi-Strauss talks about the excitement of exogamic alliances. I have noted that Lévi-Strauss's drive to find in the Genji a corroboration of the script of his Elementary Structures is reminiscent of the scene in which Alice stares at her drawing-room mirror to see "just the same as our drawing-room, only the things go the other way." And, of course, there is /101/ more to say about Alice. She goes through the mirror when its surface becomes "just like a bright silvery mist," and then she resolutely turns away from the image of her drawing-room, which "was quite common and uninteresting," whereas "all the rest [behind the mirror] was as different as possible" . Following this analogy, I cannot resist to speculate that if Lévi-Strauss had straddled across the Durkheim-inspired mirror of classificatory thinking he might have perceived a fluid way of thinking as different as possible from "domesticated" thought-a masterwork of analogical thinking, which is also a treatise in oblique references and in the semantic power of ambiguity.
Actually, as I pointed out, Lévi-Strauss the mythologist acknowledges the need to go through the looking glass, so to speak. Unlike Durkheim, he grants that savants may revert to transformative (analogical, or symbolic) thinking. We have met his claim that narrative scholars must allow the incubation of traditional schemes in their own thought processes, so as to allow "the doubly reflexive movement of two thoughts working on one another" to illuminate both ("Ouverture" 58). Fascinatingly, this description matches Alice's experience in Looking-glass Land. To put it in a nutshell, Alice dreams of the Red King (and the other characters in Looking-glass Land) even while she is being dreamt of by the Red King. Thus, Carroll implies that going through the mirror amounts to dreaming about someone who is dreaming about you, which is a decent match for Lévi-Strauss's portrayal of the mythologist's work as the doubly reflexive movement of two thoughts working on one another.
All this smacks of paradox, of course. Doniger sagaciously notes that the theme of the dreamer dreamt is about recursion, and she relates it to the problem of the receding frame in the interpretation of traditional narratives. 20 Lévi-Strauss has this problem in mind when he allows that each new interpretation of a myth will count as a variant of that myth from the point of view of the next interpretation, which entails that the traditional theme imposes itself on successive interpreters. 21 For precisely this reason, Lévi-Strauss maintains that it is a "very dangerous game" to place one's own intellectual mechanisms at the disposal of the traditional scheme. Still, as we saw, he allows that a measure of "special intuition and subjectivity" is required to understand traditional themes.
In the present case, Lévi-Strauss provides us with a precious negative lesson: he refrains from reaching out to symbolic thinking across the mirror, reads in the Genji a variant of his own script, and learns nothing. But, of course, the point is to put a positive spin on this lesson. Looking-glass scholarship and the search /102/ for meanings do not really go together; and yet, fascinating narratives from other times and places cry out for meaningful exchanges. Achieving such exchanges arguably entails placing your intellectual mechanism in the service of the alien object while also heeding the relevant scholarship. The point is, while a "view from afar" certainly helps to perceive patterns and trends, too aloft a stance may annihilate your object.
In hindsight, I think this essay suggests three things. First, in comparative narrative studies it is best to break free of homemade typologies if you mean to engage with the varieties of imaginary experience. Second, some problems can be adequately understood if one has read the Alice books. 22 And third, the Tale of Genji is one of the few truly outstanding literary masterpieces you should not miss in your lifetime.