In the last few decades scholars working in the area of pre-modern Japanese studies have become increasingly attentive to the workings of power and politics, and to the ideological underpinnings of even the most refined of courtly prose narratives and poetic compositions. Indeed, there has been a veritable explosion of academic writings that seek to analyse how gender and power operated in the historical, religious and literary texts of pre-modern Japan.

The field of medieval Japanese Studies, for the most part, has taken its cue from the work undertaken by sixties and seventies feminism, making its primary task the reinstatement of women into the historical, religious and literary narratives from which they have been systematically written out. In so doing a new generation of scholars have radically questioned the androcentric biases of earlier scholarly writings. The focus on the complex intersections between gender and class has provided new perspectives on a diverse range of interrelated themes such as family, marriage, household systems, property rights and women's labour in medieval and early modern Japan.[1] Recently, scholars have also increasingly turned their attention to hitherto neglected fields such as sexuality, and the social construction of motherhood, menstruation, child-birth, abortion and wet-nursing.[2]

The project of writing women's histories has reverberated across many other disciplines. A four-volume collection was published in Japanese in 1989 under the title Shirīzu: Josei to Bukkyō (Women and Buddhism Series),[3] and an edited volume in English, entitled Engendering Faith: Women and Buddhism in Premodern Japan[4] followed in 2002. These works heralded a new departure from previous research that had focused exclusively on religious groups formed around male leaders, or on sectarian histories and doctrinal debates within particular schools of Buddhism. They demonstrated how both the sources that were investigated, and the questions that were asked in such accounts, inevitably focused on institutions in which men played a central role, and from which women, for the most part, were excluded.

Scholars also turned to popular Buddhism, highlighting the everyday aspects of Buddhist faith and practice to demonstrate that women were indeed active participants in the religious life of pre-modern Japan. In recent decades scholarly attention has turned to the construction of gender in Buddhist discourse, and to an examination of the ways in which women were systematically denigrated and represented as inferior to men in a wide range of medieval Japanese literary/Buddhist texts.[5]

Academic writings on classical and medieval works in a wide range of genres ranging from romance fiction (monogatari), poetic diaries (nikki) and waka poetry, to popular tales (setsuwa, otogizoshi) have also sought to unveil the workings of gender and the operations of power that often lie behind the aesthetic, humorous, religious, or supernatural dimensions of courtly and popular texts.[6] The Heian classic, the Tale of Genji (Genji monogatari) is no longer read simply as a tale about the romantic escapades of an idealised hero, but rather as a particularly powerful instantiation, by a female author, of the workings of unequal gender relations and the asymmetries of power that are at the heart of sexual liaisons.[7]

Scholars have focused in particular on women's suffering in the polygamous world of the text, arguing that many of the female protagonists in the tale were victims of the oppressive conditions within which they lived. At the same time, these scholars have attempted to highlight the ways in which some of the female protagonists of the tale tried to resist male domination, or at the very least, expressed the desire to be liberated from it. It is in this context that the concept of 'agency' has acquired critical purchase, imbuing the term with a moral, emotional, and political charge.

As this far from comprehensive survey of recent academic literature suggests, the question of gender has now become central to medieval Japanese studies, and invaluable work has emerged that demonstrates that gender was constitutive to the making of the economic, social and cultural life of that age and that it significantly shaped its textual tradition. This is therefore a propitious moment for a critical reappraisal of current approaches to the study of gender in the medieval Japanese literary/Buddhist texts, which for the most part, have not engaged with recent debates within feminism itself, and have therefore failed to challenge our commonly held understandings of gender, and of categories such as woman, body and agency which are closely associated with it.

Greater attentiveness to the cultural and historical variability of these concepts is required if we are to avoid an unreflexive transposition of contemporary understandings of these terms to the Japanese past—an approach that has produced anachronistic readings of the texts we seek to analyse. It is by focusing both on questions of historicity, and on the problem of treating our core categories of analysis as foundational truths, that I hope to contribute to recent debates on gender in the medieval Japanese textual tradition.

The idea of the body as self-evident, a given with fixed meanings, is no longer tenable for, like childhood, death, madness, sex and so on, the body is now seen as having a history. Scholars coming from a wide range of perspectives have sought to demonstrate that even within the West, bodies are contingent and changing formations that are historically variable. Michel Foucault, for example, has challenged the claim that the body 'obeys the exclusive laws of physiology and that it escapes the influence of history'. Through a reading of the multiplicity of regimes to which the body is subjected, he argues that 'nothing in man—not even his body—is sufficiently stable to serve as the basis for self-recognition or for understanding other men'.[8] Historians such as Carolyn Bynum have questioned the commonly held assumption that western thought has been consistently dualistic, that the body has always been constituted in opposition to the soul/spirit/mind, and that as the unprivileged term in the binary, it has been subjected to systematic denigration or neglect. She argues that far from despising the body and wishing to escape from it, medieval Christian theologians in fact 'assumed the flesh to be the instrument of salvation'.[9]

Historicising the body has come from many quarters and even within the western tradition we can no longer assume that the significations accorded to the body have been constant and unchanging. For it was only from the eighteenth century in Europe that both the body and nature came to be seen as passive and inert entities, disconnected from the cosmos and divorced from the soul, and the mind came to be privileged as the sole repository of thought and mental processes. This conception of the body to which we are heirs was produced within the specific context of the post-Renaissance West, and hence cannot serve as the ur form from which all other bodies are seen either to derive or deviate.

In the case of medieval Japan we are doubly removed from the particular debates and questions that circulated around the body and its significations in the West, and hence there is no warrant for assuming that conceptions of a body that emerged within a specific European tradition are amenable to being transposed seamlessly to the literary/Buddhist textual tradition of medieval Japan. It is only by being attentive to the larger epistemic framework of what we might loosely call the East Asian medical, religious and philosophical traditions that we can properly discern what meanings bodies were accorded in medieval Japanese texts. Below, I consider some of the central elements that characterise conceptions of the body in medieval Japanese narratives, to demonstrate the historically and culturally contingent nature of the ways in which bodies are inhabited, imagined and visualised.

First, it is worth noting that the dualism that was at the heart of the mind/body debates in the Western philosophical tradition finds no counterpart in Daoist or Buddhist thought. In the East Asian tradition, the assumption was was that the body and mind were integrally connected, and the central question became one of working out how the two could function most effectively together as a mind-body complex. Mental and affective processes were mutually intertwined, and the body, far from being defined as pure materiality, was seen rather as a psychosomatic process, 'something done, rather than something one has'.[10] In medieval Japanese texts, 'thought' did not function as the other of 'feeling' or emotion—the verb omou encapsulated both feeling and thinking.[11] Both material and mental/emotional processes were integrally linked and central to the constitution of a meaningful body/self. The word mi (身) in the Japanese medieval lexicon that corresponds to the term 'body' makes no distinction between the physical body and what we might call the psychic, social, or cultural body, and hence one of the most common usages of the term mi is to signify a person's status or standing in the world.

In a work such as the Tale of Genji for example, love and desire—central themes for a romance narrative—are not generated through descriptions of the lines or curves of the body. Both the physical and psychic attributes that go into the making of the body find expression in the robes within which the body is enveloped. Robes, which are metonymically linked to the body, are not mere embellishments that adorn, cover and enhance the beauty of the body: they are part and parcel of embodied being, and it is the two together as an ensemble that have the power to generate erotic and affective desire in the Tale of Genji.[13] The ultimate raison d'ȇtre of clothes in the game of erotics in Japanese literary and visual texts does not lie in unveiling the clothed body as in a striptease, for the body without clothes signifies abjection.

This is because in the medieval Japanese worldview, nature and culture were not two distinct entities—a meaningful self was always understood as one that was constituted 'socially'. To put it in another way, there was nothing outside of what we would call the cultural body that existed prior to the social meanings with which it was inscribed. Thus, bodies writhing in pain as they are tortured in hell, or corpses that have been killed in war are depicted in medieval picture scrolls without their clothes, signalling that these bodies/selves have lost their status as social/cultural entities. Likewise, when men remove or lose their courtly headgear, kamuri or eboshi, this signals exposure and humiliation. When a man loses his hat this arouses derisory laughter from those who see or hear about it. Far from being simply a culturally prescribed form of adornment, headgear, like robes, was seen in medieval Japanese visual and literary culture as part and parcel of male embodied being: ideally it was never to be removed, even when a man was alone.[14]

Third, neither the body nor nature in the medieval context was something set in stone, seen as inert and passive matter with immutable attributes. Medieval bodies were granted transformative powers that rendered the boundaries between gods, humans, men, women and beasts porous and fluid. All bodies, even those of women, were conceptualised as active agents that could defy common expectations and perform miraculous transformations, thereby attesting to the power of the Buddhist faith.

In the Devadatta chapter of the Lotus Sutra for example, Prajnâkutâ is sceptical of Mañjusrî's claim that his disciple the eight-year-old dragon princess has, at such a young age, attained perfect Enlightenment. The elder Sāriputra, likewise, expresses doubts about the dragon princess possessing the necessary requisites for attaining Buddhahood on the grounds that the female body is a 'filthy' thing, subject to the five obstructions. The dragon princess does not engage in dialogue with them, but acts swiftly, transforming herself into a man and achieving Buddhahood.[15] What is striking for our purposes is that the body was conceived of as a malleable and changeable entity, making the boundaries between men, women, dragons and Buddhas fluid and open. The transformation of the dragon girl served, within the context of a Buddhist canonical text such as the Lotus Sutra, as a powerful instantiation of the temporary and provisional nature of all that seems real in the mundane world of samsāra.

As a conceptual category, widely deployed in feminist writing, the term gender, from the outset, served a political function. It was a response to the biological determinism that was at the heart of the claim, made since the eighteenth century, that sexual difference was something inscribed on the body, a fact of nature, that could not be changed, and the reason why women were innately inferior to men. The category 'gender' emerged precisely as a way of combating this logic, by arguing that social roles were not necessarily bound to sex, and that sexual or biological difference did not determine intellectual and other differences.

It has now become part of our common sense to assume that what distinguishes men from women is sexual difference and that this difference is biologically determined. This distinction is founded on the idea that there are two distinct domains—nature and culture—and that the body and sex are aligned with the former, whilst gender, a social construction, belongs to the domain of culture. Scholars working on pre-modern Europe have challenged the assumption that the binaries of sex versus gender, and nature versus culture were universal and ahistorical. Carolyn Bynum, for example, has argued that gender imagery in medieval Europe was marked by an extraordinary degree of fluidity.[16]

As in medieval Europe, so too in pre-modern China and Japan, sex and gender, which are premised upon a division between natural attributes and social roles, had little meaning given that social relations were seen as mirroring the same principles that applied to the body and to the cosmos. Gender did not function here as a 'social' category posited against the 'natural' fact of sex. When medieval texts speak of onna no mi they mean more than the physical and sexed body that makes for womanhood; for both her mental and emotional attributes as well as her relationship to others as a social being are involved in the constitution of what we might call the female body/self. The terms onna (女) and otoko (男) in medieval Japanese cannot be conflated with the modern words for woman/women, josei (女性) and man/men, dansei (男性). As the character sei or sex (性) demonstrates, this new understanding of 'man' and 'woman' as constituted through their bodies and in particular their sexual organs signalled an epistemic shift brought about by the adoption of the newly imported western discourses of sexuality to Japan. Not unlike the world of medieval Europe, so too in medieval Japan, women were undoubtedly the incomplete, lesser, inferior versions of men, but the body and sexual difference were not the grounds for affirming this preordained hierarchy, and there was no distinction made between 'natural sex' and 'social gender'.[17]

The feminist project of the sixties and seventies in the West worked with the assumption that women were constituted through their sexual attributes, that they formed a self-evident grouping, and that they were subjected to systematic neglect or vilification on the basis of their sex. This was the impetus for engaging in the project to 'retrieve' women, who were seen as having been silenced and written out of official historical, religious and philosophical discourses. This has been the dominant approach taken by scholars working in the area of medieval Japan.

While the project of retrieval has been invaluable, it is also fraught with problems for it assumes that 'Woman' or even the more capacious, lower case, pluralising amendment 'women', is a stable and pre-given transcendental category already constituted as such—a view that feminists since the 80s have increasingly questioned.[18] The destabilisation of 'women' as a fixed category has consequences for understanding what we mean by gender. The Japanese 'medieval imaginary', as I have argued, was one in which man and woman were not defined primarily through their sexual organs, and their world was not organised through the distinctions of nature/culture and sex/gender. This may explain why Judith Butler's work, which challenges these binaries can be fruitfully pressed into service for an analysis of gender in medieval Japanese texts.

Judith Butler argues that gender 'ought not to be conceived as a noun or a substantial thing or a static cultural marker', but rather as 'an activity, a becoming.'[19] Gender here is performative, and the binaries of sex/gender, man/woman and hetero/homosexuality, are not facts or truths about the world that are made manifest through our actions. Rather it is the other way round: it is through the endless repetition of certain acts that we create the illusion of the stability of gender. In pre-modern Japanese literary texts gender functions precisely as a kind of script, and it is the specificity of the gendered performance, that is to say, the particularity of the script that is enacted that gives substance to the categories 'male' and 'female'. Indisputably, gender difference was central to the hierarchical ordering of both the cosmic and social order of medieval Japan, and women within it were without question positioned as inferior to men. However, this difference was constituted not through 'man' and 'woman' understood as fixed and stable categories, defined through their bodies. Gendering in the medieval Japanese textual tradition, as I hope to demonstrate, was produced and stabilised primarily through stylised performative modes and stances.

Gender difference in a work such as The Tale of Genji does not register overtly through the body, clothes and ideals of beauty; this does not mean that men and women in the text are indistinguishable one from the other, and that their relationships are unmarked by the play of power. How gender comes to be coded, however, depends on the context in which it is performed. The shifting contexts within which amorous encounters take place are always imbricated in a variety of asymmetries and hierarchies, which sometimes work together, and at other times pull in different directions, thereby attesting to the inadequacy of treating either gender or class in isolation, outside of the setting or stage (bamen) upon which they are brought into play. It is the possibilities offered by different ways of 'doing' gender that the text explores through the many romantic encounters that are central to the tale.

Let me illustrate this point with an example of the way in which robes, which often serve to generate or dampen amorous desire in the Genji, also serve to reveal in the process, the complex interplay of gender and status. In descriptions of Genji's clothing, we often find the presence of terms that convey a sense of casualness, careless ease, languor and informality that characterises his attire.[20] Indeed, a conspicuous feature of the text is that it is Genji's appearance, not in formal garb, but rather in a state of deshabillé that arouses amorous desire in those who view him. This is in contrast to the way in which Genji assesses Utsusemi's step-daughter Nokiba no Ogi's appearance when he happens to spy upon her. He notes that she is attired in thin white robes covered with a gown worn carelessly, exposing her breasts all the way down to her scarlet trouser cord. While he cannot deny that she cuts an attractive figure, he is unable to take her seriously because her casual appearance strikes him as slovenly and coarse (hōzoku nari). Genji's judgment of Nokiba no Ogi could no doubt be readily explained by the fact that a different set of protocols applies to men as opposed to women's appearance. In other words, the foundational difference of gender can be made to account for the very different impressions created by the same modes of bodily attire and comportment. However, what is at issue for Genji here is that by allowing herself to be seen in such a careless get-up, Nokiba no Ogi has revealed her lack of class and refinement, and her inability to follow the example of her step-mother Utsusemi, who keeps herself carefully hidden from view. Undoubtedly, both gender and class are implicated here, but gender alone is an inadequate guide for reading the asymmetries that are constitutive of all relationships in the text.

For it is perfectly possible, in other contexts, for a noblewoman to evoke feelings of tenderness and desire even if she is dressed, like Genji, in informal garb. This is the case in the scene where Kaoru is treated unexpectedly to the pleasure of seeing the First Princess, a lady of an exalted rank, in thin summer robes. Kaoru's intense attraction to the First Princess is tied not only to her own noble lineage but also to his intense awareness of his own relatively inferior status. Within the dense social world of the Genji, gender, status, age and a highly aestheticised form of erotics often crisscross, creating a multitude of possibilities.

Gender is performative and not fixed and given, and thus how it is performed 'what constitutes being a woman' is itself shaped by class, which again, far from being stable, functions as a dynamic and fluid category. The figure of the serving woman, for example, is authorised to alter the prescribed performance of gender that her mistress is expected to follow. Her responses point to the possibility of another mode of engagement in which the stylisations so central to amorous play can take a different form, opening up a space for the exchange of suggestive poems, lively banter, and the possibility of striking up sexual relations with men freely—often even with the noblemen who court her own mistress. Her performance of gender, so markedly different from that of her mistress, points to the inadequacy of treating 'woman' as a stable category whose meanings are fixed regardless of differences of status and class.[21]

The performative aspects of gender are most clearly discernible in waka poetry where gender is disassociated from the body and sex, and is principally a matter of certain prescribed stylisations of performative roles. Waka poetry's dominant themes are nature and love, both of which are expressed through a prescribed repertoire of images and vocabulary. The central figures of love poems are otoko (man) and onna (woman), who appear in the poems through terms such as kimi (you) or hito (male or female lover) or ware (I). These terms are used for both men and women alike. They indicate nothing about the gender, identity, or social status of either the poet or the one who is being addressed.

When a poem is described as being a woman's poem (onna no uta), what is at issue is not the sexual or personal identity of the composer of the poem, but rather the particular stylised role or persona to be adopted by the poet that is consonant with woman, not as a real, living being, but rather as a trope or an idea. Even when a poem is marked as anonymous, or when there is no headnote explaining the circumstances under which it was composed, it is possible to infer which persona a poet has adopted. This is because in the waka tradition, woman is always positioned as the one who waits and pines for her male lover, while man is the one who visits at night, and departs before dawn. He is the one who initiates the affair, and composes the morning after poem to which his lady is expected to respond.

A poet, regardless of his/her biological sex (a category that has no real meaning in this context) can slip seamlessly into the persona of the waiting female or the male who visits. This is what makes it possible for a male poet to become the waiting female, and allows even a monk who has ostensibly renounced the world to enact the role of a woman pining for her lover. To suggest that gender is fluid and a matter of performance in waka poetry does not mean that how gender is performed is arbitrary or a matter of individual choice. Gender for the most part is the performance of a script that is given, rather than something of one's own making, 'a set of repeated acts within a highly rigid regulatory frame'.[22] However, what is significant for our purposes is that gender in the medieval Japanese tradition was disassociated from the body and sex. Gender, in this context, worked as a kind of script, and it is the specificity of the gendered performance, that is to say, the particularity of the script that was enacted, that gave substance to the categories 'male' and 'female'.

I have argued, that the body in medieval Japanese texts was not fixed and immutable, and that, as a consequence, the distinctions between man and woman were not predetermined by their respective sexual characteristics. All bodies, even those of women, were conceptualised as active agents that could defy common expectations and miraculously transform themselves—their shape-shifting forms defied any consolidation of 'woman' as a stable entity.

The stories in collections of popular tales (setsuwa) abound with humans, gods, bodhisattvas, buddhas and beasts who intermingle and change forms. This is a world in which women turn into foxes, men, women and beasts reveal themselves to be manifestations of the gods and buddhas, and snakes copulate with humans. These transformations cannot be dismissed as flights of fancy or manifestations of irrational superstition. Nor can they be translated into a secular idiom where they become simply metaphors for human fears and anxieties.[23] For to do so is to disregard the particular cosmology that shapes the medieval Japanese imaginary.

We can only make sense of this mutual imbrication of different realms of existence if we are attentive to the Buddhist epistemic framework within which this particular way of ordering, knowing and inhabiting the world came to be articulated. The doctrinal basis for these ideas came from different sources, among them the influential Tendai sect's formulation of the concept of the ten worlds concomitant (jikkai gogu). In this view, each of the ten realms was seen as mutually interpenetrating and encompassing such that no ontological distinction was possible between, say, the world of hungry ghosts, the world of demons, and the world of ghosts, beasts, humans, gods and buddhas. In this schema what was presupposed was that 'the nine realms of unenlightened beings possess the Buddha nature inherently, while the Buddha possesses the nine realms of unenlightened beings'.[24]

Regardless of whether or not the concept of jikkai gogu was widely known or accepted in Japan, there is no doubt that it permeated medieval Japanese texts, which thematised both the pleasures and dangers of living in a world in which humans and other beings shared a common space, and marvelled at the inexorable forces of karma that could work in unexpected ways, bringing human beings into contact with both bodhisattvas and demons.

In the twelfth-century picture scrolls of hungry ghosts (gaki zōshi),[25] for example, gaki are depicted as grotesque, skeletal figures with enormous bellies and needle-thin throats. They are driven to eating excrement and carrion, and are tortured by demons and vultures. But the realm they inhabit is imagined as being an integral part of the human world. Quietly consuming excrement while people defecate, begging food from monks, mingling with the crowds in a busy marketplace, licking water that is offered to the statue of the Buddha, they are inextricably woven into the fabric of medieval society and are not dissimilar to the starving beggars who once populated medieval towns and cities — at once pariahs, and at the same time integral parts of the community as a whole.[26]

What I have argued above has implications for how we use the term agency in our analyses of medieval Japanese texts. Agency, as I suggested earlier, has been frequently deployed to speak both of women's oppression and their capacity to resist male domination. The moral charge of celebrating agency often takes the form of treating it as conceptually interchangeable with the idea of resistance against relations of power and domination.

Ascribing agency to women, however, has been no easy task, for there is little consensus on how one might gauge the significance of representations of women's activities in medieval texts. Sometimes particular actions are read as instances of women heroically mobilising their agency by acting in ways that challenge the attempts by Buddhism and patriarchy to degrade them as women; at other times, possessed of false consciousness, they are seen as being complicit with Buddhism's ideological agenda, or simply passive victims of it.

Bernard Faure, for example, recognises the problems of treating 'woman' as a unified category and of treating gender as the only prism through which to read medieval texts, and yet the question of women's victimhood or agency looms large in his work. He claims that

The analyses of the workings of gender and agency in the Genji have been more sophisticated, but they rest nonetheless on the assumption that the work is the expression of a consistent voice, that of a female author, who through the plot device of romance, speaks as a woman, seeking primarily to highlight both women's suffering and oppression, as well as their struggle to resist male authority and desire. Women's agency here is understood primarily through the framework of subjection and resistance. To note just a couple of examples, Norma Field observes that many of the female protagonists of the tale are victims who are marked by 'a profound passivity'.[29] By placing their thoughts, feelings and forms of self-expression at the heart of her book, Field gives her 'heroines' a new voice, an agency, articulated through a specifically gendered reading of their predicament. Through a psychoanalytic reading of spirit possession, Doris Bargen argues for the agency of the possessed woman who, she claims, is not a passive victim;[30] on the contrary, spirit possession, in her view, 'was yet another way to criticize and perhaps to lessen male domination'.[31]

There are a number of shared assumptions in these different readings, which warrant examination. First, agency here is implicitly understood to signify the capacity for action that inheres to all humans, defined as autonomous individuals with free will, whose natural inclination is to strive to resist the oppressive conditions of their lives regardless of temporal and cultural differences.[32] This conception of agency is based on a modern humanist idea born of liberal thought, which presupposes the supremacy of Man who replaces the gods as the maker of meaning in the world. Agency in this understanding is something possessed by humans alone. And yet this anthropocentric view of agency would have been unrecognisable to those who inhabited the world of medieval Japan, where humans were not the sole actors and makers of meaning: gods, beasts, demons and even dreams and material objects, worked together with humans as active agents in a shared cosmological and worldly order.

Medieval texts consistently fail to attribute the events that take place in the world solely to human intentions and will; rather they present them as effects unfolding as a consequence of a concatenation of forces, in which a significant role is assigned to the power of the divine and to karma from past lives. In the Tale of Genji, for example, the unlikely union of Genji with the Akashi Lady when Genji is in exile in Suma comes about not as a result of the personal agency of either of the protagonists but due to the intervention of a number of supernatural agents. The amorous entanglements in which men and women find themselves embroiled are viewed not through the prism of coercion or consent, but rather through the Buddhist notion of sukuse, fate or karma. These are also recurring themes in setsuwa collections such as Konjaku monogatari.[33]

Second, the moral charge of celebrating agency often takes the form of treating it as conceptually interchangeable with the idea of resistance against relations of power and domination. This means that acts, particularly religious ones, that work in consonance with social conventions rather than against them cannot be granted real agency. Scholars have singled out nunhood as one of the sites upon which women's response to Buddhism's 'misogyny' came to be played out in the medieval period. Some have seen the act of tonsure as an act of resistance to unequal social arrangements: nunhood, in this reading, becomes the space of freedom that a woman actively chooses.[34] Others, working within the same conceptual framework of agency, have claimed precisely the opposite, arguing that the practice of tonsure was 'a form of death in life'.[35]

Piety is usually given little credence in our readings of such narratives. Scholarly attention, for example, is directed to the fact that despite Murasaki's repeated wish to take the tonsure, Genji refuses to let her do so. Murasaki's reason for becoming a nun is emptied of religious content and seen solely as a way for her to deal with her husband's infidelity. Her decision to leave worldly life is construed either in very negative terms as 'death in life' or as a heroic rebellion against a selfish and self-serving husband.[36]

However, both the taking of the tonsure and the inability to do so carry multiple significations in the Genji, and neither is reducible to being seen solely through the prisms of gender and agency, understood in terms of a binary framework of domination and subordination. To do so produces anachronistic readings, which turn the Genji and other texts within the medieval repertoire into secular works where piety simply becomes a displacement or metaphor that obscures (when read through the lens of 'gender', 'agency' and 'resistance') the 'truth' of the inequality and injustice of gender relations.

The starting point for our study of medieval Japanese texts must involve historicising 'Man', 'Woman', 'Body', 'Gender' and 'Agency'—the analytic categories that are central to our readings of women in the literary and Buddhist texts of medieval Japan and recognising that these concepts do not travel seamlessly across different temporalities and cultural traditions. If personhood in medieval Japan is located in the social, and if it is not imagined as an individual and secular identity, then agency needs to be disentangled from nineteenth-century liberalism, which speaks an altogether different language of choice and self-determination.

To conclude, undoubtedly gender difference was central to the textual tradition of medieval Japan, and relations between women and men were conceived of in terms of a hierarchy, a principle that shaped both the social and cosmic order of the medieval world. However, gender in medieval Japanese texts was not associated with the sexual attributes of the body. It was aligned not with the biological but rather with the performative aspects of the body. The performance of gender was by no means uniform, and it produced a multiplicity of ways of being 'woman' or 'man'. This was in large measure because the body was granted immeasurable powers of transformation, which foreclosed the possibility of the stabilisation of 'man' and 'woman' as fixed and unchanging categories. The potential of the body to mutate and change made gender a materialisation that was never fixed or complete: it was in the nature of a verb rather than a noun.