The primacy of narrative agency: Re-reading Seyla Benhabib on narrativity

Narrative agency and narrative identity

Before addressing these implications, it is necessary to give a clear definition of narrative agency. Narrative agency is distinct from narrative identity. The first is the capacity to say ‘I’ over time and with relation to others. This is not the ‘I’ of pure reason or the doer behind the deed. It is, rather, a basic capacity. The second refers to the shifting constellation of narratives which make up a particular self. Conceiving of narrative agency as a precondition for narrative identity is an important theoretical move because it allows us to account for the endless permutations of narrative identity – its contingency, its complexity – without having to articulate a specific, fixed narrative identity. In other words, it allows us to take into account the – potentially devastating – effects of subordinating power relationships on identity formation without accepting the notion that a subject is entirely subordinated by those relationships.

This configuration suggests that a notion of autonomy, understood as reflexive critical capacity, only makes sense after the narrative agent has been identified. ⁴ We cannot presume to know the extent to which a subject’s critical capacity is limited by the relations of power into which she is born. Critical capacity ultimately depends upon the individual’s conception of herself as a coherent yet mutable formation over time within a specific socio-historical context and upon her situation within a community which recognises her. The subject cannot transcend power relationships, social context or time. Nor can we say that a subject consistently exercises autonomy in the same way or to the same extent over the course of her lifetime. But we can change power structures from within through the collective interplay of narratives: we can recognise, rearrange and reframe norms through the collective creation of meaning.

In her early work, Benhabib outlines a concept of narrative agency within a post-metaphysical critical theoretical framework, one in which normative criteria may be posited only when the validity of those normative criteria are left ever-open to interpretation and debate (Benhabib, 1986; Benhabib, 1992; see also: Habermas, 1987). For Benhabib, this contextualised version of critical theory – at which she arrives through a rigorous deconstruction and reframing of Habermasian communicative ethics – is preferable to other epistemological methods because it can furnish (provisional) norms capable of influencing various disciplines, thereby making the most immediate difference in terms of promoting mutual recognition and social justice, even as it admits that these are context-dependent goals, based on the concerns of a particular episteme. ⁵ A normative restructuring that seems emancipatory today might just as easily seem repressive or exclusionary tomorrow. In other words, the theorist can only work within the intellectual parameters already set for contemporary critique, even as she reflects critically upon the formation of those parameters.

Within the framework of this contextualised critique, I want to identify three related claims about narrative agency. First, narrative agency is primary: it precedes narrative content (narrative identity) in the sense that the former is both necessary for the creation and implies the continued existence of the latter. The subject’s capacity to say ‘I am a woman’ depends on her capacity to say ‘I’; that is, a subject’s capacity to makes sense of herself as a unique being is a precondition for identity formation. ⁶ Second, narrative agency is relational: I can only meaningfully distinguish myself as an ‘I’ in relation to others, and these others are a) also narrative agents who are unique sites of meaning creation and b) conversation partners who determine the kind of meaning I will make about myself. Third, and finally, narrative agency is generative. An agent’s uniqueness allows her to create meaning from a point of view that is hers and hers alone. New norms (which are articulated and reproduced through narratives) arise out of the exercise of narrative agency by each individual member of a plurality. These collectively constructed norms are immanent, changeable and they may be harmful as often as they are emancipatory.

The first claim, that narrative agency is primary, is perhaps the easiest to unpack. Each individual has the capacity to construct meaningful narratives about who she is with relation to others. Narrative agency, conceived as such, does not posit a static subject who possesses, over a lifetime, an unchanging or completely articulable identity. On the contrary, the subject is not an a priori formation, wholly definable, but an irreducible site of meaning creation. An individual will make sense of her own identity by exercising her narrative agency, but she can never give a ‘complete’ account of herself because her identity depends upon the changing narratives of others in and over time. Indeed, narrative agency does not assume a fixed ontological subject who pre-exists discourse. Narrative agency does not imply any other intention or motivation than to ‘make sense’, so the narrative agent need not have any moral commitments or well-informed opinions. Narrative as capacity, the capacity to make meaningful one’s situation within a web of other narratives, is a precondition for any narrative content.

Creating meaningful narratives, then, is always a relational activity, carried out within a group of individuals, each of whom is unique by virtue of a particular, unrepeatable perspective. Each individual is ‘the same, that is, human’, as Hannah Arendt writes in The Human Condition, ‘in such a way that nobody is ever the same as anyone else who ever lived, lives, or will live’ (1958: 8). ⁷ Benhabib’s concept of narrativity is built upon this Arendtian notion of plurality, in which the self is never ‘withdrawn from the world’ but is always a ‘self in the human community, an acting [and] interacting self’ (Benhabib, 1992: 127). Within the human community are formed protean, context-dependent relationships which constitute the parameters of the self in the same way that collectively agreed upon norms establish the parameters of critique. Both norms and narratives are arrived at collectively, through social interaction; by the same token, these norms and narratives can only be confronted, resisted and changed through interaction.

Narrative agency refers to the subject’s capacity to construct a meaningful narrative, and not to the actual content of that narrative. The narrative agent need not subscribe to a particular morality or be governed by a particular set of normative criteria. Benhabib makes this point in contrast with Charles Taylor (1989). For Taylor, a self is partly definable by her evaluative or moral commitments. Identity, as he puts it, ‘is defined by the commitments and identifications which provide the frame or horizon within which I can try to determine from case to case what is good, or valuable, or what ought to be done, or what I endorse or oppose’ (1989: 27). The subject possesses a sense of agency, in this view, by knowing where she stands with regard to certain evaluative commitments. Furthermore, for Taylor, these evaluative commitments are only meaningful within the context of ‘webs of interlocution’. That is, identity depends not only upon what one stands for but also upon the person to whom or with whom one is speaking (Taylor, 1989: 36).

For Benhabib, these webs of interlocution are indeed the setting for identity formation. At birth, we are thrown into a network of other people’s narratives, and we ‘become who we are by learning to be a conversation partner in these narratives’ (Benhabib, 1999: 344). While this process of collaborative individuation might entail making lasting evaluative commitments, it is entirely possible to have agency (in the form of the capacity to construct meaningful narratives) without having a set of values or an innate moral integrity. Taylor’s tendency to conflate agency with moral commitment is, Benhabib argues, a confusion of levels of analysis, and one which must be avoided. It is important to ‘think of the continuity of the self in time not through a commitment to a specific set of evaluative goods but through the capacity to take and adopt an attitude toward those goods’ (Benhabib, 1999: 364; emphasis mine). In other words, the subject’s capacity for agency must not be confused with assumptions about what a self ought to think or do – narrative capacity must not be conflated with narrative content. As Benhabib puts it: ‘it is not what the story is about that matters but, rather, one’s ability to keep telling a story about who one is that makes sense to oneself and others’ (1999: 347).

The narrative agent, conceived as conversation partner within a web of interlocution, in no way represents a fixed or static core self: she is changeable and context-dependent, but also irrepressible – for the narrative agent will always try to make sense, even out of nonsense, no matter how varied, fragmented or overwhelming such nonsense may be. The subject’s capacity to make sense is the constant which allows for a coherent sense of self to develop over time. Moreover, if the process of making sense of the self over time is conceived as inherently relational, then the subject is never solely responsible for her own story. Indeed, she must always fit her story in with and alongside the stories of others. ‘Narratives’, Benhabib writes, ‘cannot have closure precisely because they are always aspects of the narratives of others; the sense that I create for myself is always immersed in a fragile “web of stories” that I as well as others spin’ (1999: 347). ‘Making sense’ paves the way for mutual recognition: I include you in my story as you include me in yours. Parts of your story resonate with me, as parts of my story resonate with you.

It is precisely this intersubjective give-and-take which lends narrative agency its heuristic strength, for: ‘[f]urthering one’s capacity for autonomous agency is only possible within a solidaristic community that sustains one’s identity through listening to one, and allowing one to listen to others, with respect’ (Benhabib, 1999: 350). For Benhabib, narrative agency and communicability are ‘two sides of the same coin’ (2007: 15). Stories are produced culturally, as well as individually, and individual stories have no meaning unless they are shared:

Only if somebody else is able to understand the meaning of our words as well as the whatness of our deeds can the identity of the self be said to be revealed. Action and speech, therefore, are essentially interaction. They take place between humans. Narrativity, or the immersion of action in a web of human relationships, is the mode through which the self is individuated and acts are identified. (Benhabib, 1992: 127)

Conceptualising the creation of meaning through communicative action represents an important point of tension between Benhabib’s narrative model of agency and other prominent feminist definitions of agency, especially those inspired by Judith Butler’s much-cited performative model (1993, 1997a). For Butler, speech is an action that reiterates a norm or set of norms. Reiteration and subsequent resignification, in this model, constitute a linguistic transformation, a ‘speech act’. The result of this transformation is a new linguistic configuration, one which may subversively rearrange the terms of discourse in such a way that the subject is able to express her agency. Benhabib argues that this conceptualisation of language in action does not go far enough because it cannot account for the ‘surfeit of meaning, creativity, and spontaneity’ (1999: 341) that arises out of communicative interaction. As she puts it, these ‘speech acts are not only iterations but also innovations and reinterpretations’ (339). In other words, resignification and reinterpretation create some new meaning apart from the norms or sets of norms that they reproduce.

The surfeit of meaning in language arises out of the mutual recognition (between unique and unrepeatable individual perspectives) involved in the sharing of narratives. Breaking apart and reconfiguring norms of discourse, as the performative model does, cannot fully explain the experience of participating in or appreciating these performances which create moments of humour, irony, pathos, etc. This creation of shared meaning cannot be explained by performance alone. ‘The narrative model’, as Benhabib puts it, ‘has the virtue of accounting for that “surfeit of meaning, creativity and spontaneity” that is said to accompany iteration in the performativity model as well but whose mechanisms cannot actually be explained by performativity’ (341). The meaning generated by the sharing of narratives, in other words, amounts to more than the sum of its parts. An individual, by this account of narrative agency, is always able to tell some kind of story about how she is situated in the world into which she is thrown. Even if the content of that story is rife with contradiction or belies a commitment to subordinating attachments, the capacity to tell a story about oneself remains, and with it the capacity to change that story as circumstances, beliefs and normative commitments change.

The gender objection

As mentioned, much of the literature on agency in feminism protests that a narrative model remains overly indebted to a rationalist conception of autonomy, one which presupposes the possibility of critical self-awareness from a place beyond the confines of power relations. ⁸ Amy Allen (2008b) gives one such critique. She argues that Benhabib’s conception of narrativity retains an implicit reliance upon a subject with robust autonomy in the form of critical reflexivity. According to Allen, Benhabib posits the existence of an ‘ungendered core self’ capable of ‘making sense’ of subordinating gender relationships before becoming gendered. To choose how to relate a narrative about gender, a self must first have the autobiographical capacity necessary to ‘make sense’ at all, but an individual does not gain such a capacity, Allen reasons, until that individual is already gendered. Recall, however, that narrative agency, conceived as an individual’s capacity to make sense out of nonsense, does not need to posit the extent to which an individual can gain reflective distance from her situation within society. An individual will make sense of the gender norms into which she is thrown, but the way she will make sense of those norms can be neither predicted nor guaranteed. As we have seen, the individual’s capacity to form a narrative in the first place constitutes her agency, regardless of the content of that narrative.

Allen’s broader goal is to find a framework for feminist subjectivity which allows for the possibilities of both agency, understood as the capacity to act in the world, and autonomy, understood as the capacity to reflect critically on action, in the face of power ‘in all its depth and complexity’ (2008b: 2). Thus, her aim in looking at the narrative model of agency is to determine whether it falls victim to an unsupportable rationalism in the face of ubiquitous power relationships. She finds that the narrative model does not take sufficiently into account the severity of the influence of subordinating gender norms on the process of individuation.

Allen argues that the ‘I’ who ‘chooses’ her narratives is always already gendered. But even the gendered ‘I’ has the capacity to make choices, and the content of those choices is secondary to this capacity. As we undergo the process of socialisation, we are introduced to the claims of culture, many of which are uncomfortably subordinating and marked by unequal distributions of power. These claims constitute narrative identity, but, in themselves, they do not account for every aspect of narrative identity. The capacity to form and reform attachments to these claims first depends on the unique and irreducible role of the individual as narrative agent.

It is this differentiation between the capacity to say ‘I’ and the content of the choices that ‘I’ make, however, of which Allen is sceptical. In her view, the capacity to say ‘I’ is preceded and thus at least partially determined by gender identification. As she puts it: ‘If the roots of gender identity lie deeper than those of the narrative ability that Benhabib views as the source of spontaneity, creativity, and agency, then interrelated assumptions about gender difference and gender dominance are so basic to our sense of ourselves that they are likely to be extremely resistant to critique and to change’ (Allen, 2008b: 170). For Allen, Benhabib posits a universal narrative agency at the expense of recognising the gendered constructs which shape even the child’s first exercise of this narrative agency. But does the reality of gender subordination, even at the earliest stages of individuation, compromise our definition of narrative agency as the irrepressible capacity to make sense over time and in relation to others? And does this concept of narrative agency really carry with it a harmfully rationalist endorsement of autonomy vis-a-vis gender norms?

In the model of narrative identity detailed above, gender identification, like all other narrative content, depends upon the capacity to say ‘I’. Narrative agency (capacity) and narrative identity (content) are inextricably bound together. To conceive of narrative agency as primary is not to suggest that the self exists in some fixed state apart from the narrative formation of life stories. The narrative self is the site of making sense, and as such, is fragile and always in flux. Benhabib does not argue for a static individual who may step outside of her position in the world and ask: How ought ‘I’ to confront this or that norm? Narrative agency, as we have seen, is the capacity to make sense of one’s position within a web of narratives and not the capacity to get outside of this web in order to choose whether or not to accept its terms. Furthermore, the narrative agent is not the master of her own narratives. Her access to meaning-making is always mediated not only by external norms and other stories but also by the very early unconscious internalisation of these norms.

Allen refers to several studies which suggest that gender identity forms before a child develops either autobiographical memory or the narrative capacity through which to share information about herself (Stern and Karraker, 1989; Nelson, 1993; Fivush and Golombok, 1994; Buckner and Fivush, 1998; Fivush, 1998; Fivush and Schwarzmueller, 1998). By age two, it seems, children have a firm grasp of the concept of gender difference, but they do not begin to develop the capacity for autobiographical memory until the age of three or four. Autobiographical memory is different from episodic memory in that it gives the subject a sense of self over time, rather than just an ability to recall past events. Because research suggests that autobiographical memory is developed ‘through social interactions with adult caregivers’, and because those adult caregivers tend to ‘interact with infants and young children in ways that correspond to gender stereotypes’, Allen concludes that ‘autobiography is deeply gendered as well’ (2008b: 167). To support this claim, she cites a study which concluded that little boys’ and little girls’ narratives ‘tend to be different in both their content and, perhaps more significantly, their structure’ (167). Researchers found that little boys tended to relate memories that were shorter and more self-involved, whereas little girls tended to give more elaborate and relational accounts of past experiences.

Based on this empirical evidence, Allen claims that individuation, as well as socialisation, is a fundamentally gendered process. She argues that ‘gender structures not only the substantive content of our very narratives but also our very narrative capacities, thus, our narrative selves as well’ (168). According to Allen, Benhabib does not take seriously enough the extent to which gender dominance is at work on this foundational level, and thus gives ‘an overly optimistic account of what is required in order to exercise autonomy with respect to gender narratives’ (164). But this criticism misses the mark for a few important reasons.

First, Allen gives dubious privilege to the systematic dominance of gender relations over the systematic dominance of other relations of power, such as race, class and level of education. Autobiographical understanding is formed chiefly through social interaction (learning to become a conversation partner) with caretakers, but there is nothing to suggest that gender stereotypes are more influential in forming this understanding than these other important factors. ⁹ Many of the experiments to which Allen refers examine the autobiographical narratives of white, middle-class American children (see: Buckner and Fivush, 1998: 407). Surely, the development of autobiographical capacity is different in children of other cultures or other classes; thus, Allen’s singling out of gender over other developmental factors seems to give us only part of the story. The confrontation with norms, whatever these norms may be, is an integral feature of the process of identity formation: the capacity to confront these norms remains constant.

And so, it does not matter all that much whether the ‘I’ who creates narrative meaning is already gendered – what matters is the capacity to make sense of oneself as a unique ‘I’ at all. From birth, the narrative agent interacts with the (often contradictory) claims made by gender norms over the course of a lifetime. These gender norms themselves are fluid and multi-layered, not homogenous. They mean different things to different people at different times and in different cultural contexts. Thus, a subject might very easily be ‘gendered’ in that she is constituted by and embedded in the web of gender norms into which she has been born, but this does not mean that she cannot challenge and transform, with varying degrees of success, the norms with which she is confronted. Knowing that an individual is gendered, even down to her most basic conception of herself, does not lessen the individual’s capacity for narrative agency. Benhabib writes:

We always have options in telling a life story that makes sense to us. These options are not ahistorical; they are culturally and historically specific and inflected by the master narrative of the family structure and gender roles into which each individual is thrown. Nonetheless, just as the grammatical rules of language, once acquired, do not exhaust our capacity to build an infinite number of well-formed sentences in a language, so socialization and accumulation processes do not determine the life story of any unique individual or his or her capacity to initiate new actions and new sentences in conversation. (1999: 345)

And lastly, against Allen, we can say that a notion of gender subordination, even one that is deeply entrenched at the psychic level, does not compromise our concept of narrative agency. Narrative agency does not mean that a self can always identify or isolate, and then choose whether and how to take up as narrative, subordinating gender norms. On the contrary, Benhabib concedes, after Freud, that the ego is not the master of its own house. She takes seriously psychoanalytic insights into the phenomenon of subjection. As she puts it: ‘Every story we tell of ourselves will also contain another of which we may not even be aware; and, in ways that are usually very obscure to us, we are determined by these subtexts and memories in our unconscious’ (349). This unavoidable subjection does not preclude narrative agency. The individual tries to form a coherent identity for herself by relating to others, but she may never fully know what psychic forces at work within herself influence the course of this interaction. ¹¹

The deeply ingrained system of subordinating gender norms into which all humans are born will limit the individual’s capacity for autonomy vis-a-vis these norms; it will not, however, compromise the individual’s narrative agency. The agent will make sense of herself within society whether she accepts or rejects the claims of stereotypical gender norms, when she is able to recognise them at all. The narrative agent is, indeed, constituted by a host of subordinating relationships; and yet, she is not totally determined by them. While these relationships of power, many of them gendered or gender-based relationships, often limit a subject’s ability to recognise power at work in and on herself, they do not completely determine the stories a subject will tell about herself. Narrative agency, therefore, is, very simply, the capacity of a constituted subject to make sense. This narrative model does not propose a core self that transcends the subordinating claims made by systems of gender dominance, and so there is nothing in it that prevents a gendered subject from exercising narrative agency. ‘Gendered’ identity, as it plays out in the open field of choice and circumstance, is as complex and varied as identity itself.

The communicability objection

I argue above that narrative agency is essentially relational in that it arises out of conversations within webs of interlocution. Within a radically contextualised critique, there exist no static or quasi-transcendental rules which govern conversation in action; however, if we take seriously Benhabib’s notion of the creation of shared meaning through shared narratives within a plurality, then we might reasonably conceive of narrativity as a sort of engine for social change. Together, we are able to make and re-make meaningful statements about who we are. The problem then becomes how to identify whether or not such change is for the better: how to identify, in other words, moments of ‘positive’ mutual recognition and to distinguish those moments from harmful patterns of systematic subordination, miscommunication and mis-recognition.

The second major criticism of the narrative model of agency is the concern that privileging the notions of narrative and recognition may lead us to ignore the problematics of subordination in favour of a satisfying, but imagined, narrative coherence. Lois McNay (2003, 2008) makes this argument directly against Benhabib. McNay does admit that a narrative model of agency is more conducive to theorising intersubjectivity and creativity than other models put forth by feminist theorists. She sees much heuristic potential in the narrative model of agency – a subject with the constant capacity for sense-making seems to her pragmatically preferable to the disconnected, serialised and fractured subject of postmodernity because it emphasises ‘the temporal and intersubjective aspects of subjectivity and agency’ (McNay, 2003: 7). However, McNay is sceptical of Benhabib’s work on narrativity for a few reasons. First, she thinks it insupportably relies on the inherent communicability of narratives. Second, she suggests that much of Benhabib’s theory of narrativity is still haunted by the untenably rationalist assumptions of the Habermasian communicative ethics that Benhabib has worked so meticulously to contextualise. And third, she points out that narrative, focused as it is upon relationships between individuals, struggles to address systematic domination on a broader social scale.

McNay argues, then, that narratives are not inherently communicable and, further, that the assumption that ‘making sense’ is an essential human capacity keeps us from appreciating ‘the blocks, both psychic and social, to the formation of a coherent sense of self’ (7). She equates the notion that narratives are inherently communicable to Habermas’ reliance on the context transcendence of certain kinds of communication. Benhabib, according to McNay, tends to disassociate narrative from the factors which might prevent the construction of narratives in the same way that Habermas tends to disassociate communication from inherently subordinating forms of power. In both cases, there is an overly ambitious attempt to theorise identity in terms of unproblematically equal and ‘authentic’ communication. However, where Habermasian communicative ethics depends on the use of rhetorical rather than poetic language – in other words, the acceptance of another person’s point of view on predetermined terms – the process of sharing narratives is anything but rhetorical. The Habermasian project of communicative ethics and the Benhabibian notion of narrativity are thus fundamentally incompatible, according to McNay. Benhabib, by both copying Habermas’ assumption of universal communicability and simultaneously throwing out Habermas’ criteria for validity when it comes to sharing meaning, wants to have her cake and eat it too.

McNay argues that intersubjectivity and interpersonal interaction, within Benhabib’s schema of narrative identity formation, can never function according to universal rules of engagement because the commitments, identifications and perceived truths of the narrative agent are ever shifting. Thus, relationships between the individual and the other, the individual and her self, the individual and certain social norms, may never be definitively judged subordinating or mutually recognitive; and, as a result, identifying situations in which mutual recognition or positive re-signification are certain to occur becomes nearly impossible. Moreover, the process of telling stories is not always helpful; McNay argues that it can also be disruptive, mythologising or reifying. According to McNay, Benhabib is blind to this problem of systematic inequality because she implicitly relies on ‘Habermas’s model of a communicatively symmetrical intersubjectivity’, a reliance which ‘results in the deployment of a syncretic and over-generalized idea of narrative identity’ (2).

For McNay, one result of this implicit methodological reliance is Benhabib’s assumption that all experience can lend itself to narrative. Such an assumption grants a disproportionate authority to stories about the self and promotes an insupportable primacy of the said, a linguistic monism that does not take into account the myriad aspects of selfhood that are difficult, if not impossible, to put into narrative form. In other words, McNay thinks that Benhabib copies the Habermasian ‘assumption of the unproblematic transmissibility of inner nature’ (10). She argues that Benhabib is overly ambitious in her judgment that the self is always capable of constructing a coherent narrative identity.

Indeed, McNay claims that Benhabib posits an irreconcilable difference between the fragmented self and the coherent self and in doing so ‘avoids addressing important issues, such as the nature of the boundary between the sayable and the unsayable, or of the passage of experience from a pre-discursive to a discursive level’ (9). The problem here is not that Benhabib posits a core self; rather, it is that she presupposes an essential relationship between the experiences that constitute a self and the process of making sense of those experiences. Experience, according to McNay, is not so universally translatable. Many experiences, such as the everyday episodic experiences of eating lunch or riding the bus are so mundane that they ‘resist incorporation into a meaningful account of the self’ (9). Conversely, experiences of rape or abuse are often so traumatic that an individual will disassociate from them completely and have no way of rendering them meaningful at all.

Benhabib, by this account, also privileges narrative accounts of lived experience to the extent that she sees them as unquestionably authentic. McNay sees this privileging as a kind of fetishisation of meaning – she refers to Benhabib’s problematic preoccupation with narrative’s ‘normative redemptive force’ (2). Privileging the stories we tell about ourselves is harmful because the ‘[n]arration of identity may involve the reification as much as a clarification of the self’ (11). When we rely on someone else’s memory or someone else’s perception, we often tell ourselves stories about who we are in order to fit in with the agreed upon notion of who we ought to be. The telling and retelling of such codified narratives can ‘create crevasses, ruptures, emptiness and deep wells of non-being’ (10). An individual trying to make sense of a traumatic event, for example, might tell a story in which she identifies herself as victim, where the telling of that story results in a reified, alienating narrative identity of victimhood.

Finally, McNay is concerned with the now-familiar tendency in critical theory to reduce systematically subordinating power relationships to identity politics. Gender issues, especially, are far too often conceived of in terms of identity or recognition rather than as systematically maintained forms of oppression. McNay argues that the narrative model suffers from this oversight because of its emphasis on immediate interpersonal interaction instead of trends of subordination on a broader scale. A narrative subject might be able to identify or confront the subordinating norms with which she comes into direct contact, but, at the same time, be unable to theorise a ‘public’ or ‘impersonal’ undercurrent of oppression. In the context of gender, for example, an individual might, because of her personal experience of, say, excelling as a woman in a male-dominated field, incorporate into her life story a narrative of widespread and growing gender equality while failing to recognise the ways in which gender is still responsible for various (quasi)permanent asymmetries of power within society as a whole. McNay is concerned that the ‘understanding of gender in terms of narrative identity and mutual recognition obscures the systemic levels at which gender inequalities are perpetuated and thereby renders them invisible’ (14). Emphasising the importance of these overarching objective asymmetries of gender subordination is clearly an indispensable task of feminist theory.

McNay’s critique requires that we elaborate upon the relational and generative aspects of narrative agency. First, McNay’s charge that a narrative theory is overly optimistic about the communicability of experience demands a careful definition of what ‘narrative’ is and what it is not. ¹² In her work on narrative agency and practical identity, Catriona Mackenzie argues that narrative need not follow a predetermined structure: narrative, when it applies to self-constitution, is not the same thing as literary narrative. It does not need to follow the arc of beginning-middle-end, does not need to be coherent and is not completely under the agent’s control. Indeed, narrative identity might take any number of forms: ‘patterns of action, attention, or emotional response, in bodily dispositions and habits, in moral commitments, or in one’s personal relationships’ (Mackenzie, 2008: 1). The narratives that emerge in narrative identity formation, as we have seen, are contingent and unpredictable. Their existence does not depend upon a fixed system of communication. A narrative theory allows that experiences which resist conventional narrative structure, such as ‘fragmentation and collage, the senselessness of being next-to-each-other in space in time’, are just as authentic as more straightforward life stories because they express a ‘material and lived reality’ – these experiences are still classifiable as narratives in that they involve exercising the capacity of narrative agency, or ‘making sense’ (Benhabib, 1999: 345). The construction of narrative content is an open-ended process of remembering and retelling within a web of other narratives.

Furthermore, the conversations which constitute this web almost always take place on unequal ground. Certain aspects of selfhood, especially the effects of trauma and abuse, may elude narrative awareness. But we can extend Freud’s metaphor of the ego who is not the master of its own house even further: the self may be imagined as a crowded household in which all of the petty bickering of a ‘family brawl’ (Benhabib, 1999: 349) is apt to be acted out. No relationship, whether intrasubjective or intersubjective, is free from confusion, contrast or argument; and, of course, there is no possibility of extricating oneself from the web of relationships into which one is born. Some conversations are mutually beneficial, but they often fail. Importantly, individuals and collectivities continue to weave identities for themselves out of the web of interlocution into which they are thrown, regardless of how repressive or subordinating relationships within the web may be. McNay reads Benhabib’s insistence on the resilience of narrative agency as a reliance on the inherent communicability of ‘narratives’, but we should, instead, understand the resilience of narrative agency as the certainty that some narratives will be communicable, some of the time.

Within a theory of narrativity there is room for the idea that narrative is repressive and exclusionary, but there is also, simultaneously, room for the idea that narratives can be productively communicable. The process of creating narratives is indeterminate, marked not by the construction of any specific narrative, but by the uniqueness of each narrative agent. Inherent in the formation of each narrative is the potential generation of a new meaning, unique by virtue of its unrepeatable perspective. McNay gestures towards a problematic privileging of ‘meaning’ in this formulation of narrative construction, but a theory of narrativity does not have any intractable attachments to what constitutes ‘sense’. Rather than read a theory of narrative agency as emphasising meaning itself, it is more profitable to read such a theory as interested primarily in the capacity to formulate and reformulate meaning. It is precisely this open-endedness, this unpredictability, that is foundational for the formation of narratives within a plurality. ¹³ The individual child’s construction of narrative content is essentially relational. Her idea of what ‘makes sense’, though unique, is informed by and limited to collectively decided upon and ever-shifting ideas of ‘sense’.

Embedded in this landscape of relational identity and collectively decided upon norms is the individual’s capacity to confront and change these norms through autonomous reflection. Recall Benhabib’s argument that ‘[f]urthering one’s capacity for autonomous agency is only possible within a solidaristic community that sustains one’s identity through listening to one, and allowing one to listen to others, with respect’ (350). An individual’s speech and action have no meaning unless they are recognised by others (and it is in this sense that all action is interaction). The meaning they do have once recognised is fragile and dependent upon a specific constellation of collectively constructed narratives and norms. As circumstances change, so too does the construction of narrative and normative meaning. These changes can occur for the better, under the aegis of respectful mutual recognition.

The reification of certain narratives alluded to by McNay can thus be challenged and changed under the correct conditions. The rape victim who feels she cannot escape the narrative of victimhood always has the potential to confront and reconfigure this narrative within a solidaristic community, however small. The generation of new meaning through such a supportive and mutually respectful conversation (or, more likely, series of conversations) will be extraordinarily difficult. The point, however, is that the potential for transformative mutual recognition is always there, no matter how difficult to achieve. Because narrative agency is an irreducible capacity, narrative identity never refers to a static subject who is the same over time. It always refers to an open and unfinished process full of potential but, as of yet, unrealised meaning.

Feminist politics, to be truly powerful, must recognise the importance of belonging to and being recognised within such a community. The collective capacity to identify and resist oppression first depends on the capacity of discrete, individual agents to communicate meaning to one another. The narrative identities of nations, cultures and other groups are also ‘woven out of tales and fragments belonging both to oneself and to others’ (351). ¹⁴ Collectivities, at their best, may derive a coherent sense of who they are through this dynamic of generative mutual recognition. By these lights, we can pursue a politics that does not define itself according to static or pre-existing identity narratives; in other words, we can endorse a communicative politics that is concerned with the interests (contingently) held in common by complex narrative selves. Such a politics does not insist on permanent categories of oppression, but is defined by, as Marieke Borren puts it, ‘spontaneous emergence, associative action, revolutionary pathos […] the very urgent sense that something new and empowering is happening, and […] a short-lived existence’ (2013: 207). Many different feminist politics seem to share these characteristics. This is not to say that there is no common thread connecting one feminist politics to the next; rather, feminist politics spring up differently for different collectives, depending on which issues or problems are pressing for those collectives. Each politics is feminist because it addresses a ‘particular worldly issue that affects women differently from men’ (Borren, 2013: 207), but each collective deals with or acts out this issue on its own terms and is, importantly, made up of a plurality of unique individuals concerned with the issue in question but whose identities are not fully defined by this issue.

Similarly, the collective identity of the Occupy movement, to take a somewhat dated but particularly pertinent example, was not definable in terms of its mission statement or its demographics; rather, it was a historically situated and contingent confluence of many narratives. Its transformative potential depended upon the collaborative construction of its own identity by discrete, individual actors. Diverse individuals within the movement confronted and resisted deeply entrenched systematic oppression through a series of actual conversations. The collective agency of the Occupy movement depended on these ‘dynamic and open-ended conversations’ rather than on ‘collective identity as a product’ (Kavada, 2015: 875). A plurality of real, unique, individual agents with diverse points of view was the source of the capacity of the movement itself to create new narratives, which may or may not have been sufficient to effect lasting political change. Overturning systematic oppression over a specific period of time, and according to a set of static normative commitments, applicable to everyone everywhere, was not the point. Indeed, the possibility of general consensus among the social movement and everyone else (the state, the banks, those inherently unsympathetic to the movement for any number of reasons) is an unnecessary, and often harmful, illusion. ‘[I]t is less significant’, Benhabib writes, ‘that “we” discover “the” general interest, but more significant that collective decisions be reached through procedures which are radically open and fair to all’ (1992: 9). This view of solidarity – as the mutually recognitive collective pursuit of an interest held in common – seems to me a very productive way of framing confrontations with what McNay refers to as systematically maintained forms of oppression.

Conclusion

Thinking of agency in the way I have suggested here, as the irreducible capacity at the heart of identity formation, allows us to posit an idea of a self who is always capable of change. An agent might incorporate the claims of subordinating gender norms into even her most basic idea of herself, and yet she remains capable, through the creation of new meaning from a unique point of view, of shifting those norms in surprising ways. An agent might harbour the seemingly inarticulate harm caused by deep trauma, and yet she remains capable, through continual attempts at conversation with people who respect her, of changing the extent to which that trauma defines her. Recognising this capacity to make sense of oneself through time and in relation to others as primary does not mean that we should expect people to tell coherent, authoritative stories about who they are; it means, rather, that we should be prepared for unexpected stories to emerge from each individual, no matter who he or she is and no matter what we may think we know about him or her.

Feminist critical theorists can, as Allen suggests, encourage positive self and social transformation by ‘acknowledging recognition as an ethical ideal and understanding it as a permanent—though temporally fleeting—possibility of human relationships’ (2008b: 95). There is no way to transcend the subordinating power structures we, as feminists, seek to change, but ubiquitous subordination does not compromise narrative agency. On the contrary, it is only by positing narrative agency as a constant that we are able to argue that the possibilities of mutual recognition and autonomy are permanent. Further, the identification of this irreducible capacity as a constant does not mean that we should endorse an individualistic feminist politics; on the contrary, the notion that agency is the unique individual’s capacity to make meaning is inescapably relational. The transformative and generative potential of identity (whether collective or individual) first depends upon the capacity of each human being to make sense of herself as an ‘I’ within a web of other ‘I’s.

Mutual recognition and reflexivity are plausible normative aims of an admittedly contextualised feminist critical theory. These aims are inherent in narrative practice – by virtue of narrative agency I understand myself in time and in relation to others, and through this understanding of myself I come to recognise the same narrative agency in others. Equally inherent in narrative identity, however, are the problems of subjection and misrecognition – some narratives will always be reductive, harmful, alienating and misrepresentative. Tracing possible pathways for positive social transformation thus involves attending to moments, however transitory, when new meaning is created within a plurality of unique selves.