Grief, self and narrative

ABSTRACT Various claims have been made concerning the role of narrative in grief. In this paper, we emphasize the need for a discerning approach, which acknowledges that narratives of different kinds relate to grief in different ways. We focus specifically on the positive contributions that narrative can make to sustaining, restoring and revising a sense of who one is. We argue that, although it is right to suggest that narratives provide structure and coherence, they also play a complementary role in disrupting established structure and opening up new possibilities. We add that both of these roles point to the importance of interpersonal, social and cultural factors in shaping the trajectory of grief. We conclude by briefly considering the implications for distinguishing between typical and pathological forms of grief.


Introduction
In this paper, we consider how narrative contributes to grief over the death of a person. 1 This does not involve providing a straightforward account of how one thing, narrative, relates to another thing, grief. The term 'narrative' accommodates considerable diversity, while grief is plausibly a multi-faceted, temporally extended and highly variable process, to which narratives of different kinds might relate in any number of ways. We thus begin by endorsing a wider complaint about narrative-talk voiced by Peter Lamarque (2004), who points out that bold and sweeping claims made on behalf of narrative invariably apply only to certain types of narratives. In contrast, few, if any, interesting claims apply to narratives in general. Lamarque adopts a very thin definition of narrative, according to which narratives are told rather than discovered, include at least two events that are related in a non-logical way, and have a temporal structure. So, even brief statements such as 'Bob passed the ball to Sue' would count as narratives. However, we will show that, even with more restrictive definitions, his point applies to the case of grief. Any claim to the effect that 'narrative does x' must be further qualified if it is to be informative. We will not venture a strict definition of narrative here, as our aim is to tease apart differences rather than identify commonality. However, we think of narratives as stories that are Goldie adds that grief is also 'experienced as a process'; its unity is phenomenologically accessible. But what does that unity consist in? At this point, Goldie turns to narrative, suggesting that what unifies grief is 'the coherence of a narrative of the process'. The unfolding of grief is, he says, 'narratable', with narrative corresponding to the distinctive temporal 'shape' of grief (2012,(61)(62). The narrative in question need not be something communicated to, constructed with or shared with others; it might be limited to a pattern of linguistic thought that is not imparted to anyone else. Goldie does not claim that narrative is the only way of identifying relationships between different features of grief. He allows that a causal account of the process may also be possible, but adds that narrative is a much richer resource to draw on, capturing aspects of grief that elude appeals to mechanistic causation. In particular, narrative, like grief, can include the interplay of 'multiple perspectives', such as those of a character and a narrator (2012,65).
There is much to be said for this approach, especially for the suggestion that multiple perspectives are important in grief. Nevertheless, several issues remain unresolved. First of all, it is not clear whether Goldie takes grief to be partly constituted by narrative or merely amenable to conceptualization and description in narrative terms. On the one hand, he maintains that the best account available to us is a narrative, rather than causal, account, suggesting that he is concerned with description. On the other hand, he indicates that narrative is what unites the various heterogeneous ingredients of grief, such that they comprise a singular, diachronic pattern. Hence, B's experience p at a particular time is only part of grief in virtue of a larger narrative that ties it to q, r, s and numerous other experiences.
Another possibility to consider is that however many components together constitute grief in virtue not of grief is a process of this kind, narratable in this way, that its parts "hang together into a coherent whole"' (Goldie 2012, 69). However, this position is problematic for three reasons. First of all, it is not clear how actual unity could be achieved by potential narration. It is more plausible, one might suggest, that grief is narratable because its parts already hang together. That would leave us back where we started, with the question of what unifies grief. Second, being narratable is arguably too restrictive a requirement for being a constituent of grief. It seems arbitrary to preclude the possibility of experiences that are integral to grief but not amenable to the narration. Indeed, it has even been suggested that certain experiences of grief involve a breakdown of narrative capacity, due to an altered sense of time and consequent disruption of narrative structure (Riley 2012;Ingerslev 2018). Of course, a permissive account could be offered of what it is to be narratable, so as to accommodate even these experiences. However, there is then the risk of endorsing something too accommodating to be informative. If all forms of experience turn out to be narratable, the observation that experience p is narratable will tell us nothing specific about experiences of type p. There has to be something distinctive about the temporal sequence of experiences and activities that constitutes grief, which makes it narratable in a way or to a degree that various other phenomenological sequences (which might similarly be interrupted by periods of distraction and unconsciousness) are not.
A third problem with Goldie's approach relates to when the relevant narratives are constructed and how many narratives are constructed. Presumably, one need not wait until a grief process has concluded (if it is even possible to identify an endpoint) and then construct a singularly appropriate, comprehensive narrative. Goldie's account instead concerns narratives that are constructed during grief. However, these are likely to undergo considerable revision over time. Suppose q is part of a grief-pattern in virtue of its inclusion in Narrative A. Narrative A is then abandoned and conflicting narrative B, which rejects q as a component of grief, is adopted instead (e.g. 'I thought q stemmed from my grief, but then realized I had been feeling like that for years'). Alternatively, q might be inserted into several successive, conflicting narratives, which all construe q as part of grief but ascribe it a quite different significance. In virtue of which inclusion or revision is q integral to grief? 3 In light of this, we might also wonder what is more important here: the inclusion of something within a particular narrative or the process of narration (where the latter may involve constructing, revising and abandoning any number of complementary or competing narratives). Goldie (2012, 3) acknowledges the need to distinguish narrative thinking from narrative as a product. We will further suggest that this distinction is central to appreciating the roles of narrative in grief. Goldie's account does not distinguish two importantly different roles, both of which are integral to a larger grief process; while specific narratives can aid in providing or sustaining structure and coherence, narration plays an important role in destabilizing structure and thereby enabling new structure to take shape.

What grief does
We agree with Goldie that grief should be construed as a perduring process, rather than an episode or enduring state. However, his account of grief's pattern does not capture the manner in which typical grief also has a direction. Narrative may well have some role to play in lending coherence to an emotional process. However, the roles played by narrative in grief are more fruitfully considered in relation to something that the process does. In brief, grief ordinarily involves comprehendingover a period of timewhat has happened, negotiating a disturbance of life structure, and restoring coherence to one's sense of self and world. 4 A process of this kind is implied wherever it is the case that (a) the reality of the death eventually comes to be fully acknowledged by the bereaved, and (b) the bereaved is then able to live a meaningful life, one that is consistent with the death having occurred. Although the phenomenology of grief is not exhausted by this, the process of reorientation does feature heavily in most first-person descriptions of grief. Its prominence varies, depending on the extent to which the person who died was integrated into one's life structure. In the absence of any competing accounts of grief that exclude this process, we take it to be part of grief than a mere accompaniment to it. We further suggest that the contrasting roles of narrative can be made clearer by identifying their different contributions to the process of recognition, reorientation and adjustment.
Grief is sometimes said to involve a disturbance or loss of self. It is not uncommon for people to say things like 'part of me has died', 'I died with her', 'I'm incomplete without him', 'I don't know who I am anymore', and so forth. Although such talk may seem clichéd on occasion, it can also be informative. In an important way, profound grief does impact on self-experience, on the sense of who one is. To a large extent, the structure of one's life can come to depend upon a relationship with a particular individual. This dependence encompasses various categories with which one identifies, each of which attaches to norms, expected patterns of activity, and ways of interacting with other people. Such categories partly specify who one is, in relation to particular people, larger groups or others in general: I am a wife, husband, mother, father, daughter, son, teacher, business partner, and so forth. In the event of a bereavement, one can cease to be a member of certain categories that were previously central to the structure of one's life. However, this sense of identity is not restricted to category membership of the form 'I am an x'. It also extends to a range of projects, commitments, pastimes and habits, expressed as 'I am someone who believes, does, strives to do, enjoys doing, or is committed to doing x'. Numerous commitments, pastimes and habits may only be sustainable relative to an interpersonal relationship. For instance, in the case of a major life project, it could be that we do this particular thing together in order to sustain or enhance our life together, I do this for you, or I do this because it is something that we care about. As for habits and pastimes, going to the cinema might be something that we enjoy doing, and walking in the park is something that we do together. Consequently, the structure of a life can be disrupted in all manner of ways by bereavement.
What we have in mind here is similar to what Christine Korsgaard (1996, 101) calls practical identity: 'a description under which you find your life to be worth living and your actions to be worth undertaking'. For Korsgaard, practical identity includes the likes of profession and religion, along with statuses such as being a parent, spouse or friend. Together, they comprise a sense of who one is, which regulates activities by specifying reasons for action, prohibitions and obligations. However, whereas Korsgaard (1996, 102) refers to this as a 'conception' of oneself, what we have in mind is broader in scope. With the inclusion of much that falls under 'I am someone who does … .', it becomes clearer that not all aspects of one's practical identity are explicitly conceptualized, at least unless and until they are disturbed by life events. 5 Phenomenologically speaking, much that is integral to who one is manifests itself not as an explicit conceptualization of self, but in the guise of an experiential world that one's thoughts and activities take for granted. Thus, as Kathleen Higgins (2020, 9) writes, 'profound loss catapults the bereaved person into an alternate "world" that differs in marked ways from the world we usually occupy, an alternate world lacking even the basic coherence we need to function'. But how could practical identity (permissively construed) manifest itself as a world? In fact, the answer is fairly straightforward. First of all, in thinking and acting, we take some things to be significant, to matter, while others do not. Things matter to us in different ways; they might offer the potential to enable or obstruct projects, to threaten us or those we care about, to provide relief, to elicit happiness or to inflict suffering. What we take to be significant reflects what we value, at least to the extent that our emotional lives are coherent. And these values are integrated into wider networks of values that together constitute our practical identities. The next step is to acknowledge that we do not, for the most part, first experience things and only afterwards assign one or another kind of significance to them. We experience the significance of things without recourse to inference. Smoke emanating from one's computer appears immediately concerning, as does the sight of one's child falling from a bike onto a hard surface. More generally, our experienced surroundings are imbued with all manner of significant possibilities, which are integrated to the extent that our projects and commitments are stable and coherent. 6 With a bereavement, the whole structure of one's world can become unsustainable (Ratcliffe 2019(Ratcliffe , 2020(Ratcliffe , 2022. There are two inextricable, but also importantly different, aspects to how this is experienced. First of all, experience of self and world does not alter instantly so as to assimilate the loss. Roles, statuses, projects and pastimes that are no longer sustainable may still retain their place for a time, in the guise of one's experienced surroundings and habitual inclinations. Consequently, there is a gulf between a world that endures, to varying degrees, and the explicit acknowledgement that someone has died, with all that the death entails. Making sense of the death is partly a matter of coming to comprehend all that no longer applies. However, it is also important to acknowledge a second aspect of the experience. Even if what has happened could be integrated instantaneously into one's practical identity, the result would not amount to successful adjustment. This is because, in the early stages of bereavement, there is nothing to replace the life structure that has been lost. One would be consigned to a world without established norms and patterns of significance, of the kinds that more usually elicit, shape and guide patterns of thought and activity.
A process of gradual rather than swift adjustment is therefore unavoidable. 7 While there is a need to reconstruct a world that cannot continue as it is, abandoning it all at once would involve casting oneself into a realm of indeterminacya world that lacks familiar norms, where things no longer appear significant in specific, unambiguous, cohesive ways and situations no longer elicit familiar patterns of activity. Nevertheless, even gradual adjustment involves experiencing some degree of indeterminacy and thus facing situations that no longer provide a basis for acting in one way rather than another. This sense of indeterminacy also includes tensions and ambiguities relating to the self: I am still who I was, and yet that identity is no longer sustainable; I cannot be that person anymore, but have no other identity; I am somewhere between a person who was and someone I cannot yet envisage.
One concept we can draw on in seeking to better understand the predicament is that of 'transformative experience', introduced by L. A. Paul (2014). According to Paul, certain decisions in life are both epistemically and personally transformative. We are unable to foresee their effects on our preferences and epistemic dispositions, where those effects can be substantial enough to change 'what it is like to be you, deeply and fundamentally' (2014, 3). Of course, for this to apply to grief, it must be further acknowledged that transformative experiences do not just arise due to our own decisions, but also due to contingent life events over which we have no control, as well as longer-term processes that involve complex mixtures of activity and passivity (Carel and Kidd 2020).
Importantly, some such transformations do not involve simply moving from A to B, where one cannot appreciate B from the perspective of A. In the case of grief, one also inhabits, for a time, an indeterminate realm between A and B. Thus, instead of simply contrasting the worlds of before and after, we also need to emphasize what lies in the middle, something that Ami Harbin (2016) calls 'disorientation'. During grief, one is not just unable to anticipate who one will be. One may also find that one's current sense of self is profoundly eroded; one is neither who one was nor who one will be. Norms and significant possibilities that would ordinarily constrain or even specify patterns of thought and activity are lacking. There is no fact of the matter concerning what is to be done next, how to continue. 8 As Harbin (2016, 2) writes, disorientations are 'temporally extended, major life experiences that make it difficult for individuals to know how to go on. They often involve feeling deeply out of place, unfamiliar, or not at home'.
The experience of profound grief can involve a tension. One is unable to sustain the self or world that was, but one cannot abandon it either, since there is not yet anywhere else to go or anyone else to be. Navigating upheaval therefore involves a balancing act between retention, repair, revision and loss of life structure. Some structure is required in order to sustain activities in the face of indeterminacy and serve as a basis from which to develop new structure. But the transformation of one's experiential world also requires the destabilization, over time, of established patterns. In what follows, we will suggest that narrative and narration are implicated in both, playing seemingly opposing but potentially complementary roles in a larger process that is fraught with tension and ambiguity. As we will see, several more specific roles for narrative can be discerned within the two overarching categories of sustaining/imposing and disrupting life structure.

Providing structure
In considering how narratives can provide structure and coherence during times of upheaval and uncertainty, it is helpful to distinguish three broad categories of narratives: 1. The established narratives of a culture. 2. Fairly stable, enduring narratives constructed by the bereaved person and/or others. 3. Dynamic, ephemeral narratives constructed and revised during interactions with others.
Our focus in this section will be on (1) and (2). We will then go on to suggest that narratives of type (3) sometimes play a very different role. In the cases of both (1) and (2), there is a distinction to be drawn between an established story and the token narratives via which it is told or performed, given that the same story can be told in different ways on multiple occasions, by one or more people. However, that distinction does not apply so obviously to (3), where a token narrative is collaboratively and spontaneously assembled, never to be repeated. In addressing (1) and (2), our use of the term 'narrative' concerns both established stories and their telling. 9 In the case of (3), however, we emphasize the dynamics of narration.
First of all, let us turn to (1). The recently bereaved are exposed to various narratives that do not apply specifically to their own experience. Rather, such narratives are established, shared ways of understanding bereavement and what might be expected of people in this context. They span a range of sources, including novels, plays, films, religious texts, self-help guides and the linguistic ingredients of established rituals and practices. Hence, they are quite diverse in structure, content and level of detail.
Established, shared narratives, we suggest, can serve as a form of scaffolding, which guides and structures activities on the part of the bereaved and others, during times when familiar norms may have ceased to apply and it is unclear what to say or do. Jerome Bruner (1990, 38-50) suggests that the 'established cultural meanings' integral to institutions and practices serve to guide our activities, and that much of what we do in everyday life conforms to a 'canonical cultural pattern'. But things do not always go as expected. So, interpretive procedures are also needed for dealing with actions that deviate from expected norms and established practices. Shared narratives, Bruner proposes, play an important role in re-integrating the 'exceptional' with the 'ordinary', making sense of what people say and do.
The point applies not only to seemingly anomalous utterances and activities, but equally to the task of understanding life events that disrupt norms and expectations. Narratives and associated prescriptions for interactions between people can contribute to the interpretation and navigation of events that involve loss of life structure and consequent indeterminacy. They can also involve scripts, performances and interpretative frameworks for others to follow when interacting with the bereaved. Gilbert (2002, 223) thus observes that narratives help people to 'make sense of disorder'. They are a part of what Neimeyer (1999, 66) describes as a 'complex process of adaptation to a changed reality', a process that is not merely personal, but 'relational' and 'cultural'. 10 Established narratives have the potential to promote coherence in two closely related ways. They can provide interpretive frameworks, which sustain or impose structure and guide action during times when practical identity is compromised. In addition, narratives have the potential to facilitate and support patterns of interpersonal interaction that themselves contribute to re-establishing life structure. The Danish philosopher, Knud Løgstrup, remarks on the extent to which we are porous to other people's words, gestures, expressions and actions. When we interact, we affect one another to varying degrees and in different ways. In fact, it is impossible to interact with another person without affecting her in some way, however subtly. Conventions, Løgstrup writes, 'facilitate our relationship with one another, making it smooth and effortless, not least because they protect us against psychic exposure'. He adds that, in the absence of this shared structure for regulating our exchanges, interacting with other people would be 'unbearable' (Løgstrup [1956(Løgstrup [ ] 1997. Without a context of established norms and performances to fall back on, both parties would be over-exposed to each other's influence, with nowhere to hide from feelings of awkwardness and vulnerability generated by unconstrained interaction. Whether or not this is true in general, it is plausible in the case of grief. It is widely recognized that feelings of social awkwardness and discomfort are commonplace when responding to others' bereavements (Higgins 2013, 11). This, we suggest, is partly due to a lack of shared norms and patterns of significant possibilities that are more usually taken for granted by both parties, shaping and constraining interactions between them.
The narratives that people draw upon have various different foci, including the event of the death, how one has been affected by the death, relations with other people, the challenges that are now faced and/or the grieving process itself. One important theme is how people do and should think about and relate to those who have died. Here, differing and sometimes conflicting narratives might be utilized for interpretation and guidance. Grief is sometimes said to culminate in 'letting go' or 'moving on'. However, other narratives emphasize the establishment of altered but enduring relationships with those who have died. A substantial literature on 'continuing bonds' supports the view that bereavement need not and often does not involve simply letting go or moving on. In fact, most people retain connections or bonds with the deceased and maintain certain aspects of the relationship (Klass, Silverman, and Nickman 1996;Klass and Steffen 2018). Trying to make sense of the fact that someone is no longer part of one's life in many ways, but still part of one's life in others, adds to the challenge of negotiating ambiguity and upheaval. For many, it will involve a range of experiences that are unsettling, disorienting, distressing, bewildering and difficult to interpret (Ratcliffe 2021(Ratcliffe , 2022. This is especially so when an interpersonal relationship is integral to who one is, given that renegotiating the relationship is inseparable from a process of sustaining and to some extent altering one's practical identity. 11 Stories and associated practices can contribute by providing exemplars to follow and more widely applicable frameworks through which to make sense of one's experiences. Some of these, we have suggested, are culturally established, providing a common structure for people to draw on. Others, however, will be idiosyncratic. In addition to relying on generic narratives, people construct and refine their own narratives with others, which further contribute to the tasks of reshaping a relationship with the deceased and negotiating indeterminacy. The point applies equally to those narratives that focus on various combinations of what has happened, who one is, and how one relates to other people. According to Korsgaard (1996), if there is no practical identity, there can be no reasons for action; all norms cease to apply. Although the predicament we have described here does not involve complete loss of life structure, it still presents a profound challenge to it. In the face of this, co-constructed stories about the bereaved and the deceased may provide cohesion and direction: this is what has happened; this is how it has affected me; this is where we are now; this is what I need to do; this is what others should do; this is what we should do together; here is how I should think about the person who has died. However, the distinction between idiosyncratic and cultural narratives of bereavement should not be drawn too sharply, given that even a personal story will draw on culturally available resources. 12 For the most part, the life structure that shapes our experiences, activities and thoughts does not consist in 'narrative'. Even if Lamarque's (2004) minimal definition of 'narrative' is endorsed, experiencing things as significant, habitually doing various things and acting in ways that reflect one's projects do not consist in narratives, although they might be articulated as such. Nevertheless, when life structure is eroded by bereavement, the imposition of an explicit narrative concerning who one is, what one does and why, how one relates to others, and what one seeks to achieve can serve as a temporary substitute for that structure, analogous to a cast on a broken bone. 13

Opening up possibilities
We have suggested that narratives of different kinds can help keep at bay the indeterminacy of grief, providing provisional or enduring structure to make sense of what has happened and what is happening. 14 Insofar as this structure is prescriptive, it amounts to a form of scaffolding, facilitating patterns of activity when the once familiar world has been disrupted and no longer specifies what matters, how to relate to other people, and what is to be done. It is thus tempting, when addressing the role(s) of narrative in grief, to focus exclusively on its organizing roles: narrative imposes a pattern, adds coherence, constructs meaning. However, we should also consider the dynamics of narration and, in particular, its interpersonal dynamics. In doing so, it becomes apparent that destabilizing established patterns and opening up new possibilities are equally important. With this in mind, let us return to the issue of how narratives might contribute to continuing bonds. Walter (1996) considers the part they play in keeping the dead within our lives, going so far as to suggest that the 'purpose' of grief is 'the construction of a durable biography that enables the living to integrate the memory of the dead into their ongoing lives'. Such narratives, Walter adds, are most often constructed collaboratively. Grief is an 'intrinsically social process', which involves developing a story about who someone was, how they died and what they meant to people (Walter 1996, 7-13).
The suggestion that constructing a biography is the purpose of grief seems implausible. Granted, it is something we might do during grief, but it is one among many aspects of grief and perhaps a contingent one at that. In fairness to Walter, though, he does not seek to generalize and only takes himself to be offering an account of how grief sometimes operates. Perhaps a case could be made for this more limited claim. Nevertheless, there is a further objection, which concerns the claim that narrative construction is directed towards producing an 'accurate picture of the deceased' (Walter 1996, 13). In fact, accurate portraits can be self-defeating during grief. Striving to produce a detailed, singularly appropriate picture of the person who died and her relations with others may not lead to an enhanced sense of connection or improved recollection, but instead to losing connection with the person, ceasing to remember them properly (Ratcliffe 2021(Ratcliffe , 2022. For example, in Camera Lucida, Barthes ([1980] 2000, 66-109) describes trying to 'recognize' his mother, who had recently died, in numerous photographs. Although he could distinguish her from other people easily, he could not 'find' her. Eventually, he discovers his mother, not in an accurate image of her features, but in a photograph of her as a child that somehow captures her, enabling him to 'discover' her. What he saw in the photograph was her 'kindness', not in any particular perceived properties, but in a nonlocalized 'air' that is not reducible to an assortment of details.
What is the difference between this air and an accurate image? We suggest that it is attributable to the former consisting in an appreciation of various relational possibilities involving a person, an openness to being affected by her in distinctive ways (Ratcliffe 2020(Ratcliffe , 2022. This essentially involves being receptive to that person's spontaneity and thus to the potential for surprise. An accurate image can actually erode this uncertainty and openness, replacing it with something determinate, inflexible, inanimate. The point applies not only to images, but also to the roles of narrative with respect to continuing bonds. What is important in sustaining, restoring and altering one's sense of a particular person and one's connection with that person is not a single, definitive account, but instead the possibility of endless revision and novelty. A good story about a person is somehow incomplete. As Bruner (1990, 53-54) observes, 'to make a story good [… .], you must make it somewhat uncertain, somehow open to variant readings'. Narratives do not just constrain interpretation; they also generate new possibilities for interpretation. Complementing this, Gilbert (2002, 225) emphasizes the 'ongoing and evolving nature of personal narratives' in grief. Higgins (2013, 175) goes further, to suggest that a dynamic, interpersonal process of narrative construction can play the role of 'symbolically reanimating the dead'. It is not a matter of collaboratively pinning down who a person was, but of exchanging stories that enrich participants' conceptions of the deceased, each story revealing its predecessor to be limited in certain respects, unable to capture all of the relational possibilities associated with the person. Those involved consequently experience what Higgins calls an 'enrichment' of perspective, revitalizing their sense of what a person was like in ways that essentially involve the story remaining incomplete, forever in flux.
This contrast between narratives that impose structure and narration that disrupts it can also be applied to the impact of bereavement on one's life structure, the sense of who one is. As well as animating the dead, the process of narrating and re-narrating can serve to disrupt and recontextualize life structure, contributing to a fluidity that is required for a shift in practical identity. The course of grief involves an ongoing interplay between the two roles, as well as different balances between them. Goldie draws a comparison between memory in grief and free indirect style in literature, where the perspectives of character and narrator are combined. He suggests that episodic memory in grief can involve 'the psychological correlate of free indirect style'. For instance, 'you remember the last time you saw the person you loved, not knowing, as you do now, that it was to be the last time' (Goldie 2012, 65). The memory is infused with the perspective of the present, with internal and external perspectives that are to some degree discordant, and yet integrated. 15 In our view, this is better construed in terms of the act of narration than the establishment of a fixed narrative. As one reflects on past events, memories are dislodged from their habitual places and take on new significance. This is part of a larger reorganization of life meaning, whereby habitual patterns are disrupted, opening up the potential for reinterpretation and reorganization. Hence, where one's own life and also one's relationship with the deceased are concerned, the production of narrative (as distinct from the establishment of a particular narrative) can involve the generation of new possibilities. We therefore suggest that Goldie's account, which emphasizes narrative coherence, in fact, accommodates two different things: narratives that lend coherence and narration that disrupts.
The proposal that narrative plays an important disruptive role is consistent with maintaining that it also mitigates disruption. There is no conflict between disrupting x while stabilizing y. Disrupting certain things, while retaining or imposing structure elsewhere, is how we negotiate a route from a world that no longer applies to one that does. Nevertheless, there are often tensions between the two. Consider this passage from Julian Barnes' memoir of bereavement, Levels of Life: … .the grammar, like everything else, has begun to shift: she exists not really in the present, not wholly in the past, but in some intermediate tense, the past-present. Perhaps this is why I relish hearing even the slightest new thing about her: a previously unreported memory, a piece of advice she gave years ago, a flashback of her in ordinary animation. I take surrogate pleasure in her appearance in other people's dreamshow she behaves and is dressed, what she eats, how close she is now to how she was then; also, whether I am there with her. Such fugitive moments excite me, because they briefly re-anchor her in the present, rescue her from the past-present, and delay a little longer that inevitable slippage into the past historic. (2013,108) Here, Barnes is describing a process that involves consigning his relationship with his wife to memory and accepting a narrative according to which she is gone, in the past. However, this involves gradually losing her, in a way that jars with narratives that are generated via conversational interaction with other people. The latter re-instill a sense of her spontaneity, bringing her back into the present. A tension thus arises between the consolidation of one narrative, according to which his wife and his relationship with her reside in the past, and interpersonal narration processes that restore a sense of connection. Consequently, she inhabits an unstable realm between the two.
So far, we have mentioned the roles of narratives that focus on one's own life (past, present and/or future), the deceased or one's relationship with the deceased. However, the contrast between imposing and destabilizing structure also applies when we turn to the emotional experience of grief itself: what is it that I am feeling; what are these emotions? Where emotional experience is concerned, the point relates to language more generally (e.g. naming), rather than just narrative. A consistent theme in the work of the phenomenologist Maurice Merleau-Ponty is that language does not merely express pre-formed experiences and thoughts. In some instances, it completes them, by rendering them more determinate: 'speech does not translate a ready-made thought; rather, speech accomplishes thought' (Merleau-Ponty [1945. More recently, Giovanna Colombetti has suggested that the expression of emotional feelings, including their verbal expression, sometimes lends precision to them. For instance, naming emotions 'squeezes complex feelings into something compact' (Colombetti 2009, 17). This can be conceived of in terms of the linguistic completion of emotion, rather than the expression of an emotional experience that is already fully formed. 16 One way in which language can add determinacy to an experience is by resolving its contentwhat an emotional feeling is about. However, it can also contribute to the sense of which kind of emotion one is experiencing. This latter role is not exclusive to specifically emotional experience. Consider, for instance, the phenomenological distinction between perceptual experience and memory and, more specifically, the sense of something as present or past. As illustrated by the earlier quotation from Barnes (2013), certain affectively charged experiences can fall between the two. It is not fully clear whether one experiences the deceased as present or as past, as here or as gone. Narrative, and language more generally, can contribute to making temporally vague experiences more determinate, influencing whether something crystallizes as a memory or a perceptual (or perception-like) experience.
However, language also contributes to emotional experience by disrupting established names, norms and narratives, opening up the possibility of reinterpretation. Drawing inspiration from Saussure, Merleau-Ponty (1964, 1973 distinguishes between 'language' and 'speech', where the former follows paths specified by established, shared norms, while the latter disrupts them, enabling new possibilities for experience and thought: We may say that there are two languages. First, there is language after the fact, or language as an institution, which effaces itself in order to yield the meaning which it conveys. Second, there is language which creates itself in its expressive acts, which sweeps me on from the signs toward meaningsedimented language and speech. (1973,10) In the context of grief, what Merleau-Ponty calls 'speech' can be ascribed a role in disrupting entrenched patterns of thought that presuppose what is no longer the case, thus opening up new possibilities. As Merleau-Ponty (1973, 17) remarks, 'in its live and creative state, language is the gesture of renewal and recovery which unites me with myself and others'. In this way, it may also contribute to how grief itself is experienced. For instance, recognizing the profundity and novelty of emotional experiences could involve releasing them from established labels and familiar narratives of emotion. 17

Pathological grief
To summarize, we have suggested that narrative plays multiple and contrasting roles in grief. These fall into two broad and potentially complementary categoriesnarrative pins some things down and stirs other things up. Although we have construed both of these in a positive light, it is important to add that narrative will not always make a positive contribution to processes of adjustment. It can equally serve to interfere and obfuscate, again in a variety of different ways. For instance, Strawson (2016) introduces the notion of 'narrative bypassing': relying on a narrative conception of self in ways that facilitate detachment from and avoidance of certain emotions. Bypassing could arise via any of the routes we have identified. Deciding to follow or being compelled to follow cultural narratives of grief could alienate a person from her own grief, by providing an alternative course of action to that of comprehending and engaging with her individual situation. One might also be compelled to follow a shared narrativeperhaps within a familythat ascribes various qualities to the deceased, in such a way as to obfuscate one's own understanding of the person who has died and bypass the idiosyncrasies of the relationship. One could also hide behind certain self-narratives, constructed alone or in collusion with others. In addition, certain kinds of narrative exploration could involve sustaining connections with the deceased that distract from the realities of the death and the task of reorientation. Bypassing can also feature in how others interact with the bereaved; by adhering to established performances, it is possible to avoid any real confrontation with what has happened and what someone is feeling.
Reflecting on the roles of narrative can also contribute to the more specific task of distinguishing the variety of 'typical' grief experiences from forms of grief that are deemed pathological. To be more specific, our account points to the conclusions that (a) currently established conceptions of pathological grief accommodate importantly different kinds of experience and (b) all such experiences ought to be conceived of in interpersonal, social and cultural terms, rather than exclusively in terms of processes that are internal to the individual. The term 'complicated grief' has been in use for some time; it refers to a form of grief that is claimed to be pathological due to its association with intense, enduring emotional pain and longer-term functional impairment. Prigerson et al. (1995, 69) list the following symptoms: … .preoccupation with thoughts of the deceased that would make it difficult to do things one normally would do, anger over the death, distrust and detachment from others as a consequence of the death, pain in the same parts of the body as that experienced before the death, avoidance of reminders of the deceased, feeling that life is empty without the deceased, auditory and visual hallucinations of the deceased, survivor guilt, loneliness, bitterness about the death, and envy of others who have not lost someone close. 18 More recently, a consensus has formed around the adoption of the alternative term 'prolonged grief disorder'. This was proposed and later endorsed for inclusion in the World Health Organization's ICD-11 (2018). Subsequently, prolonged grief disorder was also approved as a diagnosis for inclusion in the 2022 text revision of the American Psychiatric Association's DSM-5. 19 The DSM-5-TR and ICD-11 diagnostic criteria are not identical, and both are more specific than earlier descriptions of complicated grief. Even so, the different terms and associated diagnostic criteria are generally regarded as more or less effective ways of identifying and referring to the same phenomenon, rather than as attempts to identify different phenomena (Ratcliffe 2022, Chapter 8). And all of them, we suggest, remain sufficiently permissive to accommodate importantly different forms of experience.
For example, consider the ICD-11 description of prolonged grief, which refers to an enduring preoccupation with or longing for the deceased, painful emotions of various kinds, inability to accept the death, a sense of having lost part of oneself, social disengagement, a duration of at least six months, departure from relevant cultural norms and impairment of function. 20 It is not our aim here to arbitrate over whether such experiences are indeed pathological. Instead, we want to suggest that the various descriptions of complicated and prolonged grief are compatible with at least two importantly different forms of experience, involving contrasting imbalances between the retention and disruption of life structure. We have suggested that narrative and narration are integral to a larger adjustment process, whereby an unsustainable world is destabilized over time, structure is preserved and imposed so as to facilitate negotiation of indeterminacy, and one's relationship with the deceased is reshaped. Without significant disruption of established patterns, one would be stuck in an unsustainable world, unable to reorientate. Conversely, with an excess of disruption, one would lack the life structure required in order to comprehend what has happened from a stable vantage point, integrate it into a life and develop a new structure. In both cases, one would be stuck, but in different placesa world that is no longer sustainable or an indeterminate realm between worlds.
One way of conceptualizing this contrast is by appealing to the influential dual-process model of bereavement proposed by Stroebe and Schut (1999). According to this model, grief has a dynamic structure that involves 'oscillating' between loss-and restorationoriented coping, where the former involves engaging with the loss of a person, while the latter involves reorienting oneself in the world. Stroebe and Schut add that there is a further movement between engaging in coping activities of one or the other kind and taking 'time out' from both (Stroebe and Schut 2010, 278). As they suggest, we can think of pathological grief in terms of different imbalances in this movement. In one scenario, the bereaved person is so fixated on loss as to be practically disengaged from their surroundings. Consequently, life structure is not revised in a way that accommodates the reality of the death. In another scenario, the person does reengage with the world, but in a way that does not take adequate account of what has happened or its implications for life structure. In both cases, we have what Stroebe and Schut call 'disturbances of oscillation ' (1999, 217). A process whereby one moves between worlds past and present, leading to gradual adjustment, is skewed too far towards one or the other extreme. But these imbalances need not amount to distinct forms of grief; a single grief process could involve alternation between the two.
In addition, Stroebe and Stroebe (1999, 218) identify a predicament where the oscillation process breaks down rather than leaning towards one or another pole. There is, as they say, a 'disturbance of the oscillation process itself'. It is this distinction between a lack of oscillation and a breakdown of oscillation that corresponds to the two broad roles we have identified for narrative: providing and disrupting life structure. Some forms of experience involve an excess of structure, which obfuscates comprehension of and adaptation to loss, whereas others involve acknowledging loss without the establishment of the new structure. One loses so much that there is not enough left to build from; there is no coherent orientation from which to reengage with the world. Importantly, the various roles played by narrative do not involve the solitary construction of narratives, but rather their co-construction against a shared backdrop of stories, rituals, practices and norms. Thus, insofar as narrative processes (or their absence) are implicated in prolonged grief, this is not to be construed in intrapersonal terms, but also in terms of lacking certain interpersonal and social interactions. Due to whatever cause, the bereaved person is cut off from interpersonal processes that are integral to the renegotiation of life meaning. As Neimeyer (2005, 28) suggests, it is the 'unsuccessful struggle for meaning' that distinguishes 'complicated forms of grief' from others.
The interpersonal, social and cultural aspects of grief are especially evident when we turn to 'disenfranchised grief', a term coined by Kenneth Doka to refer to a kind of grief 'experienced by those who incur a loss that is not, or cannot be, openly acknowledged, publicly mourned or socially supported' (Doka 1999, 37). Although distinct from 'pathological' grief, disenfranchisement is said to be an important risk factor. Contributing factors can include the nature of one's relationship with the deceased, the involvement of a form of loss that is not widely recognized, ways of grieving that transgress established norms, and the circumstances of a death. The common theme is that others fail to acknowledge or legitimate one's grief, in ways that interfere with the grieving process, sometimes culminating in prolonged grief. 21 Central to disenfranchisement is the privation of forms of interpersonal interaction, including linguistic interaction, which those who are grieving might otherwise expect. With this, the bereaved are also deprived of access to narratives and forms of narration that might otherwise shape and regulate grief. 22 Disenfranchisement could involve any combination of the following: (a) alienation from or lack of access to shared, cultural narratives, along with associated practices; (b) inability to share stories with others in ways that would otherwise have shaped and reshaped a sense of one's relationship with the deceased; (c) lack of access to the provisional and longer-term narrative structure for negotiating indeterminacy; (d) lack of interpersonal interactions, including linguistic interactions, that would otherwise shake up habitual patterns of experience, thought and activity.
Hence, we suggest that the topic of pathological grief is best approached from a perspective that emphasizes the roles played by interpersonal relations and interactions. Amongst other things, our relations with others facilitate the construction and reconstruction of stories spanning what has happened, the effects on one's own life, one's relationship with the deceased and the experience of grief itself. Grief is a diverse, multi-faceted and fragile process, the course of which is not attributable solely to individual psychology. Grief processes are shaped by interactions with other people, against the backdrop of a social and cultural context that isto varying degreesshared. Disturbances of grief might well arise from a range of sources: self; particular people; others in general; contingent events; society; and culture. However, regardless of where these disturbances originate, they are typically interpersonal and social in structure. Notes 1. Although we emphasize the case of bereavement, much of what we say here is also applicable to experiences of loss more generally. 2. A limitation of our discussion is that we are concerned for the most part with non-fictional narratives. We also limit ourselves to the first-person singularto the roles of narrative in the lives of individuals who are grieving. There is more to be said concerning narratives that others tell about those who are grieving and, in those cases where narratives involve more than one grieving individual, about the dynamics between 'I', 'you' and 'we'. 3. See also Brunning (unpublished) for a critique of Goldie's account that raises similar concerns. 4. Unlike Goldie (2012), we are not concerned principally with what unifies grief. However, the answer, we suggest, relates to the manner in which a human life is integrated. The various aspects of a life are not disrupted in an atomistic fashion. Rather, there are relationships of mutual implication. For instance, Project A may be unintelligible without Project B, which is itself unintelligible in the absence of a particular person. Grief is unified insofar as it engages with a unified disturbance of life structure (Ratcliffe 2017b). 5. See also Cholbi (2019) for an account of grief that emphasizes its impact upon one's practical identity. Cholbi further claims that grief provides us with a valuable opportunity for selfinterrogation and the acquisition of self-knowledge.
21. Rinofner-Kreidl (2016) further suggests that even typical or healthy forms of grief involve degrees and kinds of disenfranchisement. 22. Consistent with this, Higgins (2013, 173-174) observes that 'one of the problems caused by disenfranchised grief is that the lack of social support also makes it difficult for the bereaved to find occasions for talking, constructing narratives, about the loss'.