Shared emotion without togetherness: the case of shared grief

Abstract

I offer a philosophical account of shared grief, on which it is a process, undergone by a group, of recognising and accommodating significant possibilities that are lost to that group. In setting out from an understanding of grief’s distinctive characteristics, a philosophically interesting, metaphysically undemanding, and practically useful account of shared grief comes into view, that has broader consequences for understanding shared emotion.

1 Introduction

In this paper, I offer a philosophical account of shared grief. On this account, shared grief is a process, undergone by a group, of recognising and accommodating significant possibilities that are lost to that group. Philosophical accounts of shared emotion do not typically seek to explain the sharing of a particular emotion such as grief or fear, being framed instead as accounts of shared emotion in general. For instance, whilst Max Scheler’s example of parents grieving a child is often used to illustrate discussions of shared emotion—especially more recent discussions explicitly informed by work in phenomenology—these discussions have not concerned themselves with the characteristics of grief (I will introduce this example in Sect. 3). Here, I demonstrate the fruitfulness of departing from this approach and focussing on the sharing of a particular emotion, in this case, grief. In setting out from an understanding of grief’s distinctive characteristics, a philosophically interesting, metaphysically undemanding, and practically useful account of shared grief comes into view. Furthermore, despite the focus on a particular emotion, there are lessons to be learned for understanding shared emotion more generally. The view might be used as a model to understand a range of other shared emotions and demonstrates that neither similar feelings nor a sense of ‘we-ness’ are required for genuinely shared emotion.

I will proceed as follows. In Sect. 2, I outline an account of the nature of the grief of individuals. On this account, the object of grief is a loss of significant possibilities. The grief of an individual is a process of recognising and accommodating this object. In Sect. 3, I show that there can be a philosophically interesting form of shared grief, since the process described in Sect. 2 can occur in shared form. In Sect. 4, I develop this view of shared grief further, showing that it not only permits dissimilar feelings amongst the participants in a shared grief process but that such feelings can enable shared grief. As a result, the account of shared grief accurately and usefully reflects the reality of grief in groups. Finally, in Sect. 5, broader lessons are identified.

2 What grief is

To show that grief can be shared, we need, first, to say what grief is. To that end, I will in this section outline an account of the object of grief and the relation that a grieving individual stands in to that object. This, in turn, will allow us to understand grief’s processual nature: its status as a lengthy, heterogenous and, I will suggest, telic process. This account of individual grief is not intended to be controversial. It is, for the most part, a philosophical specification of a view of grief that is widely accepted amongst grief researchers.

The object of grief is obviously loss of some kind, although—since there are losses, such as the loss of an umbrella, that we do not typically grieve—more must be said about the kind of loss which is grief’s object.¹ The kind of loss that occasions grief involves profound disturbance to something very encompassing, which has been labelled an ‘assumptive world’ (Parkes, 1988), a ‘construction of reality’ (Marris, 1986), or a ‘meaning structure’ (Gillies & Neimeyer, 2006). Some recent philosophical work on the object of grief can be seen as further specifying this kind of view of what the losses that we grieve disrupt. For example, Cholbi (2021) has suggested that grief’s wide-ranging disruption can be understood by taking grief’s object not to be someone’s death, but the loss of a relationship with someone on whom one’s practical identity depended. On another view, developed in most detail by Ratcliffe (Ratcliffe, 2022, Ratcliffe et al., 2023) grief’s object is instead a loss of significant possibilities: in the case of bereavement, a wide-ranging network of possibilities that depended on someone’s continued, living presence.²

Additionally, according to this lost possibilities account, a network of lost possibilities is an appropriate object of grief—as opposed to, say, disappointment or regret—only if it is something on which your sense of identity depends. It is this aspect of the account that rules out feeling grief over an isolated minor loss, such as the loss of an umbrella.³ ‘Identity’, here, is not a matter of the logical identity of persons over time. Instead, it is the target of what Schechtman calls the ‘characterisation question’. To answer this question is to pick out ‘the set of characteristics each person has that make her the person she is’ (Schechtman, 1996, p.74). This captures well the meaning that ‘identity’ typically has outside of philosophy, in expressions such as ‘identity crisis’, where a person experiences painful uncertainty about the characteristics that in this way define them. Thus, to have a sense of identity in the relevant respect is to have a sense of which characteristics are truly one’s own. A sense of identity can change over time and need not be consciously articulated by a subject. Rather, it may be implicit, showing up in, for instance, the choices they make, or how they respond to various kinds of criticism or praise. The loss of a network of possibilities is a potential object of grief when one’s sense of identity thus understood depended upon it to a sufficient degree (a degree which we need not think can be made precise).

I propose to adopt the lost possibilities account of grief here. Since it has been defended elsewhere, and my primary aim in this paper is to defend a view of shared grief, I aim only to provide some sense of why the lost possibilities account is plausible enough to make use of in that view of shared grief. Furthermore, the success of this view of shared grief will not depend on ruling out alternative accounts of grief’s object. Accounts such as Cholbi’s, and others that may be developed in the future that also explain grief’s wide-ranging disruptions, may also be consistent with my account of shared grief, suitably adapted. In fact, it is a virtue of the account of shared grief offered here that accepting it is consistent with more than one possible view of the object of grief.

One salient reason to prefer the lost possibilities account to Cholbi’s is that it can more easily accommodate the variability of grief, in particular, variability in the extent to which in grief, the loss is felt for oneself. On Cholbi’s account, grief is always partially self-directed in that its object is one’s relationship with the person who has died. The lost possibilities account can also capture grief’s self-directedness, since some of the lost possibilities that constitute grief’s object may be things that the deceased made possible for you, the bereaved. This can range from the relatively small and concrete—eating a favourite meal cooked by them—to the more encompassing and diffuse—feeling at home in one’s living space or town or anticipating the future with excitement. It can also include shared possibilities: carrying out joint projects and enjoying shared pastimes, for example. However, the lost possibilities account is better placed than Cholbi’s to recognise that grief can be more or less self-directed. That’s because what is experienced as lost may also include things that had been possible for the person who died, such as seeing a project come to fruition, or overcoming some personal difficulty. The objects of different tokens of grief will vary in the balance of different kinds of possibilities lost. For instance, the long-anticipated death of an elderly, comatose relative may involve the loss of very few possibilities for the deceased. In contrast, if one were to receive news of the death of a child who one expected never to see again—imagine adoption or even long-distance space travel—one may feel the loss of few of one’s own possibilities or of possibilities shared with them. Instead, the focus might be wholly on the lost possibilities (for flourishing or happiness, say) of the deceased.

There being an appropriate object of grief does not of course suffice for grief’s occurrence. To understand grief’s nature we need also to consider the relation that a grieving subject stands in to this object, i.e., to the loss of a network of identity-relevant possibilities. According to the lost possibilities account, this relation is twofold: one of both recognition and accommodation. In recognising the object of grief, one becomes aware of the breadth of one’s loss—of the manifold possibilities that have gone. Recognition also involves the ‘sinking in’ of the loss, so that the feeling of ‘disbelief’ that often characterises early grief ebbs away (see Ratcliffe et al., 2023, p.321). In accommodating the object of grief, one ‘relearns the world’ (Attig, 2011), adjusting to what one recognises has gone, repairing or recreating that which has been disrupted. Accommodation—which might be inseparable from recognition—can involve forming new expectations, projects, and habits of thought and action, which can in turn involve changes in how one perceptually and emotionally responds to the world.

Another widely acknowledged idea about grief is that it is a temporally extended process, something captured in Wittgenstein’s remark that it sounds ‘queer’ to say, ‘for a second he felt deep grief’ (1953, p.174). Unlike pain, which can be felt ‘for a second’, grief is necessarily temporally extended. Furthermore, grief is, as Carolyn Price puts it, ‘a complex emotional process, involving a number of emotions’ (Price, 2010, p.25). In addition, this process can also involve other mental states and events such as memories, misperceptions, and imaginings. Grief is thus a heterogenous process, and conceiving of it as a process of recognising and accommodating the loss of a network of identity-relevant possibilities allows us to understand why this is so. Becoming aware of the breadth of loss, the sinking in of the loss, and the development of new habits (and so on) will involve various kinds of mental item, that do not obtain or occur all at once. Recognising and accommodating the lost possibilities that constitute the object of grief occurs slowly and to some extent severally, and the process is heterogenous in that ‘not everything that happens during the process is happening at any one time’ (Goldie, 2012, p. 63).⁴

A final point on grief’s processual nature pertains to how its varied elements hang together as a singular process. Grief, I propose, is a telic process or accomplishment.⁵ Telic processes such as walking to the shops or baking a cake have ‘built in’ end points that function as success or fulfilment criteria (see for example Crowther, 2020 for discussion). As such, if one sets out to walk to the shops and then stops, it does not follow that one has walked to the shops. Recognising and accommodating the network of significant possibilities that constitutes grief’s object is the success or fulfilment criterion for grief in the same sense. An individual’s feelings and other mental items hang together as a singular grief process in that they tend, together, towards the end of recognition and accommodation.

Before moving on to shared grief, some clarificatory points are in order. First, the claim that grief is a heterogenous process is distinct from the—popular but empirically ill-supported—claim that it proceeds in a limited number of predictable stages, such as denial, anger, bargaining, depression, and acceptance (see Kübler-Ross & Kessler, 2005 for this view, and Stroebe et al., 2017 for an important critique). Second, that recognition and accommodation of lost possibilities is the ‘end’ towards which grief tends does not imply that grievers intend or aim at this end. Instead, it is simply a fact about humans—and perhaps some other creatures—that the disruption occasioned by significant loss is typically resolved in this way, whether or not one wants or intends or even knows that this is so. Third, the claim that recognition and accommodation provide the telic end of grief is consistent with all manner of grief trajectories and views of the typical or healthy length of grief. For instance, it seems plausible that the telic end of grief is some degree of recognition and accommodation, and there may be no reason to think that this degree can be specified. Thus, on the lost possibilities account it is consistent with a subject experiencing grief that there are aspects—perhaps many aspects—of his loss that are never fully recognised, and/or accommodated. Relatedly, it shouldn’t be thought that ‘accommodation’ as I use it here entails equanimous acceptance of loss, or a state of being that is, overall, positive or to be recommended. One might in principle accommodate lost possibilities by developing unhappy expectations, projects that are undesirable even by one’s own lights, and all sorts of unpleasant habits of thought and action.

3 The shared process view of shared grief

On the account of grief’s nature just outlined, individual grief has certain interrelated features that are central to its nature. First, it concerns a loss of significant possibilities, and second, the significance of these possibilities lies in the dependence of one’s sense of identity upon them. Third, the grieving individual stands to that loss in a twofold relation of recognition and accommodation. Fourth, and as we saw this can be understood in terms of the previous three features, grief is a lengthy, heterogenous process. Fifth, and finally, the elements of this process hang together as elements of a singular grief process, in that they tend towards the telic end of recognition and accommodation of grief’s object. I will argue that there is a form of shared grief in which all these features are present in shared form, thus defending the shared process view of shared grief or, for brevity, the shared process view.

The shared process view is best introduced using an example. When a member of some suitably close-knit group dies, possibilities are lost not just for each member of the group, but for the remaining members of the group, together. For instance, in her interview study of grief in families, psychologist Janice Nadeau discusses an example in which a father and son had lost the possibility of easy communication on the death of the mother of the family, Debbie:

During the interview, they did not talk with each other. It seems likely that they had been accustomed to communicating through Debbie before she died and that they had not yet developed the skills or willingness to talk with each other directly (Nadeau, 1998, p.157).

This loss of the possibility of easy communication is not a loss just to the father, nor to the son: it is a shared loss. Now let us also suppose—imaginatively filling in details not provided by Nadeau—that having a relationship in which easy communication is possible is an aspect of this family’s sense of identity, of who they are: ‘we are a family that communicates easily.’⁶ We can further imagine, without going into too much detail, that the father and son have lost various other possibilities upon which other aspects of the family’s sense of identity depended. For instance, we can suppose that they have lost possibilities for continuing with shared projects and pastimes, and that these projects and pastimes were part of ‘who they were’ as a family. This allows us to see how there can be a shared object of grief: a loss to a group that consists in the loss to them of various possibilities upon which their shared sense of identity depended.

As was emphasised in the previous section, grief is not merely an emotion with a specific kind of object, it is also an emotion in which one stands in a certain twofold relation to that object: one of recognition and accommodation. Hence, if a group stands in that twofold relation to a shared object of grief, then that group will undergo a form of grief that it is natural to describe as shared grief. That is, in this form of shared grief, a group would be said to recognise its losses, and to accommodate them. To see that a group can indeed recognise and accommodate its losses, let us return again to the specific shared lost possibility we have been discussing: the father and son’s loss of the possibility of easy communication. The father and son might recognise this loss over the course of various failed attempts to communicate as they used to. After numerous failed attempts, it will be plausible to say of them that they recognise that they no longer communicate as they once did. In fact, the father might (uncharacteristically, given their communicative difficulties) say to the son ‘we can no longer talk as we did’, to the son’s regretful agreement. The father and son might accommodate the loss as they gradually learn to communicate with one another easily, without the mother. We might imagine that this happens as they get into the habit of communicating regularly but in some new ways (text messages, emails) that they didn’t make use of when the mother was alive. Alternatively, this loss could be jointly accommodated less happily, as they make a habit of their mutual awkwardness and adjust their sense of identity accordingly.

Of course, this is just a single lost possibility. It would be the father and son’s collective recognition and accommodation of numerous shared and identity-relevant possibilities that would count as their shared grief, on the shared process view. As in individual grief, this process will be lengthy and heterogenous, involving various mental items. Whereas the varied elements of an individual’s grief process all belong to that individual, each of the varied elements of a shared grief process might belong to individual members of the group. However, in both cases, the elements hang together as a unitary process due to their tending towards to the telic end point of recognition and accommodation of the relevant loss.

What kind of shared emotion is shared grief, on the shared process view? As John Michael has pointed out, ‘the expression “shared emotion” is…used to refer to a motley of phenomena that do not make up a single natural kind’ (Michael, 2016). Hence, what on one account qualifies as a shared emotion may not on another, different account. For instance, on Michael’s own view, a shared emotion requires, at a minimum, that one subject expresses an emotion, and that another subject perceives that emotion (Michael, 2011, p.361). The form of shared grief brought to light by the shared process view does not meet these conditions, and so is not a form of the kind of shared emotion with which Michael is concerned.⁷ It is however shared emotion of a kind that some other philosophers have explored, in that the shared process view belongs to the category of what León et al. have called ‘token identity accounts’ of shared emotion (León et al., 2019, p.4853). On a token identity account, a shared emotion is as Joel Krueger puts it shared in a strong sense (Krueger, 2015, p.263): a ‘numerically single emotion’ is had by more than one subject.⁸ In other words, the key feature of a token identity account of shared emotion is that on any such account, a token of some emotion type (for example, grief) is had by more than one subject. The shared process view is a token identity account because it has this key feature. As we saw in Sect. 2, a token of grief is a process of recognising and accommodating the loss of a network of identity relevant possibilities. On the shared process view, there can be tokens of grief—tokens of this very process of recognising and accommodating the loss of a network of identity-relevant possibilities—that are shared. As we have also seen, these tokens of grief are shared in that their object (a network of identity-relevant possibilities) is shared, along with the process of recognising and accommodating this object.

As Krueger points out, the kind of shared emotion described by a token identity account is especially ‘philosophically intriguing’ (2015, p.269) because it challenges the intuition that emotions can only be had by individuals. Token identity accounts are also appealing due to the straightforwardness of the sharing they involve. The sense in which a single token of an emotion type had by two subjects would count as genuinely shared is easy to see.⁹ The shared process view, as a token identity account, inherits the philosophical interest and appeal that all such accounts have. Furthermore, the shared process view is metaphysically undemanding. For instance, it does not require a demanding notion of plural subjectivity. As we have seen, a group undergoing shared grief must have a shared ‘sense of identity’. However, the shared sense of identity involved in the shared process view is quite commonplace and ordinary. It is a feature of everyday thought and talk that a group may have a ‘sense’ of being collectively kind, generous, abstemious or socially liberal for example. What this everyday thought and talk picks out is all that is meant by the group’s ‘sense of identity’. And what constitutes a group’s sense of being, for example, a socially liberal group may be no more than individual group members being disposed to behave in certain ways in certain circumstances: expressing disapproval at or punishing behaviour that diverges from the group sense of identity as socially liberal, for example. The individual members of the group may also agree, when the topic comes up, that they are socially liberal. However, as we have already seen in the case of an individual’s sense of identity, a group’s sense of identity need not be consciously articulated.

In contrast, some token identity accounts of shared emotion do incur what seem to me additional metaphysical commitments, beyond what is involved in the everyday notion of a group’s sense of identity just described. For instance, according to Schmid’s token identity account of shared emotion, a shared emotion consists in ‘plural self-awareness of a shared affective concern’ (2014b, p.10). Such shared emotions are had on this view by plural subjects with ‘plural pre-reflective self-awareness’ which, Schmid argues, ‘plays the same role in the integration of a group, as individual self-awareness plays in the integration of the individual mind’ (2014a, p.17). The shared process view incurs no commitment to plural pre-reflective awareness, nor therefore to the kind of plural subjectivity that requires it. Furthermore, the ‘sense of identity’ required by the shared process view does not play a metaphysical role analogous to (Schmid’s) plural subjectivity at all. That is, the sense of identity’s place in the shared process view is not that of integrating a group so that it can be a subject of experience analogous to an individual mind. Rather, this sense of identity comes into play in the shared process view in the explanation of the significance of a loss, and thus its appropriateness as a target for grief. Hence, the shared process view is a metaphysically undemanding token identity account.

Furthermore, the shared process view succeeds as a token identity account due, in part, to its rootedness in an account of the type of emotion—grief—to which it applies. As I said at the outset of this paper, accounts of shared emotion do not typically seek to explain the sharing of a particular emotion such as grief or fear, being framed instead as accounts of shared emotion in general. However, not attending to the nature of any type of emotion can be an obstacle to the success of a token identity account, the key feature of which is as we have seen that a token of some emotion type (for example, grief) is had by more than one subject. To see this, consider the following example, owing to Max Scheler, which has often illustrated recent philosophical discussions of shared emotion:

Two parents stand beside the dead body of a beloved child. They feel in common the “same” sorrow, the “same” anguish. It is not that A feels this sorrow and B feels it also, and moreover that they both know that they are feeling it. No, it is a feeling-in-common. A’s sorrow is in no way an “external” matter for B here, as it is e.g., for their friend, C, who joins them and commiserates “with them” or “upon their sorrow”. On the contrary, they feel it together, in the sense that they feel and experience in common, not only the same value-situation, but also the same keenness of emotion in regard to it. The sorrow, as value content, and the grief, as characterizing the functional relation thereto, are here one and identical. (Scheler, 1954, pp.12–13)

Scheler describes a brief emotional episode, something that occurs as the parents stand beside their dead child. Grief, however, is the lengthy heterogenous process described in Sect. 2. Because it is silent on grief’s processual nature, Scheler’s example does not serve its purpose well as an example of shared grief, which is what it is supposed to be.¹⁰ Therefore, it is not a good illustration of the sharing of a token of some type of emotion. In contrast to Scheler’s case, the shared process view describes a lengthy process, as opposed to a brief episode. Furthermore, the process thus described has all five of the central features of a token of individual grief, according to the lost possibilities account of grief described in Sect. 2 and summarised at the beginning of this section. Hence, there should be no question that this shared process is a token of grief. Since a token of grief is a token of some emotion type, the shared process view is a successful token identity account.

The value of close attention to grief’s nature in generating an account of shared grief can be further emphasised by reflection on a token identity account of shared emotion in general which is formulated using Scheler’s case. According to Joel Krueger, the grief of A and B in Scheler’s example is shared in that they are engaged in a kind of mutual emotion regulation which determines what it is like for them collectively. This is enabled, according to Krueger, by the fact that each is poised to off-load regulatory processes on the other, due to their bodily and spatial intimacy at the time of the episode, and their history of interaction over a longer period. Each is also vulnerable to such offloading, due to a breakdown in inhibitory processes (2015, p.272). As it is intended to be a token identity account, Krueger’s view is open to the objection that he has not explained the occurrence of a shared token emotion, but only some especially intricately causally related individual emotions, and/or an episode of mutual emotion regulation, neither of which is sufficient for a shared token emotion. Indeed, an episode of emotion regulation, even in the individual case, is not itself an emotion (grief or otherwise), though it may contribute to one. The shared process view of shared grief is not vulnerable to this ‘insufficiency problem’. Since it posits a shared token of the very same kind of process that suffices for grief in an individual, there is no question that this shared process is a token of grief, and thus that it is a token emotion that is shared. (I will return to this insufficiency problem in Sect. 4.)

I have suggested that the shared process view succeeds as a token identity account, and that token identity accounts have certain virtues. In particular, they have a distinctive kind of philosophical appeal and interest. Furthermore, the current token identity account is quite metaphysically undemanding. However, this is not to say that the only worthwhile accounts of emotional sharing are token identity accounts. As we have already seen, different philosophical accounts of shared emotion pick out quite different phenomena and so need not compete with one another. And several such accounts do not aim at identifying shared tokens of emotion types. For example, according to Szanto, a virtue of Stein’s view of emotional sharing is that whilst entailing something stronger than the possession, by several people, of emotions of the same type, it does not entail the sharing of a token emotion. On her model of emotional sharing, which Szanto develops, a group shares not an emotion, but ‘a pattern of emotional experience and regulation’ (Szanto, 2015, p.510). Similarly, Salmela, though not arguing that the ‘straightforward sharing’ involved in token identity accounts is impossible, suggests that this issue is of ‘secondary importance’ (2012, p.33), instead identifying three other forms of emotional sharing short of token identity that are ‘important phenomena in our emotional lives as social beings’ (2012, p.44). It is no part of my purpose to deny the occurrence nor the significance of these or other sorts of shared emotional phenomena. Neither do I want to deny that these other sorts of shared emotional phenomena might take forms that could with justification be described as ‘shared grief’. Nevertheless, any such ‘shared grief’ would have to be clearly distinguished from the strong, straightforward sharing of a token of grief that the shared process view yields.

Similarly, the form of shared grief identified by the shared process view certainly does not rule out, but also needs to be distinguished from, grief that takes the form of a ‘group-based’ emotion. A group-based emotion is an emotion that is felt by an individual—or by multiple individuals, separately—in virtue of their membership of a group (Menges & Kilduff, 2015). As such, group-based emotions, whilst in some sense non-individual emotional phenomena, can be had by a single subject, and in complete isolation from other members of the group. Although I will not develop the idea in detail here, it seems plausible that there could be tokens of grief that are group-based emotions in this sense, and that this form of ‘group-based grief’ could be spelled out in terms of the lost possibilities account of grief. That is, a subject might, quite alone, undergo a process of recognising and accommodating the loss of a group’s significant possibilities without, as in shared grief, engaging in a process of recognition and accommodation that is shared with members of that group. For instance, we might re-imagine the details of the father and son example in this way, so that rather than recognising and accommodating the loss of their ability to communicate easily together, one or both of them recognises and accommodates this loss separately, in a way that does not depend on any contribution from the other. This might happen, for example, if the two lose contact altogether shortly after the mother’s death, or if the father’s energies are devoted to grieving his individual losses, while the son primarily feels the losses incurred by the family. However, this acknowledgement of the feasibility of a process view of group-based grief is no threat to the view defended here. That grief can occur as a group-based emotion doesn’t show that it does not also take the more strongly shared form I have identified here. Nor does acknowledging the potential occurrence of group-based grief diminish the interest of the form of shared grief identified by the process view. That is, it does not undermine the claim that this form of shared grief is philosophically interesting, appealing, and yet metaphysically undemanding.¹¹ In the next section I will develop the shared process view further, and in a way that will allow us to see that its value is not merely intellectual.

4 Divergent feelings in shared grief

As we saw in the previous section, Scheler’s case of purported shared grief does not reflect the processual nature of this emotion. In fact, it might be argued that in side-lining the processual nature of grief this example, since it is so frequently used in recent philosophical discussions of shared emotion, has had a deleterious effect, leading to a focus on the sharing of brief, episodic emotions and the neglect of emotions that take temporally extended, processive form.¹² To introduce the main claim of this section, I first single out another way in which this example misleads, namely, in its failing to reflect the common reality of grief in groups. The parents in the example feel, as Scheler puts it, ‘the same value-situation, but also the same keenness of emotion in regard to it’: they feel, that is, the same way. Now consider, by way of contrast, the poem ‘Home Burial’, in which Robert Frost describes a quite different interaction between two people whose child has died.¹³ A man notices that his wife always pauses at the same place on the stairs and realises that she can see from there the graveyard where their child is buried. A tumult of misunderstanding follows and is left unresolved in the poem. She is distressed by his voicing his realisation, and he baffled by her distress:

He said twice over before he knew himself:

Can’t a man speak of his own child he’s lost?

It transpires that as she sees it, he cannot thus speak, by which she means that he does not grasp the significance of the loss. In particular, she is unable to comprehend how he could on the day of the child’s funeral have dug the child’s ‘little grave’, talking afterwards about ‘everyday concerns’. After he suggests (the reader winces) ‘there, you have said it all and you feel better’, she makes to run from the house, and he promises to ‘follow and bring you back by force’. Sadly, Home Burial seems to be at least as representative of the experience of a couple losing a child as Scheler’s case.

More generally, when two or more people are co-bereaved, ‘dissimilar or incongruent grief…appears to be the norm’ (Gilbert, 1996, p.275).¹⁴ In her interview study with ten families, Nadeau (1998) found that family members disagreed even about the basic facts of what had happened, such as when the death occurred and whether it had been avoidable. They diverged in how they characterised the deceased, in their views of what would happen to them after death, and in how they understood the significance of events surrounding the death. Intensity of grief can also vary amongst family members, for example, in that ‘some may see the loss as devastating; others may see it as distressing; yet others may find it a relief’ (Gilbert, 1996, p.276). The divergent nature of grief in families more generally is enough for Gilbert to insist that ‘Families do not grieve. Only individuals grieve’ (1996, p.273).

In the co-bereaved heterosexual parents of Scheler’s case and Frost’s poem differences associated with gender might also be of significance. It has been suggested that women, and thus mothers, ‘tend to be more loss oriented than fathers’ (Stroebe & Schut, 1999, p. 218). Couched in the language of the dual-process model of grief, this means that women are more likely to spend time engaged in ‘loss-oriented coping’, which is to say ‘concentration on and dealing with, processing of some aspect of’ the bereavement itself (p.212) such as ‘rumination about the deceased, about life together as it had been, and the circumstances and events surrounding the death’ (ibid.). Men on the other hand have been said to be more likely to engage in ‘restoration-oriented coping’: managing the secondary stressors associated with bereavement, including taking on new roles, attending to life changes and distraction from feelings of loss (ibid.). Even if the impact of gender on grief is overstated, there is reason to think that individuals can have somewhat different grieving styles. For example, Doka and Martin propose an ‘intuitive’ style, which favours ‘unfettered, uninhibited experience of the affective component of grief’ (2010, p.45), and an ‘instrumental’ style, characterised by strategies that are ‘cognitive and active’ (p.4). The way in which a subject of any gender grieves will fall somewhere on a spectrum with these two styles at its poles. In summary, Scheler’s case is misleading in that it presents the co-bereaved as feeling similarly, which they often do not.

Furthermore, the thought that shared emotion necessitates that subjects have similar feelings is common in philosophical discussions of shared emotion. More specifically, philosophers who have provided accounts of shared emotion of kinds that involve multiple subjects have typically thought that if A and B are sharing an emotion then they will be feeling similarly. This is true of the two token identity accounts of shared emotion (Krueger’s and Schmid’s) mentioned above, but also of some accounts that are not token identity accounts.¹⁵ For example, according to Salmela, emotional sharing requires that the experiences of the individuals concerned must be ‘synchronised’, by mechanisms such as ‘attentional deployment, emotional contagion, facial mimicry, and behavioural entrainment’, all of which imply that the participants in a shared emotion will have similar responses (2012, p.43).

Now, it is itself no objection to a view of shared grief (or shared emotion more generally) that it requires that—as in Scheler’s case—those who share grief have more similar responses to a bereavement than they often do. However, the frequent dissimilarity of the responses of the co-bereaved does mean that, on such a view, shared grief would be somewhat rare. Hence, a view that took similar feelings to be necessary for shared emotion would have to say that frequently, the co-bereaved experience only individual and not shared grief. In contrast, on the shared process view of shared grief, there is no requirement that participants in a token of shared grief have similar feelings. So, it is consistent with their undergoing shared grief that they differ in their individual grief and in how they feel more generally.¹⁶ This is for several reasons. For one thing, on the lost possibilities account of grief that the shared process view presupposes, the individual grief of the co-bereaved can be expected to differ. That’s because although A’s grief and B’s grief may be occasioned by the death the same person X, the objects of their individual griefs will be losses of different significant possibilities associated with X, which will generate some differences in feeling.

Secondly, because the shared process view acknowledges that grief, whether individual or shared, is a heterogenous process, there is no requirement that the parts of the process of shared grief are qualitatively similar. This is notably true over time in the case of individual grief. Experience e¹ had at time t¹ and e² at t² can be part of the same individual grief process of subject A, despite e¹ and e² being qualitatively dissimilar.¹⁷ Put more straightforwardly, a subject recognising and accommodating significant loss can feel differently at different times. In the case of shared grief, the parts of the process, distributed over different subjects, can also be qualitatively dissimilar at any one time. That is, it is possible in shared grief that at t³, e³ had by A, and e⁴ had by B, are also part of the same shared grief process and that e³ and e⁴ are qualitatively dissimilar. Put again straightforwardly, the claim is that the members of a group who are engaged in recognising and accommodating their loss collectively, need not feel the same way at the same time.

We can illustrate this again with the father and son from our earlier example, who are undergoing a shared process of recognising and accommodating the particular loss of their accustomed way of communicating. Frozen at some moment during the process of accommodating the fact that they have lost the possibility of easy communication, the son might at that time feel frustrated while the father feels sad or even hopeful. Nevertheless, the process they undergo will be a shared process to the extent that reaching grief’s telic end of recognition and accommodation depends on both their contributions. This does not necessitate their feeling similarly. Hence, the shared process view does not require that those who undergo a shared grief process feel similarly, and thus the frequently dissimilar feelings of the co-bereaved do not on this view entail that shared grief is rare.¹⁸

In fact, I want to suggest, the divergent feelings of the co-bereaved can in fact enable shared grief, contributing to the shared process of recognising and accommodating that which has been collectively lost. Here is another (fictional) example, call it ‘Christmas’, to illustrate this point. Suppose that family members A, B, and C have lost their relative D. Each of the family members individually encounters the loss of the possibility of spending this Christmas—the first after her death—with D, and for each of them, this lost possibility is included in the object of their individual grief experience (a network of identity-relevant possibilities). But A, B, and C diverge in their responses to these individual, parallel losses: A wants to recreate every element of a Christmas with D in her absence, whilst B would find this too painful and thinks they should celebrate Christmas together in a different way. C plans not to celebrate Christmas at all. What is most important about this example for my purposes is that A, B, and C cannot help but navigate each other’s divergent responses, which is to say, get through or past them. It is not possible for the preferences of any pair of them to be met at once, and something must be done, since time will pass, and Christmas will come. I propose that it’s precisely in the way that they navigate these divergent responses to their individual losses that the family members in this case might recognise, as well as those individual losses, an additional shared loss. For example, if they navigate their divergence intransigently, they might experience the loss of a shared possibility, that of ‘our spending Christmas with each other this year’. This could be a constituent of a significant—grief-worthy—loss because it is sufficiently close to an aspect of this family’s shared sense of identity, say, as a family that celebrates together. Having recognised the shared loss, there are numerous ways they might accommodate it: resolving to postpone family celebrations until New Year, weakening the dependence of their family sense of identity on joint celebrations, and/or by habitual mutual resentment, for instance. Recognising and accommodating this particular shared loss might then form part of this group’s token of shared grief, whilst being enabled by the group member’s dissimilar feelings.

Here is another illustration (call this one ‘Siblings’) of divergent feelings enabling shared grief, that draws on several real-life examples. Suppose now that A and B are siblings. Furthermore, they have a shared conception of their sibling relationship as supportive and strong. But when their parent, D, dies, they find themselves frequently at odds in numerous ways. For instance, they disagree over how to disburse D’s belongings or about their significance.¹⁹ And, they are inclined to tell somewhat different stories about the events surrounding the death: A does and B does not see signs of D’s presence in the sun having come out at the funeral,²⁰ whereas B does and A does not think that D’s death could have been avoided.²¹ A tends to ‘sanctify’ D, in part by portraying her final illness as a valiant battle with death. B lacks this tendency.²² The conflicts which result, and which together constitute a navigation of each other’s divergent individual grief, give rise to the recognition of various shared lost possibilities, over and above what they might have lost individually. As in the father and son example from Sect. 3, the process of accommodating these shared losses (in the sense of acquiring new expectations, projects, and habits for instance) might go in various directions. The participants in a token of shared grief may find ways to put their relationship on a new and better footing, or their relationship may come to be characterised by continued uneasiness or conflict. Either way, recognising and accommodating what they have lost together is enabled in ‘Siblings’ as it is in ‘Christmas’ by the dissimilar feelings of the individuals in the group.

This point about the role of dissimilar feelings in enabling shared grief allows us to clarify another aspect of the shared process view. We have said that the recognition of shared loss can occur as we navigate each other’s dissimilar responses. This navigation—and the recognition and accommodation of shared loss more generally—could be explicit and conscious: a matter of ‘talking it out’ and coming to an avowedly shared view of some matter, perhaps even with a therapist. But—as in the case of individual grief—this could all be largely implicit. In ‘Christmas’, A, B, and C might grasp their shared loss and its consequences during other kinds of interaction the aim of which is not to come to any shared view of, nor to accommodate, what has happened. Consider, for example, a circumstance in which A—to the dismay of B and C—starts unpacking D’s Christmas decorations and C, seeing what this means, announces that he will be travelling abroad in the last week of December. Each is likely to feel as a result that there will be no family Christmas this year, and when C quietly leaves the room and A is surprised by C’s response, a process of accommodating this loss by—let us imagine—developing new family traditions has begun, whether the participants in this process know it or not.

At this point, it might be objected that the shared process view does now fall foul of the ‘insufficiency problem’ that I said in Sect. 3 faced Krueger’s token identity account of shared emotion. He has not explained, so goes the objection, the occurrence of a shared token emotion, but only that of some especially intricately causally related individual emotions, and/or an episode of mutual emotion regulation. It might be thought that the examples just assayed of navigating one another’s divergent responses are also examples of merely causally related individual emotions, or of mutual emotion regulation. Hence, so it might be objected, the shared process view is open to this ‘insufficiency problem’ too. My response to this objection has two parts. First, recall that the scenarios described in ‘Christmas’ and ‘Siblings’ are not intended to suffice for grief. In these cases, the subjects’ navigation of each other’s divergent responses contributes to recognising and/or accommodating some aspect or part of a loss of a network of shared possibilities that constitutes the object of shared grief. For (for example) A, B, and C to undergo shared grief, the situation described in ‘Christmas’ will have to be part of an overarching process of recognising and accommodating this more expansive object. The second part of the response is to underline again the relevance of the nature of this overarching process. The shared process view will only entail that the interactions between A, B, and C in ‘Christmas’ or between A and B in ‘Siblings’ count as part of their grief (i.e., a token process of the type grief) if they are parts of a shared process the telic end point of which is recognition and accommodation of significant loss which, according to the shared process view, is precisely what a token of shared grief is. The problem for Krueger is not that causally interacting emotions or episodes of mutual emotion regulation cannot contribute to shared emotion, and to shared grief in particular. Rather, the problem is that the occurrence of such phenomena does not suffice for a shared token of grief (nor, without further details, for the sharing of any token emotion) and that some story needs to be told about how such phenomena contribute to a shared token emotion, given that they do not suffice for it. The shared process view provides such a story and so does not face this same problem.

One final point of clarification relates to the relationship between individual and shared grief on the shared process view. There is no reason in principle—although there may be reasons relating to the circumstances of specific groups or specific losses—why those who grieve together in this way must be at similar ‘stages’ with respect to their individual grief. That is to say, A and B may both be involved in a process of recognising and accommodating their shared loss, even though A has got about as far as she is going to when it comes to recognising and accommodating her own loss, whilst B is still numb with disbelief. In fact, this kind of divergence in feeling may be one of those the navigation of which constitutes part of A and B’s shared grief. And, A and B’s shared grief may outlive or may not outlive A’s grief, or B’s grief. In one imagined version of this scenario, while A’s individual grief has become merely intermittent, A and B’s shared grief continues, as they struggle to accommodate their collective losses, including the loss of the ability they previously had to empathise with each other’s responses.²³

We are now able to see the value that the shared process view might have, beyond the intellectual value identified in Sect. 3. Part of the potential value of a philosophical account of grief lies in providing grieving people, and those who support them, with a way of understanding what is happening to them. This is especially salient in the case of grief which may not be recognised as such by its subject, such as grief over an ‘ambiguous’ loss (Boss, 1999). Core examples of ambiguous loss include kidnapping, where there is uncertainty about whether the person will return, and advanced dementia, where the uncertainty relates to whether the person is still present. As Dwyer puts it, unrecognised grief can have a ‘chameleon effect’, ‘affecting one’s thoughts, feelings and behaviours without being recognised as such’ (Dwyer, 2020, p.110). As we have seen, on the shared process view, those who are grieving together can feel quite differently. This in turn opens the possibility that subjects—such as the suffering parents in Frost’s poem—for whom this is an apt description are undergoing a form of shared grief. Hence, the shared process view may allow those who have suffered a shared loss to recognise that despite their differences, what they are undergoing is, indeed, shared grief.

Understanding not only that what is happening to one is shared grief but what shared grief is has additional value. The lost possibilities account of individual grief allows one to appreciate that one is undergoing a process of recognising and accommodating all those significant losses of possibility that have depended on ‘your person’. This might help someone to better tolerate the wide-ranging emotional and cognitive upheaval of grief, which can otherwise be distressingly puzzling. It might also contribute to helping a grieving person, or someone supporting them, to guide the process towards happier, healthier, more equanimous modes of accommodation. Similarly in the shared case, and here it can be especially useful to recognise that interpersonal conflict and tension amongst the co-bereaved might be an aspect of their shared grieving. One study found that the perception that one’s partner’s grief over the death of a child differs from one’s own is correlated with lower relationship satisfaction, an effect that increases over time (Buyukcan-Tetik et al., 2017). And, in their interviews with bereaved parents, Gilbert and Smart (1992) discovered that, as Gilbert reports in her later paper, ‘the expectation that bereaved couples would grieve in the same way…added to the stress they felt’ (Gilbert, 1996, p.276). If a family or other group recognises that alongside their individual griefs they have shared losses to be recognised and accommodated in part in the way they navigate each other’s differing responses, some of this additional distress may be better tolerated, or avoided in the guidance of the process.²⁴ In the next and final section, I identify two broader consequences of this view. The first pertains to its scope, the other to the light it sheds on the necessary conditions for shared emotion, or shared emotion of certain kinds.

5 Beyond shared grief

As I said at the beginning, accounts of shared emotion are typically intended to be quite general, covering the sharing of all emotions, rather than the sharing of some specific emotion type. I hope to have illustrated that it can be profitable to narrow one’s scope, in that doing so has in this case yielded an account with virtues of various kinds. Before suggesting that the account can nevertheless be extended to some other shared emotions, note that the account should be of interest even if it applies only to the grief that follows a bereavement. Such grief is not only extremely common but also, sometimes, a profoundly life-altering emotion, and so there is value in an account of it, even if this account sheds no light on other emotion types. Another reason to welcome some narrowness of scope is that, as mentioned above, it seems likely that we need a variety of models of sharing if we are to understand the diverse phenomena described by philosophers as shared emotion. In particular, we should seek a quite different view to the one defended here if we are to think of short-lived episodic emotions as shared. Nevertheless, I think the account I’ve offered might be extended to a quite broad range of cases beyond the examples of co-bereaved family members focussed on here.

First, there is reason to believe that there can be grief over losses other than deaths. This includes, for example, involuntary childlessness (Ratcliffe & Richardson, 2023), losses associated with illness and injury (Byrne, 2022; Ratcliffe and Cole, 2022), losses incurred due to pandemic-related social restrictions (Richardson & Millar, 2022), and losses associated with environmental destruction (Fernandez Velasco, 2024). If there can be grief over such ‘non-death losses’—and the lost possibilities account of grief makes room for it—then more of our emotional lives, individual and shared, may be constituted by grief than we might otherwise have thought. In consequence, there would be a broader range of experience to which the shared process view might directly apply.

Second, my focus here has been on the shared grief that might be had amongst small and intimately related groups such as families or groups of close friends. Talk of collective grief is very often (perhaps more often than not) concerned with much larger groups than a couple or family.²⁵ An institution, nation, or city may be said to grieve deaths caused by a natural disaster or terrorist attack, or to grieve the death of a person of importance. Large-scale collective grief over non-death losses is also sometimes given an explanatory role in, for example, explaining political divisions (for example, Hochschild, 2018). Some such talk is clearly to be taken with a pinch of salt, e.g., as hyperbole. And some might pick out parallel group-based grieving, in which members of a group each, individually, grieve for the losses incurred by the group. However, the shared process view has the potential to ‘scale up’ to large groups. Even a large group might have the relevant kind of sense of identity and its members a need to navigate each other’s divergent responses to a shared loss to recognise and accommodate it.

Thirdly, the kind of account offered here, on which a shared emotion is the sharing of a heterogenous process, could extend to emotion types other than grief. It might extend to other temporally extended emotional processes such as (amongst those that are easily named) falling in or out of love or of friendship, or prolonged forms of shame, guilt, regret, or resentment. It seems likely that these, too, can be heterogenous processes whereby a subject stands, over time, in some specified (and perhaps multifaceted) relation to a proprietary object. In principle then, as in the case of shared grief, such processes might be shared in a way analogous to the way in which I have argued that grief can be shared. For instance, there may be a form of shared regret whereby a group recognises and comes to terms with the fact that they have ‘acted in a way that has deprived them of something they care for, even though they could have chosen to act otherwise’ (Price, 2020, p.147). Exploring the details of any such account will have to wait for another time. Philosophy of emotion—both individual and shared emotion—tends to be preoccupied with short-lived emotional episodes—say, of fear or anger. I do not propose that the sharing of such episodic emotions could be modelled by the shared process view of shared grief, and I have done little, here, that speaks to the plausibility of any account of the sharing of episodic emotions. However, it is possible that a more profitable approach to shared and collective emotion would shift its focus from episodic emotions to processive ones.

As well as, potentially, applying to a range of emotions beyond bereavement grief, the account I have offered reveals something about the necessary conditions of shared emotion. First, as we have seen, it should not be thought, as it sometimes is, that shared emotions that involve multiple subjects require that those subjects feel similarly. A second point can be approached via a potential objection. As we have already had cause to acknowledge, the term ‘shared emotion’ picks out multiple phenomena, rather than a single natural kind. However, it might well be urged that I have overlooked something central to the very idea not only of shared emotion, or many types thereof, but shared mentality more generally: a sense of ‘we-ness’ or ‘feeling of togetherness’ (Zahavi, 2015, p.90). As Thonhauser puts it ‘genuinely shared emotion requires that we share the same emotion’, which means that ‘we do not feel it as mine and yours, but as ours’ (Thonhauser, 2018). This view has been quite widely accepted by philosophers of shared emotion, and not only by those who defend the strongest token identity accounts. For example, according to Szanto, without an experience of the emotion as being ours ‘one would have a hard time distinguishing emotional sharing from simply being causally affected by other’s emotions…or from automatically mimicking those emotions, i.e., from emotional contagion’ (2015, p.508). If this were right, it would generate a problem for the shared process view of shared grief, since there is nothing in the shared process view that entails that the subjects of a shared grief process feel their shared grief as ‘ours’.

One somewhat conciliatory response to this worry is to allow that shared grief might, perhaps even typically, involve a sense of ‘we-ness’, even though it is not necessary that it does. For example, a sense of we-ness may emerge when the subjects of a shared grief process are less divergent in their grief, or when they navigate their divergent responses explicitly. Furthermore, Thonhauser elsewhere (2022) allows that there can in fact be a kind of genuinely shared emotion without a sense of we-ness. Shared emotion can take the form of what he calls ‘emotional segregation’ when, for example, it is ‘too painful’ to experience the emotion together. In such circumstances, subjects do not acknowledge the collective nature of an experience that is nevertheless collective. On his view though, they could acknowledge it: ‘thorough reflection…would reveal that the collective emotion has been shared all along’ (2022, p.9). I can allow that too, in the case of shared grief without we-ness. Thorough reflection could reveal to a group with no sense that they are experiencing grief as ‘ours’ that they have in fact been undergoing a shared process of recognising and accommodating loss.

Nevertheless, a less conciliatory response ought also to be part of the picture, and this is the thought that it might well be a mistake to take the phenomenology of we-ness to be central to shared emotion involving multiple subjects, even when that sharing is of the strongest possible kind. We saw in Sect. 3 that one way of motivating the centrality of we-ness is Schmid’s claim that it serves to integrate a group mind (or plural subject) in the same way that self-awareness integrates an individual mind. But that self-awareness plays this role in the individual case is controversial.²⁶ Furthermore, we have the resources to respond to Szanto’s claim that a sense of we-ness is required to distinguish genuinely shared emotion from a situation in which A’s and B’s emotions causally interact or from emotional contagion. On the shared process view, A and B’s shared grief is a genuinely shared token of grief, distinct from A’s individual grief and B’s, because it has a distinct shared object, and is a distinct, shared process of recognising and accommodating this distinct shared object, involving both A and B. Hence, it is not clear what a sense of we-ness is needed in this case to explain. That is not to deny any role to the sense of we-ness in work on shared emotion: for instance, it is sometimes a central explanandum in discussions of shared emotion (for example, Zahavi, 2015), from which its exclusion would thus make little sense. But we ought not to confuse the compellingness of the project of explaining the sense of we-ness when it occurs with the idea that we-ness is the ‘mark of the shared’. To do that, is to make some shared emotion invisible, including the kind of shared grief I have identified and explained in this paper.

6 Conclusion

It can, as I hope this paper shows, be useful to explore a way in which just one kind of emotion—in this case, grief—can occur in shared form. For to proceed in that way, one must attend to the nature of the emotion in question, which has allowed us to identify a form of shared grief on which multiple subjects can undergo a single token of that emotion. This form of shared grief is philosophically interesting, in running counter to the idea that an emotion is something had by an individual, but metaphysically undemanding. The account also has potential practical use. Furthermore, though the account is of shared grief (rather than shared emotion more generally) and of a particular form that shared grief might take (rather than everything that might be called shared grief) it has as we have seen some broader consequences for theorising about shared emotion. It casts into doubt the helpfulness of a certain, frequently-used example (Scheler’s) as well as—more importantly—the centrality to shared emotion of a feeling of ‘we-ness’. Finally, it might be adapted to bring to light shared forms of some other lengthy and processive emotions, thus shifting attention away from those brief episodic emotions that are more often the concern of philosophers of shared emotion, and emotion more generally.