Sensed presence without sensory qualities: a phenomenological study of bereavement hallucinations

This paper addresses the nature of sensed-presence experiences that are commonplace among the bereaved and occur cross-culturally. Although these experiences are often labelled ‘‘bereavement hallucinations’’, it is unclear what they consist of. Some seem to involve sensory experiences in one or more modalities, while others involve a non-specific feeling or sense of presence. I focus on a puzzle concerning the latter: it is unclear how an experience of someone’s presence could arise without a more specific sensory content. I suggest that at least some of these experiences consist in a dynamic and non-localized experience of significant and salient possibilities. This can amount to the sense of currently relating to a particular individual and, by implication, a sense of that person’s presence. Where an experience of this kind also includes sensory qualities, they are inessential to the sense of relatedness and perhaps symptomatic of it.


Introduction
The term "bereavement hallucination" refers to a perceptual or perception-like experience of someone who has died, usually a partner, family member, or close friend. Such experiences are sometimes described in terms of specific sensory modalities: one might see, hear, or feel the touch of the deceased. However, the most common form of experience is a non-specific sense or feeling of presence. Conceptualizing these phenomena might seem straightforward enough: they are non-veridical experiences in one or more sensory modalities, which resemble -to varying degrees-veridical perceptual experiences of a particular individual. In other words, they conform to an orthodox conception of hallucination: an experience of p that, although inappropriately caused, resembles or is even identical to a veridical experience of p.
Perhaps this captures some of the relevant phenomena. Nevertheless, it remains unclear what a non-specific feeling or sense of presence consists of. We are presented with a puzzle. If first-person accounts are taken at face-value, the content of the experience is quite specific: the presence of a particular person. However, this does not originate in more concrete sensory contents, such as hearing or seeing that person. What, then, does it involve? Of course, one could insist that there is in fact a sensory content, which is either sufficient to specify a given individual or at least disposes one to infer the presence of that individual. It could be added that people do not report this because they either lack reflective access to it or find it hard to describe. However, this paper will adopt a different approach. I will show how it is indeed possible to experience a particular person in a way that does not depend on sensory contents such as visual images, sounds, smells, or tactual feelings. There is a distinctive way of being affected, which can be further analysed in terms of the experience of possibilities. This amounts to the sense of being with a particular person. I will conclude that, although it is unlikely that all perception-like experiences of the deceased take this form, some of them plausibly do.
2 The presence of the dead Bereavement hallucinations are reported to be commonplace and, in most instances, benign. An early and widely cited study by Rees (1971) involved 227 widows and 66 widowers in Wales. Nearly half of those interviewed reported experiencing the deceased. The most common experience consisted simply in "feeling the presence of the dead spouse" and was reported by 39% of respondents. This was followed by visual hallucinations (14%), auditory verbal hallucinations (13.3%), and then tactile hallucinations (2.7%). Rees regards the majority of these occurrences as "normal and helpful accompaniments of widowhood" (1971, pp.37-8). More recent studies report similar findings (e.g. Bennett and Bennett 2000;Keen et al. 2013;Castelnovo et al. 2015). There are consistent references in the empirical literature to a sense or feeling of proximity that does not originate in a more specific sensory content, at least not one that is readily identified. For example, Keen et al. (2013, p.390) write that the most common type of experience is "a sense or feeling that the deceased person is close by without experiencing them in any sensory modality", while Longman et al. (1988, p.44) note that "an overwhelming sense of presence was often expressed indicating that the subjects felt they were not alone". Steffen and Coyle (2012, p.35) similarly state that "people report that they can somehow sense or feel the physical proximity of the deceased loved one". However, descriptions of the experience tend to be cursory and it is not at all clear what it actually consists of (Castelnovo et al. 2015, p. 271).
Given this lack of clarity, there is the methodological concern that studies may be using the same terms to address different phenomena, without making those differences explicit (Datson and Marwit 1997, p.133). In addition, where different terms are used, it is not always clear whether they have a common or overlapping referent. For instance, Dannenbaum and Kinnier (2009) consider "imaginal relationships" with the dead, rather than hallucinations or sensed presence experiences. Such relationships may encompass some bereavement hallucinations (those that involve an experience of connection), but not others (which are more detached). There are also questions regarding the nature and extent of cultural variability. According to Keen et al. (2013), up to 90% of bereaved spouses in some cultures experience the presence of the deceased, but prevalence varies considerably. Furthermore, differing cultural attitudes influence how people interpret and respond to their own and others' experiences, in ways that may well affect how those experiences unfold. For instance, whether or not a sensed-presence experience is positive and ultimately beneficial may hinge, to a large extent, on whether it conforms to cultural norms and whether it is interpreted in accordance with culturally accepted practices (Steffen and Coyle 2012). Given the degree and types of potential variation, it is debatable whether a core, underlying experience can be identified cross-culturally and, if it can, what that experience consists in. 1 Differing interpretive frameworks are also explicitly or implicitly adopted by researchers. Consider the term "hallucination", which suggests a discrete experiential content that is aberrant in failing to track what is actually the case. For example, Castelnovo et al. (2015, p.266) define what they term "post-bereavement hallucinatory experiences" as "abnormal sensory experiences that are frequently reported by bereaved individuals without a history of mental disorder". However, this is in tension with first-person interpretations, which often regard the experience as valuable and / or as a source of information, integrating it into a wider-ranging account of the world. Moreover, it is not always clear what the criteria are for deeming an experience normal or abnormal. A non-veridical experience might well be a normal reaction to certain events, meaning a reaction that is not only commonplace but also situationally appropriate according to some criterion. One option is to construe abnormality in specifically epistemic terms: a type of experience is abnormal when it is invariably misleading. However, matters are not so straightforward, as an experience with non-veridical elements could still serve to reveal certain truths about oneself, one's relationships, and one's values.
An alternative way of conceptualizing these phenomena is suggested by continuing bonds approaches, which maintain that grief does not simply culminate in letting go and severing one's ties with the deceased. Instead, most people in most cultures continue to relate to the deceased in a variety of ways. Relationships are reorganized rather than altogether lost and may continue to play important roles in people's lives (e.g. Klass et al. 1996;Klass and Steffen 2018). Which perspective we adopt will have a bearing on how sensed presence experiences are conceived of, including whether or not they are deemed aberrant or pathological (Sanger 2009). In particular, I want to emphasize the difference between a hallucination (construed as a perceptual experience of p that arises in the absence of p) and a sense of connection (where the emphasis is on relating to someone rather than merely experiencing their presence). 2 I will suggest that we can better understand what certain sensed-presence experiences involve by considering their relational phenomenology, rather than seeking some component of the experience that adds up to an elusive quality of presence.
I will adopt a provisional distinction between two broad and overlapping categories of experience: those with contents traceable to one or more sensory modalities and those involving a less specific sense or feeling of presence. 3 Some interpretive caution is needed here, given the likelihood that the latter are sometimes described in terms of the former. More generally, terms such as "see", "hear", and "touch" are used in a range of ways that do not refer to sensory phenomena: "let's return to the point you just touched on"; "I hear you loud and clear"; "I see what you mean". It could be that people also resort to such terms when attempting to convey an unfamiliar form of experience (which does not appear to originate in a particular sensory modality) in a more familiar way. Here, I am concerned with those experiences that do consist in a non-specific sense of presence, regardless of how they might be described.

Senses of presence
In addressing the nature of sensed-presence experiences, we need to consider what is actually meant by "presence" in this context. Presence might be thought of as a straightforward matter of spatiotemporal coincidence, as suggested by definitions such as "feeling that the deceased person is close". Thus, several other people would be present to me when we are all doing our shopping in the same supermarket at the same time. However, what counts as sufficiently close proximity for presence depends on the situation. "We were both present at the concert in Wembley stadium" allows for greater spatial distance between two parties than "we were both present at the job interview". So, in addition to spatiotemporal distance, presence seems to hinge on having access to a common object of perception or attention, such as an interview or a concert. The requirement of shared access is more apparent in examples such as "twenty people, who were present at the scene of the accident, have provided statements", where the point is not just that they were there but also that they witnessed something. Actively engaging with something can also be a factor. For example, "she was present at the exam" suggests participation in the exam. Standing just inside the door of the exam room for two minutes would not suffice. Being present may further involve acting in accordance with certain norms, as in "his presence at the event was required". Other uses of the term relate to more diffuse ways in which a person engages with and influences his social surroundings, as in "his presence was toxic in every way". The presence of something can also involve being affected in some manner by it or somehow relating to it: "she was overwhelmed by the presence of God".
"Presence" thus takes on a range of different connotations, and geographical proximity is not always sufficient or even necessary. It not sufficient when one stands at the door of an exam room and it is not necessary when one is present at a Skype meeting. Hence, when one is said to experience the presence of the deceased, it cannot be assumed that close physical proximity is most central to the experience in question. Instead, I will emphasize a form of relatedness or connection. In conjunction with this, it is important to note that what is present is not just any person but this particular person. Experiencing someone in a distinctively personal way involves recognizing an essential particularity that is lacking from our encounters with other types of entities. Suppose you somehow sense the presence of a coffee cup. Perhaps it is a particular coffee cup -the one your grandmother gave you. On the other hand, the question "which coffee cup did you experience?" may not have an answer. It could be just any old coffee cup. The question "which coffee cup?" might be met with bemusement, as might "which sausage roll", "which brick", "which tadpole", "which coaster", and "which paving stone". 4 However, in the case of a person, there is always the further question of who that person is. Granted, there are many cases where the question "which x?" is appropriate for an impersonal entity: it matters which house we are heading to, which plate we are eating from, which car we own, and so forth. However, this applies only to certain specific members of those kinds. Where other people are concerned, there is always the potential to engage with someone as a particular person, as a who.
It might be thought that experiencing a particular person just amounts to having an experience with a sensory content specific enough to pick out that individual or, at least, a sensory content that is usually associated with her. This raises difficult philosophical issues concerning the nature and scope of sensory perceptual content. Perhaps that content is itself rich enough to constitute a sense of being in someone's presence. Alternatively, an initial sensory perceptual experience might be supplemented by inference or interpretation. Another option is to think of the sensory experience itself as having a dynamic structure, such that a rich interpersonal experience somehow crystallizes out of an initial experience with a less determinate content, as opposed to being an inference or interpretation that follows it: "While the meaning-making part could be seen as a retrospective attribution, it could be viewed as a translation of the initial sense, an unpacking of its implicational meaning" (Steffen and Coyle 2011, p.593).
All of these alternatives are consistent with the assumption that either sensory content alone or sensory content supplemented by something else suffices to specify a given individual. However, I will suggest instead that the sense of presence has a kind of relational structure, which constitutes the sense of being with a particular person. This cannot be captured in terms of sensory experience, regardless of whether we appeal to rich perceptual content, inference, interpretation, or crystallization. The relevant phenomenology is thus obscured by the term "hallucination", which suggests a non-veridical, sensory experience of something or other.
How could one experience the presence of a particular person, without that experience originating in sensory experiences that indicate more specific properties? To address this question, I want to emphasize the distinction between being in close proximity to another person and being with her. The latter is not exhausted by the experience of a specific object of perception occupying a precise location. What, then, does it consist of? We can get the beginnings of an answer by drawing on work by the phenomenologist Maurice Merleau-Ponty. He suggests that a kind of presence is experienced when the bereaved person encounters the world in a habitual way that continues to presuppose his relationship with the deceased (Merleau-Ponty 1945/2012, Part One, Chapter I; Ratcliffe 2019). To explain further, how we experience our surroundings reflects both our capacities and our concerns. As I look around my office, the computer screen and keyboard appear salient and significant insofar as they reflect my current writing project, along with the bodily capacities that I take for granted in pursuing that project. More generally, the surrounding world, as we experience it, is imbued with a range of different significant possibilities, which are integrated in ways that reflect the coherence of our various projects. Regardless of whether or not such possibilities are to be regarded as specifically perceptual in nature, it is at least plausible to maintain that they are integral to pre-reflective experience of one's surroundings, rather than being products of inference from experience. This is illustrated by those occasions when something in particular or the world in general appears strangely devoid of its usual significant possibilities, as sometimes happens when we are ill, jetlagged, hungover, or undergoing emotional upheaval (Ratcliffe 2008(Ratcliffe , 2015.
Our projects, Merleau-Ponty further indicates, can involve a particular person in such a way that the possibilities we experience imply his actual or potential presence. For instance, if a project is geared around doing something for him, the various activities that comprise it are only intelligible in relation to him. The same applies to some instances where I pursue a project with him. The point applies not only to goaloriented projects, but also to habitual routines and pastimes. Consider how one's home might be experienced in ways that implicate a partner: the sofa appears as something for us to sit on; these are the films we enjoy watching together; this is our bed. The suggestion is that, when the partner dies, a habitual world may remain largely intact for a time, despite an explicit, propositional recognition of the death. Things still appear as they did before and therefore imply the partner's potential or actual presence, in a diffuse, non-localized way that cannot be attributed to a more specific inventory of perceived properties.
To elaborate on Merleau-Ponty's discussion, this form of experience could be fleeting or enduring and localized or non-localized. For instance, a non-localized sense of presence might be punctuated by periods of absence, as when one encounters a scene in light of the person's potential appearance, only to realize that she will never appear again. The dynamic can be construed in terms of an ongoing tension between two worlds, a past that one continues to inhabit and a present that is drained of meaning and seems curiously distant (Fuchs 2018). There may also be tensions between pockets of the world that continue to allow for the person and other parts of it that do not. Consider this passage, from Simone de Beauvoir's account of her mother's illness and death: As we looked at her straw bag, filled with balls of wool and an unfinished piece of knitting, and at her blotting-pad, her scissors, her thimble, emotion rose up and drowned us. Everyone knows the power of things: life is solidified in them, more immediately present than in any one of its instants. They lay there on my table, orphaned, useless, waiting to turn into rubbish or to find another identity … . (1965, p.98) The contents of the bag remain imbued with possibilities that relate to her mother, even though other parts of her surroundings are not. It therefore appears strange, somehow out of context. Hence, what we have here is not a straightforward experience of presence or absence.
Experiences of these kinds are not specific to bereavement; they encompass all manner of scenarios where habitual ways of experiencing, thinking, and acting endure to some degree, despite the explicit recognition that something has changed. Examples range from the mundane, as when one changes the password for one's email account or reorganizes the kitchen utensils, to the extreme, as when one loses a job to which one has been dedicated for many years.
It could be argued that some of those experiences labelled as "hallucinations" are best thought of in this way. For instance, participants in a study by Steffen and Coyle (2011, p.586) described a range of sensed-presence experiences, including "feeling the deceased is standing close by", "feeling the deceased is walking alongside", and "feeling the deceased is around". These, one might suggest, do not involve engaging in some activity plus experiencing the deceased nearby, but engaging in activities and experiencing one's surroundings in ways one did when with that person. The person is not simply there, in the guise of a discrete object of perception, but implicated in a nonlocalized way by what is there.
However, the phenomenon identified by Merleau-Ponty is not quite what I am looking for. He suggests that preserving the habitual world is a way of not confronting the loss. It involves the avoidance of situations where one might expect to encounter the person who has died and have one's expectations negated. Hence, this emphasis on one's habitual world does not accommodate the manner in which that person's presence can be a conspicuous aspect of the experience. Feeling that someone is walking alongside suggests something more than the preservation of practices that implicate the deceased. It is not merely that, as one walks, the surrounding world is imbued with possibilities that somehow presuppose a particular person. A personal presence is itself a salient part of the experience, amounting to more than the forgetting of absence. Furthermore, the experience of walking with someone is dynamic and changeable. As one walks, the relationship shapes how one's surroundings appear, in ways that vary from moment to moment. Sensedpresence experiences of this kind can also involve an appreciation of ongoing "communication" and "mutuality" (Steffen and Coyle 2011, p.589). They are thus importantly different from a project that remains frozen in time after someone's death. Unlike such a project, a companion with whom one walks can imbue the surrounding world with new and changing possibilities; things seem more alive with her than when alone. (This is what is sometimes meant when we talk of enjoying someone's company.) What should be retained from Merleau-Ponty's account, I suggest, is his emphasis on the experience of significant possibilities. It might well be that some bereavement hallucinations and sensed presence experiences consist wholly or partly in nonveridical sensory experiences. Others, however, involve a dynamic, self-affecting experience of possibilities, something that is to be identified with the sense of being with a particular person. In these latter cases, sensory experiences associated with one or more externally directed sensory modalities might also be present. However, it would be wrong to assume that these are principally responsible for the sense of being with the person. They could be accompaniments to a core experience that is different in kind or sensory imaginings elicited by it.
More generally, interpersonal experience is not simply a matter of perceiving certain physical properties and inferring the presence of an internal mental life lurking behind them. Again, it is informative to consult Merleau-Ponty's work, where a recurring theme is that we encounter the experiences of others as inherent in their expressions, gestures, and goal-directed actions. He maintains that we are able to do so because these activities always point to more than what is currently revealed to sensory perception. What they point to is not something currently hidden behind them, inside a head, the presence of which is inferred on the basis of what we do have access to. Rather, we experience gestures, expressions, and actions as infused with possibilities, including relational possibilities that are specifically personal. The experienced mental life of another person is encountered as a dynamic, unfolding configuration of possibilities. This is suggested by the following passage, which emphasizes how the sense of another person's presence cannot be accounted for in terms of specific combinations of perceived properties or something distinct from those properties that is itself perceptually inaccessible: … .the other is never present face to face. Even when, in the heat of discussion, I directly confront my adversary, it is not in that violent face with its grimace, or even in that voice travelling toward me, that the intention which reaches me is to be found. The adversary is never quite localized; his voice, his gesticulations, his twitches, are only effects, a sort of stage effect, a ceremony. [ … .] One must believe that there was someone over there. But where? Not in that overstrained voice, not in that face lined like any well-worn object. Certainly not behind that setup: I know quite well that back there there is only ''darkness crammed with organs.'' (Merleau-Ponty 1973, p.133) In Merleau-Ponty's terms, to encounter someone in a specifically personal way is to experience a distinctive type of style, an organized way in which possibilities interrelate and unfold. This, he adds, essentially involves being affected in a certain way: to experience someone else's possibilities is to have one's own possibilities altered in a subtle or more dramatic fashion (e.g. Merleau-Ponty 1945/2012. This same theme appears in the work of various other phenomenologists, albeit expressed in different ways and with slightly different emphases. The best-known example is Sartre (1943Sartre ( /1989, who construes our most fundamental sense of ''the Other'' in terms of being affected by their presence in a pre-reflective, bodily manner. With this, the world is no longer organized in terms of one's own possibilities and one instead becomes an object for them. De Beauvoir (1947/2018) similarly emphasizes the self-affecting experience of others' possibilities. However, in contrast to Sartre's (1943Sartre's ( /1989 account, she maintains that a sense of others' freedom is essential to sustaining one's own experience of an open future. We also find a complementary formulation in the work of Løgstrup (1956Løgstrup ( /1997, who suggests that we are unavoidably responsible for others, insofar as any dealings we have with another person will unavoidably affect that person, in such a way as to "determine the scope and hue of his or her world", making it "large or small, bright or drab, rich or dull, threatening or secure" (p.18). The effect he describes can similarly be construed in terms of reshaping the possibilities to which a person is open. Løgstrup goes so far as to say that relying exclusively on a determinate, image-like representation of a person, rather than letting that person "emerge through words, deeds, and conduct", amounts to a "denial of life" (p.14). There is also empirical support for such a view. For instance, it is arguable that various forms of psychopathology are most plausibly interpreted as involving alterations in how one's sense of the possible is interpersonally regulated (Ratcliffe 2015(Ratcliffe , 2017. 5 Hence, I maintain that there is a type or aspect of interpersonal experience that is neither exhausted by the contents of sensory experience nor dependent upon it in the form of an inference, interpretation, or embellishment. This accommodates not only the general experience of being with another person, but also that of being with a particular person. In short, things look different, depending on how we relate to specific individuals. The sense of being with that person involves being affected in a certain way by her unique style. A coffee cup, in contrast, is not self-affecting in a personal way and lacks this essential particularity. Now, suppose some experiences of sensed-presence takes this form. That would explain why they are difficult to pin down and describe. They do not depend on specific sensory contents, and neither do they involve being able to point to particular properties or locations. Despite this, there remains an unequivocal sense of being with a specific individual. In some cases, this might be accompanied by more concrete sensory experiences of the person. After all, it seems likely that an ongoing sense of relating to someone in a patterned way will evoke sensory imaginings, as well as memories with sensory contents. Furthermore, these may be confused with current sensory experience (by the subject of experience or by others who attempt to interpret firstperson descriptions of experience), given their close association with the sense of currently relating to someone.
If this is right, the relevant experience is not only distinct from a more concrete, determinate experience of a person; it can also be in tension with it, given that a sense of the possible is essentially indeterminate. It is therefore quite unlike an orthodox hallucination. Suppose one somehow managed to concoct a perceptual and cognitive representation of the dead person that consisted in an exhaustive inventory of all those properties associated with her; absolutely nothing is left behind. In fact, this would fail to capture her distinctive style. To experience that style is to be affected in a unique way. And this requires being open to possibilities that are not fully anticipated, to something that is not fully captured by the contents of one's own mental states. In order to make this contrast clearer, I will now turn to a concrete example.

Losing her twice
In his memoir, A Grief Observed, C. S. Lewis charts the grief he experienced following the death of his wife. His account describes, in detail, the pain of losing her again by failing to retain her in memory. This is contrasted with a sense of her presence that later returns. Although what Lewis describes is not localized, episodic, or pronounced to the extent that certain sensed-presence experiences are, his account still serves to illustrate the difference between a sense of personal presence and an orthodox hallucination.
Lewis describes how, in his sorrow, he sought to preserve his wife in memory and not let her slip away. Yet, his doing so prevents him from experiencing anything of her. She ends up being replaced by something that appears to him as his own creation: I am thinking about her nearly always. Thinking of the H. factsreal words, looks, laughs, and actions of hers. But it is my own mind that selects and groups them. Already, less than a month after her death, I can feel the slow, insidious beginning of a process that will make the H. I think of into a more and more imaginary woman. Founded on fact, no doubt, I shall put in nothing fictitious (or I hope I shan't). But won't the composition inevitably become more and more my own? The reality is no longer there to check me, to pull me up short, as the real H. so often did, so unexpectedly, by being so thoroughly herself and not me. The most precious gift that marriage gave me was this constant impact of something very close and intimate yet all the time unmistakably other, resistant in a word, real. (Lewis 1966, p.17) What Lewis is saying here is that, however many properties are held in memory, and however vividly a person's properties are imagined, a certain kind of recollection is self-defeating. It will always culminate in eradication of the person's distinctiveness. To experience his wife, to be with and feel connected to her, was to be affected in a certain way. It was to anticipate and experience a transcending of his own possibilities, including his own imaginative efforts. Without this, the sense of her is lost; what remains is experienced as integral to him. Lewis thus writes,"The rough, sharp, cleansing tang of her otherness is gone. What pitiable cant to say 'She will live forever in my memory!' Live? That is exactly what she won't do" (pp.18-19). What is lacking is a distinctive style of resistance to his own expectations and imaginative efforts. To experience her is to feel her effect on his world, something that essentially involves an openness to possibilities that are not of his own making.
Thus, not knowing everything about another person is not a contingent feature of one's relationship with her, an epistemic shortcoming. To encounter someone in a personal way is to experience her as inevitably surpassing one's determinate representations of her. To know everything about her would be to cease experiencing and relating to her in that way. Later, as Lewis's sorrow lessens, so does the intense "longing" associated with his self-defeating attempts to hold onto his wife in memory (p.40). With this, there is a renewed sense of connection, involving a different kind of presence: And suddenly at the very moment when, so far, I mourned H. least, I remembered her best. Indeed it was something (almost) better than memory; an instantaneous, unanswerable impression. To say it was like a meeting would be going too far. Yet there was that in it which tempts one to use those words. It was as if the lifting of the sorrow removed a barrier. (p.39) This "impression" is something that Lewis contrasts with a determinate representation of his wife, derived from memory, imagination, and / or sensory perceptual experience. It involves, I suggest, an interaction between one's sense of the possible and what Merleau-Ponty would call another person's distinctive style. This need not be restricted to a particular location, project, or pastime. A relationship with another person, living or dead, can pervade all aspects of one's life. As Lewis ceases to worry about imposing a false memory and replacing his wife with his image of her, "she seems to meet me everywhere", not as an apparition with determinate properties, occupying a particular location, but as "a sort of unobtrusive but massive sense that she is, just as much as ever, a fact to be taken into account" (p.44). 6 What Lewis describes is reminiscent of the Myth of Orpheus and Eurydice. As Orpheus returns from the Underworld with Eurydice, he must walk ahead and not look back. Yet he increasingly doubts her presence and eventually turns in order to acquire the determinate sensory experience he currently lacks. As he does so, her shadow returns to the Underworld, this time irrevocably. Lewis's narrative runs the other way. The image of his wife eradicates her distinctiveness and she is lost, returning only when he ceases in his attempts to preserve her in memory. In both cases, an image not only fails to capture the other person; it also renders her inaccessible.
Interestingly, Lewis describes his grief as inseparable from his relationship with God. Loss of connection with his wife coincides with the loss of a faith that now strikes him as superficial and naive. Similarly, a renewed connection with his wife spells the 6 One influential approach to sensed-presence experiences and bereavement hallucinations, developed in detail by Colin Murray Parkes, appeals to searching behaviour. There is a kind of yearning or longing on the part of the bereaved person, which leads her to search for the deceased, despite knowing that the person will never return. This disposes one to interpret sensory stimuli in certain ways, resulting in sensory experiences of the deceased (e.g. Parkes, 1970). Some bereavement hallucinations may well take the form of sensory experiences elicited by searching behaviour. However, Lewis's account points to an interesting alternative. He finds his wife only after a certain kind of search is abandoned, one that involved, in his case, memory and imagination more so than perception. What he then finds is not a non-veridical sensory image but a renewed sense of connection.
rekindling of faith. The common theme here is inexhaustibility, something resisting all of one's efforts to conceptualize and somehow contain it: "Images of the Holy easily become holy imagessacrosanct. My idea of God is not a divine idea. It has to be shattered time after time. He shatters it Himself. He is the great iconoclast" (p.55). In both cases, something is experienced as offering up possibilities that surpass one's own cognitions, affecting one in a dynamic, distinctive, living way. 7 Lewis's grief was notably solitary. However, what I have described need not be experienced in isolation. As Higgins (2013, p.175) observes, stories that we tell others about the dead or that we co-construct with others need not be aimed at preserving specific, determinate properties in memory. They can also play a role in sustaining a sense of someone's spontaneity, an ability to surpass any one narrative: "Narratives [ ….] symbolically reanimate the dead because they allow fresh insights, recalling something of the continued potential for surprise in an ongoing relationship. A story invites interpretation, and its meaning remains in flux as the interpreter reconsiders various features of it". Again, one engages with the person's possibilities, rather than assembling an increasingly elaborate inventory of her properties.

Sensed-presence without hallucination
For many philosophers of mind, a perfect hallucination consists in a non-veridical experience of p that is either identical with or at least indistinguishable from a veridical experience of p (Ratcliffe, 2017, Chapter 1). What we have here is quite different: a way of being affected that is not approximated by any sensory perceptual content, however rich and detailed. This, I have suggested, can amount to a sense of connection with a particular person. 8 Although such an experience might seem strange and unfamiliar, what I have described is in fact ubiquitous, characterizing our relations with the living and the dead. How things appear significant and salient, how we interact with our surroundings, and the possibilities we entertain in thought are all influenced in subtle, pre-reflective ways by who we are with and who we anticipate being with. Sitting in a restaurant, going for a walk, strolling round a museum, or watching a film can be profoundly different experiences, depending on who we are with. Engaging with the distinctive styles of certain people may diminish the possibilities on offer, while others imbue the world with novelty and dynamism.
One might object that, even if what I described in the previous section warrants the term "sensed presence", it differs from the type of feeling or sense of presence that is sometimes labelled as "hallucination". The latter is not a diffuse, non-localized experience of enduring connection, but something more episodic and localized, involving a feeling of the person being right here, right now. At the very least, this is more pronounced, and it may also be different in kind. However, all that is required to accommodate such experiences is some refinement. First of all, we should allow that a sense of connection may wax and wane, in ways that might be described in terms of a feeling of presence that comes and goes. Even so, what I have described is indeed less specific than the sense that someone is currently present. This is because the contrast between losing and retaining a sense of connection applies equally to memories, perceptions, and imaginings. When Lewis loses his wife for a second time, the experience is focused principally around his memories.
Under what conditions, then, does the sense of connection amount to a more specifically perceptual (or perception-like) presence? My suggestion is simply that certain ways of being affected relate more closely than others to real-time interaction with one's surroundings. How one sees this tree now, the possibilities that this painting embodies now, and the manner in which the significance of one's surroundings changes from moment to moment all implicate one's relationship with that person. Depending on which possibilities appear and how they are transformed, this adds up to an experience of relating to someone that can have a more or less specific structure.
The extent to which the relevant phenomenology approximates that of perception may vary. In some cases, it could be that kinds of possibilities that are integral to perceptual experience combine with remembered and imagined possibilities. This would generate a sense of connection that straddles the boundaries between modalities of intentionality: a relationship that is not experienced as unambiguously present or past, current or imagined. With this, there is a sense of the person as present but not fully present, or present in a way that differs from other aspects of one's environment. It can be added that the sense of presence will also be diminished to varying degrees, at least where the various possibilities associated with real time interaction, such as those of affecting and interacting with the person, are lacking. There is just the being affected. Hence, even without any dependence on sensory contents that point to particular properties, this type of experience can include degrees of specificity and, indeed, degrees of presence.
The approach can also be extended to certain experiences that are described in sensory terms, such as seeing someone or hearing his voice. For instance, in the case of being spoken to, we should distinguish experiences with an auditory and specifically linguistic content from others that involve being affected in a way that resembles communicationa change in one's experience of the possible that might be interpreted as a message. Certain other kinds of sensed-presence experiences could be thought of as privations of what I have described. Consider an experience that is bereft of certain types of interpersonal possibilities, while also involving the intensification of others. For instance, it might involve an air of inchoate menace, a current threat that is personal in nature, although not straightforwardly attributable to a specific person. Although such an experience differs from what I have described, it would be amenable to the same general approach. And this approach need not apply exclusively to bereavement; it can equally accommodate sensed-presence experiences that occur in various other contexts.
Throughout my discussion, I have adopted a fairly abstract level of description, which allows for considerable variation in how sensed-presence experiences are interpreted, integrated into shared practices, and themselves shaped and reshaped in the process. Given this, along with the widespread acknowledgement that sensed-presence experiences in general are not culturally specific, there is every reason to expect that the experience described here occurs cross-culturally. For example, in an unpublished lecture, Morioka (2019) addresses what seems to be the same type of experience, as it occurs in contemporary Japan. Morioka refers to the phenomenon of "conversing" but "without spoken language" and goes on to describe the phenomenon of an animate "persona", which is neither a hidden mind inferred from observable behaviour nor something that is apprehended via specific sensory contents. Instead, it consists in a dynamic sense of a particular person that is experienced in a self-affecting way, through one's living body.
Morioka compares this to something described by Viktor Frankl in a very different context, that of his incarceration in Auschwitz. Frankl writes of how he was sustained by an experience of connection with his wife, despite not knowing whether she was alive or dead. Although he mentions experiencing his wife's "image" with an "uncanny acuteness", he also describes a vivid experience of "mental conversation" that was not exhausted by its sensory qualities. Love, he writes, "goes very far beyond the physical person of the beloved"; it involves a sense of "spiritual being", something that does not depend on "whether or not he is actually present, whether or not he is still alive at all". Again, we come across something that is not traceable to an experience with more determinate sensory content. A way of being affected, sometimes involving reciprocity, serves to specify a particular person. Although Frankl refers to a vivid experience of her "image" or "look", he adds that it was "then more luminous than the sun which was beginning to rise" (2004, pp.48-50). This can be interpreted in terms of experiencing someone as a locus of self-affecting possibilities, rather than via the representation of however many determinate physical properties.
The sense of a dead person's somehow transcending one's image of her is not restricted to sensed-presence experiences. It is also a theme that appears more widely. Merleau-Ponty, whose work I drew on earlier, himself remarks briefly on two contrasting ways of remembering the dead: imposing one's own image on them, such that they "no longer place us in question" and continuing to respect them by retaining "the accent of their freedom in the incompleteness of their lives" (1953/1970, p.65). Consider also this passage from William Maxwell's semi-autobiographical novel So Long, See You Tomorrow, where the author reflects on a photograph of his dead mother: This picture didn't satisfy my father either, and he got the photographer who had taken it to touch it up so she would look more like a mature woman. The result was something I was quite sure my mother had never looked likevague and idealized and as if she might not even remember who we were. My mother sometimes got excited and flew off the handle, but not this woman, who died before her time, leaving a grief-stricken husband and three motherless children. The retouched photograph came between me and the face I remembered, and it got harder and harder to recall my mother as she really was. (1980/2012, pp.11-12) The problem described here is not merely that the image is inaccurate; it also interferes with a sense of this particular person that is different in kind. 9 In its more subtle forms, an enduring connection of the kind I have described may be widespread among the bereaved. In an interesting article, Allan Køster (2020b) describes a distressing sense of loss that can arise when people are unable to experience something like this, as when a child loses a parent and is later unable to summon a sense of what the person was like. That such cases are in the minority could be taken to indicate that some sense of connection is more usually sustained. However, it is questionable how often this takes on a form that might be categorized as "sensedpresence" or "hallucination" by an interpreter. Empirical work is therefore needed to determine the prevalence and variability of these experiences. However, studies to date have been insensitive to the required phenomenological distinctions. So, what I have tried to do here can be regarded as philosophical groundwork, which has the potential to inform empirical studies of bereavement hallucinations and sensed-presence experiences by providing the means to single out a distinctive form of experience that is easily misinterpreted.