Grief and Phantom Limbs: a Phenomenological Comparison

Grief and Phantom Limbs: a Phenomenological Comparison

Matthew Ratcliffe

First-person accounts of bereavement often compare it to losing part of one’s own body. More specifically, the continuing presence of the deceased is sometimes said to be like a phantom limb. One might think that these are just metaphors or analogies that serve to convey the profundity of loss. However, this paper argues that the two types of experience can indeed be structurally similar, in a number of important ways. Another person, I argue, can come to shape how we experience and engage with our surroundings in a way that resembles the contribution of our own bodily capacities and dispositions. Furthermore, the boundaries between bodily and interpersonal experience are indistinct.

Keywords: amputation; grief; habit; hallucination; interpersonal relations; phantom limb; worldexperience

1. Grief and Amputation First-person accounts of grief often state that the experience resembles that of losing a limb. Bereavement is somehow like amputation, and grief is like learning to live without an arm or a leg. 1 The following published interview excerpts are representative examples:

It’s as though I have to live without my arms or something like that – without something, but I can’t put a finger on it because it’s not visible….I have to try and learn to live without this vital you know like my sight or something, because that’s how integral my dad was.

Something I’ve kept in mind is that I really feel like I’ve had an amputation and I can’t see which limb has gone and that it’s not a visible limb, but it most certainly is an amputation – there’s no other way I can describe it. 2

1

As I use the terms here, “bereavement” refers to the short-term recognition, reaction, and response to loss, whereas “grief” refers to a longer-term emotional process. 2 From Christine Valentine, Bereavement Narratives: Continuing Bonds in the Twenty-First Century (London: Routledge, 2008), 100.

1

Similar comparisons can be found in most published autobiographical accounts of bereavement. For instance, Adri an der Heijden writes, “What else is your child but an external enclave of your own flesh and blood?....A part of me has been amputated, so how will I ever be able to say I feel at home with my body?” 3 Comparing bereavement to losing a limb serves at least to convey the profundity and painfulness of loss. 4 However, I will argue in what follows that there is much more to it than that. What we have here is not merely a set of culturally entrenched metaphors and analogies that people employ in order to stress how important somebody was to them. In fact, the two experiences can be structurally similar in more specific and philosophically interesting ways (which is not to suggest that they always are; both admit considerable diversity and we should we wary of over-generalizing). In elucidating these similarities, my principal aims are to show that (i) another person can play much the same role in shaping experience, thought, and activity as one’s own bodily capacities and habitual dispositions, and (ii) there is no clear line between the phenomenological role of one’s own body and the roles played by interpersonal relationships - the two are inextricable. Phenomenologically speaking, the boundary between subjectivity and intersubjectivity is indeterminate. In a 1975 study, Colin Murray Parkes explores, in depth, the similarities and differences between grief and reactions to the loss of a limb. He concludes that the two have much in common and tend to follow a similar temporal course:

This included an initial period of numbness, soon followed by restless pining with preoccupation with thoughts of the loss, a clear visual memory of the lost object and a sense of its presence. Defensive processes, reflected in difficulty in believing in the loss and avoidance of reminders, were also evident. 5

Both responses, Parkes suggests, centrally involve a “psycho-social transition”, an adjustment process whereby one world-view (construed not just as a conceptual representation of the world but also a way of relating to and interacting with one’s surroundings) is replaced by another. 3

Tonio: a Requiem Memoir (London: Scribe, 2015), 286. Valentine takes the comparison with amputation to convey both the “extreme nature of the pain of loss” and the “extent of the loss” (Bereavement Narratives, 100) 5 Colin Murray Parkes, “Psycho-social Transitions: Comparison between Reactions to Loss of a Limb and Loss of a Spouse,” British Journal of Psychiatry 127 (1975), 204-210, here 204. 4

2

Although there were some differences in responses to bereavement and amputation, Parkes notes that these had all but disappeared after the first thirteen months. The only exception was a “sense of the presence of the lost object”: while 56% of amputees continued to have phantom limb experiences, only 14% of bereaved subjects had comparable experiences of the deceased as present. 6 Interestingly, twelve of his interviewees had experienced both amputation and bereavement, and they further emphasized the phenomenological similarities. Parkes’ comparison between phantom limbs and the felt presence of the deceased is dismissed outright by Ramachandran and Hirstein. A phantom limb, they maintain, is to be accounted for in principally neurobiological terms rather than in terms of a psycho-social adjustment process of the kind involved in mourning. So the claim that a phantom limb is analogous to a situation where someone is “unable to believe that her husband has died” and “has a strong sense of his presence” should not be taken seriously. 7 In fact, their appraisal is rather unfair. Parkes explicitly acknowledges the obvious neurobiological differences. Indeed, he attributes the higher relative frequency of phantom limbs to the fact that bereavement is a matter of psychological adjustment, whereas both physiology and psychology contribute to the generation of a phantom limb. However, as I will make clear in what follows, this is to concede too much. The similarities that I will address are not restricted to the relevant phenomenology; the physiological effects of bereavement can also be similar, in certain respects, to the effects of losing a limb.

2. Merleau-Ponty on Grief and Phantom Limbs In comparing phantom limbs to the felt presence of the deceased, it is important to recognize that neither is adequately characterized in terms of a localized entity seeming to be present when it is actually absent; there are different ways of experiencing both presence and absence. To illustrate this, I will begin by considering Merleau-Ponty’s discussion of grief and phantom limbs in Phenomenology of Perception. 8 Merleau-Ponty suggests that the two have a common structure, involving a kind of presence quite different from that of a perceived entity situated in an already 6

“Psycho-social Transitions,” 207. V. S. Ramachandran and William Hirstein, “The Perception of Phantom Limbs: The D. O. Hebb Lecture,” Brain 121 (1998), 1603-1630. 8 Maurice Merleau-Ponty, Phénoménologie de la perception (Paris: Gallimard, 1945); English translation: Phenomenology of Perception, trans. Donald A. Landes (London: Routledge, 2012). Quotations and page numbers are from the English language edition, referred to hereafter as PP. 7

3

given world. In contrast to Parkes, he rejects additive models that attempt to account for phantom limbs in terms of distinct physiological and psychological components. Instead, he suggests, they should be conceived of in a unitary way, in terms of the “movement of being in and toward the world”. 9 To explain, when we think of a phenomenon in terms of physiological and/or psychological processes, we take for granted that the organism already finds itself in a world, where it can relate to features of its surrounding environment in one or the other way. However, the sense of being situated in a world is itself a phenomenological achievement, one that is overlooked by objective, scientific conceptions of cognition. For Merleau-Ponty, a phantom limb does not arise within a ready-made experiential world; it is integral to the constitution of a world that we take as given when we encounter localized entities as present or absent. This, he adds, applies equally to grief. Ordinarily, how we experience our surroundings reflects our bodily capacities and dispositions, along with our various projects, commitments, concerns, and values. Suppose one glances at a market stall and spots a book that one has been looking for. The book appears perceptually salient and practically significant. It may even have a certain practical allure, drawing one’s hand towards it. Its salience and significance do not just reflect a sense of one’s bodily capacities - what one can and cannot do; they depend equally on concerns that are to varying degrees idiosyncratic. Habitual ways of experiencing and interacting with one’s surroundings involve an amalgam of the two. Hence a change in one or the other will affect what appears significant and how. Merleau-Ponty suggests that phantom limbs arise when the habitual world is preserved despite a change in bodily capacities. After the loss of an arm, things may continue to appear salient, significant, and accessible in ways that they previously did. Although one can see that the arm is no longer there and one knows -in a reflective, propositional way- that it is gone, the surrounding world says otherwise: “To have a phantom limb is to remain open to all of the actions of which the arm alone is capable and to stay within the practical field that one had prior to the mutilation”. 10 According to Merleau-Ponty, anosognosia (denial of paralysis) can be understood in the same way: a person may be unable to move one side of her body but her practical field remains intact. Its retention depends in part on her avoiding situations that would draw attention to the loss, something that applies to grief as well:

9

PP, 80 PP, 84.

10

4

We only understand the absence or the death of a friend in the moment in which we expect a response from him and feel….that there will no longer be one. At first we avoid asking the question in order not to have to perceive this silence and we turn away from regions of our life where we could encounter this nothingness, but this is to say that we discern them. The anosognosic patient likewise puts his paralyzed arm out of play in order not to have to sense its degeneration, but this is to say that he has a preconscious knowledge of it.” 11

In the cases of both bereavement and limb loss, the correlate of an enduring system of practical meanings is a continuing sense of presence. But this does not involve an entity appearing to be here, now when it is actually not. Rather, it consists in a variably specific set of practical dispositions, which are implicated in how the surrounding world appears. Experience continues to be permeated by possibilities that depend on having specific bodily capacities or on being able to relate to and interact with a particular individual:

The amputee senses his leg, as I can sense vividly the existence of a friend who is, nevertheless, not here before my eyes. He has not lost his leg because he continues to allow for it, just as Proust can certainly recognize the death of his grandmother without yet losing her to the extent that he keeps her on the horizon of his life. The phantom arm is not a representation of the arm, but rather the ambivalent presence of an arm. 12

Merleau-Ponty also describes these experiences in temporal terms: they involve a “previous present that cannot commit to becoming a past”, a system of possibilities that continue to take the form ‘p is currently significant in this way’ and thus to specify patterns of activity, rather than being experienced as possibilities that have been extinguished. 13 His discussion sometimes reads as though phantom limbs and sensed presence experiences are products of intention or choice: one actively strives to preserve a lost world, in a way that resembles psychoanalytic repression. However, Merleau-Ponty also emphasizes that they involve an aspect

11 PP, 82-3. I focus on phantom limbs here and do not consider the phenomenology of anosognosia. Even if it turns out that Merleau-Ponty’s analysis does not apply to anosognosia, I will suggest that it has at least some bearing on phantom limb experiences. 12 PP, 83. 13 PP, 88.

5

of experience that operates below the level of intention (conscious or otherwise). They are integral to the structure of a world within which we act and within which we form intentions of one or another kind. Even so, they are not merely mechanistic in nature and involve a kind of bodily purposiveness. Of course, neither grief nor phantom limbs can be understood solely in terms of striving to preserve an impossible world. The habitual world also changes over time. The speed, extent, and nature of adjustment vary considerably, and explicit, effortful choices plausibly have at least some role to play. Consider an essay by Oliver Sacks on experiences of losing sight and adjusting to blindness. These, he observes, can take a number of different forms. In the case of “deep blindness”, a person eventually forgets what it was to see; even visual imagination is lost and he comes to inhabit a world bereft of the possibilities offered by sight. In contrast, others actively, willfully preserve visual imagery and even continue to utilize it in goal-directed activities. 14 Merleau-Ponty makes some complementary, albeit briefer, remarks on differing experiences of blindness, thus acknowledging that the practical field can be preserved to varying degrees and reconfigured in different ways. 15 Similarly, where grief and limb loss are concerned, it is not just a matter of “arrested time”. As we will see, the worlds of before and after interact with each other and are altered in the process.

3. Varieties of Phantom I will agree with Merleau-Ponty that phantom limbs and felt presence experiences share a common structure. They can involve systems of significant possibilities that are integral to the experienced world. The limb and the person are present in an indeterminate, diffuse way - they are implicated in situations rather than being perceived constituents of situations. Nevertheless, this story is far from complete, and there is considerably more to be said about both. MerleauPonty distinguishes between image-like experiences (or, if you like, representations) of the body and the phenomenological role of the body as that through which we experience our surroundings. In other words, he distinguishes the body “image” from the body “schema”, from

14

Oliver Sacks, “The Mind’s Eye: What the Blind See,” in Empire of the Senses: the Sensual Culture Reader, ed. David Howes (Oxford: Berg, 2005), 25-42. 15 PP, 81.

6

how the body structures practically engaged perception. 16 Phantom limbs, he maintains, are to be understood in terms if the latter; they involve retention of habitual dispositions that manifest themselves in the guise of a significant environment. However, contrary to this, first-person reports of phantom limbs indicate that they can and often do have image qualities. In a 1997 exhibition entitled After Image, Alexa Wright interviewed amputees and then produced photographic images of what their phantoms looked like. 17 These included quite specific characteristics, such as reduced diameter, partial retraction, or being frozen in a particular position. Of course, the relevant experiences are not themselves visual, but the point is that there can be a proprioceptive awareness of the limb that is sufficiently image-like for it to be described in fairly precise spatial terms. This is difficult to square with the proposal that phantoms consist of diffuse, ambiguous experiences of presence that permeate the surrounding world. It should be added, however, that phantom limb experiences are multi-faceted and diverse. So it could still be that Merleau-Ponty’s analysis captures some of them, or at least an important aspect of some of them. Ramachandran and Hirstein provide a fairly comprehensive account of the different variants. 18 All phantoms, they observe, involve a vivid sense of presence, but this core experience accommodates considerable diversity. While between 90% and 98% of those who lose a limb experience a phantom almost immediately afterwards, the sense of presence may fade within days or persist indefinitely. When phantoms do fade, they sometimes become shorter and/or change shape. Phantoms can also involve pain or cramping. For some, the limb remains rigid, perhaps stuck in an uncomfortable position, while others report experiences of voluntary movement. Others describe habitual, unthinking responses to situations, such as reaching out with a phantom arm to shake somebody’s hand. To further complicate matters, phantoms are not specific to limbs; they can also occur after the loss of a breast, a part of the face, or the penis. Ramachandran and Hirstein offer an explanation that appeals to “plasticity in the somatosensory system” and processes of “remapping”. 19 Parts of the cortex associated with the missing limb are taken over by sensory input from elsewhere in the body, making it seem as

16 For a detailed discussion of the image/schema distinction, as employed by Merleau-Ponty and others, see Shaun Gallagher, How the Body Shapes the Mind (Oxford: Oxford University Press, 2005). 17 See https://www.alexawright.com/after-image Last accessed 8th September 2017. 18 “Perception of Phantom Limbs”. 19 “Perception of Phantom Limbs,” 1608-9.

7

though the limb is still present. This is consistent with the finding that tactual stimulation of another body part often generates sensations in the phantom. The rapid onset of phantoms indicates that this process does not or at least need not involve neuroanatomical changes. Rather, patterns of synaptic activity that were previously eclipsed by input from the missing limb become more salient; they are “unmasked” by its loss. 20 However, Ramachandran and Hirstein concede that this explanation cannot accommodate every aspect of phantom experiences. For instance, it does not account for experiences of voluntary and involuntary movement. 21 In order to simplify my task here, I will exclude the phenomenon of phantom pain from further consideration. Many phantom limb experiences do not involve pain and, where there is pain, it is not clear that its nature has to be addressed in order to understand other aspects of the experience. 22 In setting aside pain, I do not wish to assume that there are no informative parallels between the pain of a phantom limb and the “pain” of grief. Perhaps there are. But, even if there are not, it remains the case that the two broad types of experience are similar in other philosophically informative ways. Even without a consideration of phantom pain, it looks as though Merleau-Ponty’s analysis is in trouble. A habitual world that is retained in the face of incapacity does not account for a limb that is frozen in position or a hand that is vividly experienced as protruding from one’s shoulder. Aplasic phantoms, which arise despite the congenital absence of a limb, pose a further problem. If phantoms involve the retention of bodily capacities, which are to some degree innate but also habitually entrained, how can we account for the appearance of a phantom where no such capacities were ever present? One response is to suggest that aplasic phantoms are different in kind from others. For instance, Gallagher suggests that they may not concern the body “schema” at all, whether innate or habitual. Instead, they are image-like phenomena. 23 Consistent with this, aplasic phantoms, unlike post-amputation phantoms, often have a late onset. In

20

“Perception of Phantom Limbs,” 1614. For a complementary discussion of phantom limbs and cortical remapping, see also V. S. Ramachandran and Diane Rogers-Ramachandran, “Phantom Limbs and Neural Plasticity,” Archives of Neurology 57 (2000), 317-320. 22 Phantom pain is experienced by between 50% and 80% of amputees. It can have different qualities, such as “stabbing, throbbing, burning, or cramping”. See Herta Flor, “Phantom-Limb Pain: Characteristics, Causes, and Treatment,” Lancet Neurology 1 (2002), 182-189, here 182. It could be that the prevalence of pain is historically variably and that this is attributable, in part, to culturally changing interpretations of phantom limbs by the medical profession and, consequently, by patients as well. It seems that, as pleasant phantoms have become rarer, painful phantoms have increased in frequency. See Cassandra S. Crawford, Phantom Limb: Amputation, Embodiment, and Prosthetic Technology (New York: New York University Press, 2014). 23 How the Body Shapes the Mind, 92. 21

8

addition, they do not involve experiences of forgetting that a limb is missing, as when trying to walk using a missing leg. Even so, Gallagher also allows for the possibility that observing and interacting with other people may somehow activate innate components of the body schema. So it should not be assumed that even aplasic phantoms consist exclusively of image-like bodily experiences. It could also be argued that phenomena such as phantom breasts and penises are principally image-based phenomena, given that breasts and penises are not integrated into “motor programs”. However, I am doubtful of this. Although one does not use a breast to act in the way that one uses an arm or leg, it is still integrated into habitual activities in all sorts of ways, shaping a sense of one’s capacities for action as well as one’s interactions with other people. The habit-body should be thought of as a unified whole, rather than as an assortment of motor capacities that are stuck together alongside inactive components. Consider a more mundane experience, which I take to be analogous in relevant respects. Most of the time during the day, I wear glasses. However, when I take them off, I usually “forget” within a few minutes that I have done so, where forgetting takes the form of habitually pressing the bridge of my nose with my index finger, in order to adjust my glasses. It is not that I first form an explicit “image” of the glasses resting on my nose. Rather, they are integrated into wider patterns of activity, habitually taken account of. Hence even artefacts can be incorporated into one’s activities and taken to be somehow present when absent, an observation that is consistent with reports of wedding rings on phantom fingers and watches on phantom wrists. 24 If this much is conceded, then there are insufficient grounds for excluding body-parts that are not directly involved in motor action from a schema-based account of phantoms. Although Merleau-Ponty does not provide a comprehensive account of phantoms, he does at least succeed in identifying an important aspect of many such experiences. Indeed, influential work on phantom limbs by Marianne Simmel in the 1950s takes the body schematic component to be most central:

24

Ramachandran and Hirstein, “Perception of Phantom Limbs,” 1607.

9

….we regard the phantom as the symptom and result of a discrepancy between the schema and physical reality. Reality can change –a leg may be lost in an accident- but the schema persists, and the phantom is the experiential representation of this persistence. 25

Some phantoms clearly do involve retention of the practical field, as when the absence of a limb is forgotten during the course of habitual action: “Despite his knowledge that the amputation has been performed the patient may ‘forget’ and reach out with the missing hand to grasp something, or to steady himself, or he may step on the phantom foot and fall”. 26 Retention of the practical field is also consistent with findings concerning the influence of prosthetic limbs on phantoms. Those who use them tend to experience more frequent phantoms than those who do not, suggesting that continuing use of the limb and thus retention of practice is somehow implicated. 27 Furthermore, it has been observed that gradual loss of a limb and gradual loss of use prior to amputation are less likely to be followed by a phantom than sudden loss of a functional limb, again suggesting that the experience has something to do with the retention of practical dispositions. 28 Where adjustment proceeds gradually, there is no sharp contrast between experience of a wholly intact practical field and recognition of its loss.

4. Grief and the Habitual World Some analogous observations apply to experiences of grief. Grieving processes cannot be understood exclusively in terms of what happens to the “practical field”. Nevertheless, it is an important aspect of grief. In both grief and loss of bodily capacity, Merleau-Ponty emphasizes the preservation of a world that is no longer possible:

The passage of time does not carry away impossible projects, nor does it seal off the traumatic experience. The subject still remains open to the same impossible future, if not in his explicit thoughts, then at least in his actual being….New perceptions replace previous ones, and even new emotions replace those that came before, but this renewal only has to do with the content of our

25

Marianne Simmel, “The Conditions of Occurrence of Phantom Limbs,” Proceedings of the American Philosophical Society 102 (1958), 492-500, here 493. 26 “Conditions of Occurrence”, 492. 27 See, for example, C. M. Fraser, et al., “Characterising Phantom Limb Phenomena in Upper Limb Amputees,” Prosthetics and Orthotics International 25 (2001), 235-242. 28 Ramachandran and Hirstein, “Perception of Phantom Limbs,” 1625.

10

experience and not with its structure. Impersonal time continues to flow, but personal time is arrested. 29

So far as grief is concerned, this description best captures predicaments that might be labeled as “complicated grief”, which involve continuing attachment to the deceased of a kind that prevents the formation of new meanings - a failure to fully acknowledge the death. 30 As Martha Nussbaum puts it, “the pathological mourner continues to put the dead person at the very center of her own structure of goals and expectations, and this paralyzes life”. 31 In Merleau-Ponty’s terms, a system of practical meanings is preserved, despite its impossibility in the face of loss. However, when we consider grief more generally, it becomes readily apparent that Merleau-Ponty’s account oversimplifies matters. Although grief can involve periods when the habitual world is retained in the face of loss, retention is not simply to be contrasted with loss. The two interact, in ways that incorporate both reflective, effortful and pre-reflective, habitual aspects. 32 Importantly, the loss is not consistently denied, in the guise of a world that remains intact. It is also recognized, where recognition also takes the form of a pervasive, diffuse experience, rather than the simple acknowledgement that a specific entity is gone from the world. Things in general lack the significance they once had - everything appears strange and unfamiliar. As Carse writes, when someone whom we love dies, “we live in a universe that makes no sense. The cosmos has lost its fundamental order. As a result, our own lives lose their meaning”. We therefore face “the formidable task of reassembling a new universe”. 33 Descriptions of meaning-collapse are to be found in almost every published first-person account of grief. The relevant experience is conveyed in a range of different but complementary ways. For instance, Joyce Carol Oates describes it in terms of the once meaningful world being reduced

29

PP, 85. For a discussion of habit-retention in the cases of grief and phantom limbs, which refers specifically to Merleau-Ponty’s work, see also Maria Talero, “Merleau-Ponty and the Bodily Subject of Learning,” International Philosophical Quarterly 46 (2006), 191-203. 30 See, e.g., Robert A. Neimeyer, “Complicated Grief and the Reconstruction of Meaning: Conceptual and Empirical Contributions to a Cognitive-Constructivist Model,” Clinical Psychology: Science and Practice 13 (2006), 141-145, here 143. 31 Upheavals of Thought: The Intelligence of Emotions (Cambridge: Cambridge University Press, 2011), 82-3. 32 As emphasized by Thomas Attig, grief consists of both “reactions” and “responses” to loss, where the latter involve agency rather than the passive experiencing of something. See his How We Grieve: Relearning the World, Revised Edition (Oxford: Oxford University Press, 2011). 33 James P. Carse, “Grief as a Cosmic Crisis,” in Acute Grief: Counseling the Bereaved, ed. Otto S. Margolis et al. (New York: Columbia University Press, 1981), 3-8, here 5. Complementing this, Attig, in How We Grieve, suggests that grieving involves “relearning the world”, revising habitual, practical assumptions.

11

to a collection of mere things, stripped of their previous significance: “Without meaning, the world is things. And these things multiplied to infinity”. 34 Along with this, there is a pervasive sense of disconnection from the consensus world. While that world persists, one has lost one’s place within it and no longer finds oneself in the midst of meaningful, shared situations: “Planes still landed, cars still drove, people still shopped and talked and worked. None of these things made any sense at all”. 35 Thus, in addition to the retention of habitual experience and activity, grief can involve a pervasive sense of habitual expectations as negated, as no longer capable of fulfillment. As C.S. Lewis writes, “I think I am beginning to understand why grief feels like suspense. It comes from the frustration of so many impulses that had become habitual”. 36 Even when one faces meaning-loss of this kind, pockets of practice can remain relatively unscathed. For instance, Oates describes how she was still able to immerse herself in her teaching, a part of her life that was insulated from her marriage. Preservation and collapse can also oscillate. 37 One might unthinkingly immerse oneself in a habitual routine that depends on one’s relationship with the deceased, and then come to recognize -in one or another way- its incompatibility with the reality of her death. A similar sense of discrepancy can also arise with memory, where memories involving the deceased appear in a different light after the death. Peter Goldie observes how, “in grieving, we relate to our past in a special way, realizing that things as they used to be, and as we remember them, can never be the same again”. He draws a parallel between grief and free indirect style in literature (a way of writing that combines internal and external perspectives on a situation, usually that of the author and a character):

Autobiographical narrative thinking can reveal or express both one’s internal and external perspective on one’s tragic loss, so that these two perspectives are intertwined through the psychological correlate of free indirect style. 38

The act of remembering unites a world that one habitually took for granted with the reality of bereavement, yet at the same time preserves the tensions between them. Similarly, we might add, the perceived world is not simply preserved or lost; its structure is in flux, and to some degree 34

A Widow’s Story (London: Fourth Estate, 2011), 176. Helen Macdonald, H is for Hawk (London: Vintage Books, 2014), here 15. 36 A Grief Observed (London: Faber & Faber, 1966), 41. 37 A Widow’s Story, 176. 38 Peter Goldie, The Mess Inside: Narrative, Emotion, and the Mind (Oxford: Oxford University Press, 2012), 56, 66 35

12

fragmented. Experience is simultaneously shaped by conflicting systems of expectation. This can involve an erosion of space-time cohesion, where some experienced situations and places are still permeated by a habitual, cohesive set of expectations that implicate the deceased, while other parts of the surrounding world incorporate recognition of loss. Consider this passage from Simone De Beauvoir, which refers to a time shortly after her mother’s 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…. 39

On the one hand, the knitting materials are experienced in terms of a coherent system of salient practical possibilities, which together imply the actual or potential presence of her mother. That system remains inherent in them. On the other hand, it no longer fits into a world from which her mother is absent, and the patterns of activity that it points to conflict with this wider experiential context. 40 Grief thus involves various relationships and interactions between retention and revision of practical meanings, and the two are not mutually exclusive. It seems plausible to maintain that some such experiences include a kind of presence, where anticipated or actual interactions with the deceased are integrated into practical configurations of the environment. However, this presence is not constant, unwavering in the guise of an unchanging but impossible system of meaning. It can be more pronounced in some situations than others, and it can be conflicted, as when certain aspects of a situation imply presence while others imply absence. It can also be more or less localized. While a system of meanings that shapes one’s world as a whole might

39

A Very Easy Death, Trans. Patrick O’Brien (New York: Pantheon Books, 1965), 98. The profound effect that interpersonal loss can have on one’s world is consistent with a wider emphasis in the phenomenological tradition on how the world of everyday experience (and -by implication- objective, scientific conceptions of the world that continue to presuppose it) depends for its sense on intersubjectivity or intercorporeality. However, it is important to distinguish ways in which particular individuals can shape worldexperience from the roles played by other people in general or by a generic other. For a helpful recent discussion of intercorporeality in Husserl and Merleau-Ponty, see Dermot Moran, “Intercorporeality and Intersubjectivity: A Phenomenological Exploration of Embodiment,” in Embodiment, Enaction, and Culture: Investigating the Constitution of the Shared World, eds, Christoph Durt, Thomas Fuchs, and Christian Tewes (Cambridge MA: MIT Press, 2017), 25-46. 40

13

presuppose someone in a diffuse, non-localized way, other experiences may more closely approximate the sense that he is right here, right now. For instance, when walking into a person’s office, where papers still lie on the table, books stand organized on the shelf, and so forth, one might experience a sense of presence involving fairly specific patterns of anticipated activity and interaction. Such an experience could be followed swiftly by a sense of negation. Or it could be recognized as discrepant even as it arises; a localized sense of presence co-exists with a more diffuse sense of absence. So, to reiterate, felt-presence experiences such as these do not consist of experiencing something as there when it is not there. There is a less determinate sense of actually or potentially relating to a particular person, constituted by variably localized and sometimes conflicting systems of practical meanings. Such experiences are equivocal, ambiguous, and quite unlike mundane perceptual experiences of entities as present within an already given world. They are therefore difficult to express. These observations also apply to the interpretation of phantom limbs. Although many different types of phantom have been identified and distinguished, the empirical literature remains lacking in one important respect. For the most part, it is stated that a phantom appears as present or as vividly present, but nothing more is said about what this sense of presence actually consists of. It is just taken for granted that we have a sufficient grasp of what it is to experience something as present. However, we have seen that certain experiences, which might be described in terms of felt presence, are quite unlike perceiving a particular entity as here, now. Furthermore, these experiences are diverse, involving varying degrees of localization, specificity, conflict, and ambiguity. So it is not enough to observe that a limb seems to be present; too many questions remain. For example, Simmel states that a person may be “more aware of the phantom extremity -even though painless- than of the contralateral intact limb”. 41 But what does being “more aware” of it actually involve? Is one aware of it in the same way but to a heightened degree? Alternatively, does it appear as present in a qualitatively different way to the intact limb? Without a more discriminating phenomenological analyses, it is unclear what an experience of presence or heightened presence actually amounts to in any given case. Nevertheless, it does at least seem clear that not all phantoms are principally a matter of retaining a practical field. Some are more image-like, more like encountering an entity perceptually. Or, at least, they include an additional image-like aspect. This, one could add, 41

“Conditions of Occurrence,” 492.

14

distinguishes them from experiences of bereavement. But it is again arguable that the same observations apply to bereavement. I have suggested that certain ways of experiencing the surrounding environment might also be conveyed in terms of the “presence of the deceased”. However, other felt presence experiences are different in kind, more like perceiving a person through one or another sensory modality, while others involve a less specific but by no means unfamiliar experience of somehow sensing that someone is present, in this room right now. Sensed presence experiences and so-called “hallucinations” in one or another modality are common among bereaved spouses. Rees interviewed 227 widows and 66 widowers in Wales, almost half of whom reported hallucinations or illusions involving the deceased spouse. These often occurred for many years and were not associated with social isolation or depression. A sense of presence was most common, but 14% also reported visual hallucinations, 13.3% auditory hallucinations, and 2.7% tactual hallucinations. Most of those interviewed found these experiences helpful rather than distressing. 42 Subsequent studies report similar findings, although it also seems that the frequency of bereavement hallucinations and sensed-presence experiences is culturally variable. 43 Parkes notes that a sense of continuing presence (a phantom) is experienced more often by amputees than bereaved spouses. Even so, he adds, “a sense of the presence of the dead husband near at hand was described in very similar terms by over a third of the widows at the time of the first interview and three quarters (16) of them reported this phenomenon at some time after bereavement”. As with phantom limbs, the experience usually “stopped short of hallucinations” involving specific exteroceptive sensory modalities. 44 It should be added that, where experiences of the deceased are described in terms of seeing, hearing, or touching, it may well be that indeterminate experiences are being conveyed in more determinate ways that render them easier to communicate. More generally, “hallucinations” in clinical and other contexts tend not to be modality-specific in the manner they are often assumed to be. For instance, when a person refers to “hearing a voice”, she may also state that the relevant experience involves

42

W. Dewi Rees, “The Hallucinations of Widowhood,” British Medical Journal 4 (1971), 37-41. See, for example, Gillian Bennett and Kate Mary Bennett, “The Presence of the Dead: an Empirical Study,” Mortality 5 (2000), 139-157; Catherine Keen, Craig Murray, and Sheila Payne, “Sensing the Presence of the Deceased: a Narrative Review,” Mental Health, Religion & Culture 16 (2013), 384-402; Anna Castelnovo et al., “Postbereavement Hallucinatory Experiences: A Critical Overview of Population and Clinical Studies,” Journal of Affective Disorders 186 (2015), 266-274. 44 “Psycho-social Transitions,” 206-7. 43

15

receiving a communication from elsewhere in a way that is unlike perception via any familiar sensory modality. 45 It is therefore unlikely that cursory descriptions of these experiences will serve to distinguish between perception-like experiences of something or someone and the kind of experience described by Merleau-Ponty. And, in some instances, the two might be better regarded as complementary aspects of a singular felt presence experience: the room suddenly takes on the air of significance it used to have when she had just walked in and, with this, one also feels that she is present in a particular location. In fact, I doubt that the phenomenological distinction between image- and schema-based experiences is a clear one, in this context at least. A practical configuration of the environment could be so specific as to imply a person’s location in a particular place, along with certain patterns of activity. Such an experience would have a gestalt structure - her presence is not merely implied by what appears salient and significant; she is also part of the scene, integral to how it is organized. When such configurations arise fleetingly, it may be like briefly seeing a silhouette of the person. Alternatively, such an experience might take the form of presence in absence. The room appears like a picture frame without a picture, a system of expectations in the context of which someone is set to appear in a certain way. Yet he fails to do so. Perhaps even this latter experience is also described, on occasion, in terms of the person’s presence. The environment is configured so as to specify his potential or even actual presence, something that at the same time conflicts with his visible absence.

5. Bodily Experience and Intersubjectivity Although experiences of bereavement and losing a limb can be similar in several respects, it might be objected that some of the similarities are superficial. For instance, even if it is admitted that one can have a perception-like experience of the deceased and a perception-like experience of a missing limb, this is not necessarily illuminating. We have all sorts of perceptual and perception-like experiences in all sorts of different situations. Furthermore, the fact that phantom limbs are largely explicable in neurobiological terms, while grief is not, might be taken to indicate that they are importantly different. However, in this concluding section, I will show that

45

Matthew Ratcliffe, Real Hallucinations: Psychiatric Illness, Intentionality, and the Interpersonal World (Cambridge MA: MIT Press, 2017).

16

the various phenomena addressed here are not only phenomenologically similar; there are physiological similarities as well. Of course, some of the sensorimotor processes involved in the generation of a phantom limb are anatomically distinct from whatever processes are at work during grief. Nevertheless, there are functional similarities between the roles played by other people in shaping our experiences and activities, and the roles played by our own bodies. I have noted that our surroundings appear practically significant in light of our bodily capacities and projects. This point applies equally to our relations with particular persons. Something may matter to me in one or another way because of my concern for you. In addition, what appears achievable reflects not just my own bodily and intellectual capacities but also what I can do in cooperation with you, what we can achieve together. So the loss of a particular individual could impact on what appears salient and how it appears significant in ways that are equally or more profound than the loss of a limb. There are also differences, of course. Losing a limb involves losing some very specific abilities, along with equally specific forms of adjustment. Loss of a person is likely to have a less selective, more diffuse impact on one’s world. But even contrasts like this are by no means clear. For someone whose life is focused around a particular type of bodily performance -perhaps a pianist or sportsperson- loss of a limb could amount to a pervasive loss of meaning from the world as a whole. And someone whose life is structured around particular projects involving another person, such as playing in a band or running a business together, may experience the loss of some quite specific abilities following bereavement. One might worry that similar points could be made about losing one’s accountant or bank manager (at least where one has become exceptionally dependent on that person’s services), but without the relevant experience adding up to one of grief. So it should be added that grief additionally reflects certain specific types of interpersonal concern and interdependence; one cares for the other person in ways that one does not ordinarily care for one’s bank manager. These kinds of concern can permeate one’s world, shaping -in often subtle ways- the perceived significance of almost every situation. 46

46 It should be added that, even if we restrict ourselves to grief, how -exactly- the habitual world is affected will depend more specifically on the nature and closeness of the relationship, and the circumstances of bereavement. A more detailed phenomenological analysis would need to address, among other things, the differences between losing a parent, spouse, child, or close friend. It is also important to consider the person’s wider interpersonal and social relations, including how other people respond and whether/how they offer support.

17

Perhaps the similarities between grief and limb loss even extend to the relevant neurobiology. Hoffman offers an account of auditory verbal hallucinations in schizophrenia, which appeals to the concept of “social deafferentation”. These hallucinations, he suggests, are functionally comparable to phantoms limbs - both are attributable to sensory deprivation. In the case of phantom limbs, an experience of continuing presence arises -in part- because of deafferentation (loss of sensory input). Analogous experiences are associated with sensory deprivation more generally. For instance, if a person is prevented from seeing for a day or two, complex visual phenomena usually start to appear. Hoffman suggests that we are similarly reliant on sensory stimulation from the interpersonal domain and that people with schizophrenia diagnoses are often socially isolated. Hence their hallucinations may arise in the same fashion: “high levels of social withdrawal/isolation in vulnerable individuals prompt social cognition programs to produce spurious social meaning in the form of complex, emotionally compelling hallucinations and delusions representing other persons or agents”. Such experiences thus involve the “repopulating” of a “barren interpersonal world”. 47 Although Hoffman is concerned specifically with schizophrenia, his position is equally or more plausibly applied to bereavement hallucinations. Consider the case of a spousal bereavement, where the two partners’ activities have been closely integrated for many years. Here, sensory expectations involving the spouse will be multifarious, habitually entrenched, and cohesive. So bereavement plausibly involves a kind of sensory deprivation that is more sudden, extreme, specifically focused, and structured than what Hoffman refers to, thus disposing a person towards perceptual or perception-like experiences with more circumscribed contents. If this is right, then image-like experiences that follow bereavement and loss of a limb are partly attributable to a common process. Turning from image-based phantoms to habitual experience of the surrounding world, it is arguable that other people shape our experiences of salience and significance in ways that are not so different from the operations of our own bodies. What appears salient and how it appears significant depend on whether one is with other people and on what they are doing. For example, Bayliss et al. found that others’ reactions to the shared environment shape one’s own appraisals of it. 48 Objects looked at with a happy expression by someone else are subsequently liked more

47

Ralph E. Hoffman, “A Social Deafferentation Hypothesis for Induction of Active Schizophrenia,” Schizophrenia Bulletin 33 (2007) 1066-1070, here 1066. 48 Andrew P. Bayliss et al., “Affective Evaluations of Objects are influenced by Observed Gaze Direction and Emotional Expression,” Cognition 104 (2007), 644-653, here 644.

18

than those looked at with disgust. A range of other empirical findings point to the conclusion that the value properties of perceived entities are shaped by interpersonal experience and interaction:

Converging evidence from behavioural neuroscience and developmental psychology strongly suggest that objects falling under the gaze of others acquire properties that they would not display if not looked at. Specifically, observing another person gazing at an object enriches that object of motor, affective and status properties that go beyond its chemical or physical structure. 49

It is not merely that perceived entities are appraised in a particular way when another person is present. The relevant properties can endure even after the person has left; they are experienced as inherent in objects. Similarly, the kinds of practical possibilities that things offer us depend not just on what our bodies are able to do but also on what can be achieved with others. 50 There is even evidence suggesting that anticipating others’ actions can shape perceptual experience in a similar way to initiating an action oneself. For instance, when one presses a button to generate a tone, it is perceived as less intense than when the tone is produced at random. The same attenuation effect occurs when another person is observed pressing the button. 51 To speculate further, the presence and intensity of such effects surely depend more specifically on the kind of social situation one is in and who one is with. The appraisals of a spouse in the context of sustained interaction are more likely to shape perception of one’s surroundings than a brief glance at the expression of a stranger. And repeated exposure to consistent appraisals is more likely to forge enduring evaluations. Unlike watching a stranger press a button, interacting with a spouse involves elaborate and structured systems of expectation that continue to influence evaluative experiences and practical dispositions even outside of one’s interactions with her. To add to this, one’s sense of what is salient and how it is significant depends largely on projects, commitments, and concerns that are shared, many of which only make sense given the relationship. In many cases, projects are ours and, even when a project does not take this form, experience and activity continue to be shaped by concern for one’s 49

Cristina Becchio, Cesare Bertone, and Umberto Castiello, “How the Gaze of Others influences Object Processing,” Trends in Cognitive Sciences 12 (2008), 254-258, here 254. 50 See, for example, Natalie Sebanz, Harold Bekkering, and Günther Knoblich, “Joint Action: Bodies and Minds Moving Together,” Trends in Cognitive Sciences 10 (2006), 70-76; Elisabeth Pacherie, “How Does it Feel to Act Together?” Phenomenology and the Cognitive Sciences 13 (2014), 25-46. 51 Natalie Sebanz and Günther Knoblich, “Prediction in Joint Action: What, When, and Where,” Topics in Cognitive Science 1 (2009), 353-367.

19

spouse - things matter in the light of their potential implications for her. Furthermore, one’s sense of what is achievable integrates the anticipated presence and abilities of the spouse in a stable, habitual way. It is important to distinguish these points from the observation that a close personal relationship can involve “we-intentionality”, where things are experienced as appearing “for us” rather than just “for me”, and where one has the sense that “we are doing this” or “we seek to achieve this”. It is debatable what the experience of “we experience/act” amounts to and how it relates to “I experience/act”. 52 However, while I accept that a close relationship will involve weintentionality and that the relevant experience requires further clarification, the point I am making here is broader in scope. Even when a situation appears as significant in a certain way “for me”, and even when something appears achievable “for me”, the partner may still be implicated. Something appears significant for me in the light of projects and wider concerns that are ours; something matters to me given my concern for you; and something appears achievable for me in light of what I habitually anticipate from you. Hence, even what one experiences as one’s own perspective (in contrast to our perspective) is shaped by one’s relationship with the other person. This is consistent with research on interpersonal co-regulation and the kinds of dysregulation that often accompany bereavement. Hofer compares bereavement to the effects of infant separation. Other people, he proposes, play a range of important regulatory roles, even in adulthood. Interactions with a particular individual can come to regulate sensorimotor activities in ways that are functionally comparable to intra-bodily regulatory processes. Certain effects of bereavement and infant separation are therefore to be understood in terms of “withdrawal of specific sensorimotor regulators hidden within the many complex interactions of the relationship that has ended”. 53 Sbarra and Hazan likewise maintain that “multiple biological and psychological systems are regulated in the context of adult attachment relationships, dysregulated by separation and loss experiences, and, potentially, re-regulated through individual

52 For some recent discussions of this issue, see Joel Krueger, “Merleau-Ponty on Shared Emotions and the Joint Ownership Thesis,” Continental Philosophy Review 46 (2013), 509-53; Hans-Bernhard Schmid, “Plural Selfawareness,” Phenomenology and the Cognitive Sciences 13 (2014), 7-24; and Elisabeth Pacherie, “How Does it Feel to Act Together?” 53 Myron A. Hofer, “Relationships as Regulators: A Psychobiologic Perspective on Bereavement,” Psychosomatic Medicine 46 (1984), 183-197, here 188.

20

recovery efforts”. 54 So, in the human case, not all homeostatic processes are wholly internal to the individual; some are interpersonally distributed. Pursuing the comparison still further, both grief and phantom limbs follow a number of different trajectories over time. As the “continuing bonds” literature plausibly emphasizes, grief is not a finite process that ends with “letting go”. And adjusting to a world without the deceased need not involve ultimately losing all those habitual ways of perceiving, thinking, and acting that involved her. The deceased can continue to be experienced as present in various ways, and the relationship is reconfigured rather than altogether abandoned. 55 As Kathleen Higgins writes, “one’s realistic expectations regarding interaction with another person are irreparably altered by that person’s death; but one’s sense of identity continues to be constructed in part on the basis of one’s relationship to that person”. 56 Perhaps phantoms can be thought of in a similar manner. Rather than retaining the experience of an arm, one might modify one’s relationship to it, in ways that are volitional to varying degrees. Taken together, all of this points to the conclusion that another person can come to play a similar role to one’s own bodily capacities in configuring the habitual world. Furthermore, the boundary between the experience-shaping contribution of bodily capacities and the contribution made by potential, anticipated, and actual relations with another person is blurred. What I take to be my own perspective on the surrounding environment does not incorporate a clear distinction between how the world appears to me and how it appears to us; the line between intra- and interpersonal is unclear. Given this, it is tempting to take utterances such as “it is like losing a part of myself” and “it feels like part of me has died” literally. What has been lost cannot be identified specifically, as when pointing to the loss of a limb. Even so, the loss is similar in kind. Something that was previously integral to one’s ability to experience and engage with the world,

54 David A. Sbarra and Cindy Hazan, “Coregulation, Dysregulation, Self-Regulation: An Integrative Analysis and Empirical Agenda for Understanding Adult Attachment, Separation, Loss, and Recovery,” Personality and Social Psychology Review 12 (2008) 141-167, here 141. 55 See Dennis Klass, Phyllis R. Silverman, and Steven L. Nickman, eds, Continuing Bonds: New Understandings of Grief (London: Routledge, 1996); see also Valentine, Bereavement Narratives. 56 “Love and Death,” in On Emotions: Philosophical Essays, ed. John Deigh (Oxford: Oxford University Press, 2013), 159-178, here 173. Higgins adds that there are also ethical aspects to continuing bonds. Severing the connection altogether involves a failure to respect the person and one’s relationship with her. She further suggests that the construction of narratives has a pivotal role to play in maintaining and shaping one’s relationship with the deceased. While it is less clear that retaining a relationship with a lost limb involves moral commitments, it is possible that narrative-construction has some role to play in shaping how phantom limb experiences develop. Furthermore, both kinds of experience are embedded in and shaped by wider interpersonal, social, and cultural contexts.

21

to perceive things in structured ways that reflect a coherent system of projects, cares, concerns, and abilities, is now absent. Valentine thus remarks on how narratives of grief point to the conclusion that “self-identity, personhood and agency” are “relational and intersubjective”, in ways that conflict with predominant emphases in certain cultures on “separateness, independence and control”. 57 We can also appeal to “mutual incorporation” here, as conceived of by Thomas Fuchs and Hanne De Jaegher. 58 The idea is that, just as one can integrate various props into bodily activities, perceiving and acting on the world through them as though they were parts of one’s own body, interactions with another agent can involve a blurring of phenomenological boundaries between the two parties. In sustained, structured interactions, there is a “reciprocal interaction of two agents in which each lived body reaches out to embody the other”. 59 Fuchs and De Jaegher place the emphasis on losses of interpersonal differentiation that arise during tightly coupled face-to-face interactions, where the two parties remain in close proximity to each other. But the point applies equally to habitual forms of incorporation, of the kind that might develop in the context of close, long-term relationships. Here, incorporation does not depend on ongoing interaction. Even when a partner is not physically present, one’s relationship with her continues to shape practically engaged perceptual experience. Indeed, the influence of “incorporation” in such a case will be much more profound, pervasive, and enduring than the kinds of incorporation that characterize one-off interpersonal interactions and episodic or habitual couplings with items of equipment. So it is important to distinguish the kind of incorporation involved in feeling connected to someone at a given time from the interpersonal permeability that I have sought to make explicit. Suppose Person A interacts with Person B and feels connected to B, somehow in unison with B. Even in a case of mutual incorporation, A continues to distinguish -to some extent- the perspective through which she experiences B from B as an object of her experience. In other 57

Bereavement Narratives, 126. Thomas Attig makes some complementary claims concerning the extent to which we are permeable to each other. For example, “If our selves were self-contained social atoms, isolated in their development, impermeable and invulnerable, and truly independent, as so many commonly think, we would not be shattered by the loss of someone we care about or love. His or her death would simply be a matter of the disappearance of another nearby self-contained being. We would remain entirely intact. But the truth of the matter is that loss shakes our personal integrity and identity. Bereavement penetrates to the core of our being.” (How We Grieve, xlvi). 58 “Enactive Intersubjectivity: Participatory Sense-Making and Mutual Incorporation,” Phenomenology and the Cognitive Sciences 8 (2009), 465-486. 59 “Enactive Intersubjectivity”, 474.

22

words, she experiences herself as relating to B. However, at the same time, A’s perspective may be structured by her longer-term relationship with B, such that the boundaries between her own subjectivity and B’s are blurred from the outset. Two ways of experiencing B thus occur simultaneously: a sense of relating to B and a more subtle way in which A’s own attitude towards and experience of B is already permeated by her relationship with B. The latter continues to shape experience and activity even when B is absent, and can also feed into A’s interactions with other people. I concede that there remain important phenomenological and neurobiological differences between grief and losing a limb. In addition, both experiences are diverse and develop in various ways. Most importantly, love for another person is something that the comparison fails to fully capture. So my claim is that some central aspects of grief can be structurally similar to some central aspects of adjusting to the loss of a limb, not that the experiences can be mapped onto each other in their entirety. As Parkes writes:

You can’t get an artificial Dad’, said one amputee who had lost a father, and it was the irrevocable nature of the loss which was emphasized by another amputee, a woman whose husband had died six years previously, ‘If you lose a leg you can tell yourself you’re going to cope – but you never get a husband back’. 60

It can be added that the enduring presence of the deceased, in many of its guises, involves continuing to care for that person, and feeling a sense of duty towards him. 61 Nevertheless, the similarities suffice to illustrate a point that has much wider applicability: a sense of our own capacities, what matters to us, and what we might achieve can come to depend, in various ways and to differing degrees, on our relations with others. Comparisons between bereavement and losing part of one’s body are not mere analogies that convey the closeness of a relationship. The two phenomena are structurally isomorphic in a number of important respects. These serve to

60

“Psycho-social Transitions”, 206. On the other hand, Dan Moller argues for the view that a spouse can, at some level of description, be functionally replaced and that resilience in the face of loss is sometimes to be understood in this way. See his “Love and Death,” Journal of Philosophy 104 (2007), 301-316. 61 Higgins, “Love and Death”.

23

illustrate how the habitually taken for granted world is shaped by one’s bodily capacities, one’s projects, and one’s relations with other people in a unified way. 62

62

Thanks to Jonathan Cole, Kathleen Higgins, Line Ryberg Ingerslev, an audience at the conference “Affectivity and Embodiment” (Helsinki, February 2017), my research group at the University of Vienna, and two anonymous referees for helpful comments.

24