Non-subsumptive memory and narrative empathy

This article shows the relevance of a model of non-subsumptive understanding for theorising memory as a mode of sense-making that can contribute to understanding the other in ethically sustainable ways. It develops a theory of non-subsumptive memory and narrative empathy. While understanding is often seen as a form of appropriation, assimilation, and subsumption of the singular under the general, a hermeneutic approach suggests that there are also non-subsumptive, non-appropriative, dialogical forms of understanding. In dialogue with Jenny Erpenbeck’s novel Gehen, ging, gegangen (Go, Went, Gone), the article argues that cultural memorial forms, as (narrative) models of sense-making, tend to be productive when they adapt and change as they are applied to new situations and harmful when they subsume new experiences under fixed meaning templates. The article envisages memory as a resource for learning and other-oriented empathy in processes of dialogical understanding.


Introduction
Memory plays a crucial role in our sense-making processes -in our ways of orienting ourselves in the world as individuals and communities. We always encounter a new situation from a horizon of understanding shaped by our earlier experiences. Our memory even shapes what we see and do not see. Our memory, however, is never entirely our own. As has been acknowledged since Maurice Halbwachs (1925) and Frederic Bartlett (1932), we always remember in social frameworks. Memory is mediated by cultural models of sense-making that can be called cultural memorial forms. These forms affect how things are perceived in the public and how they are talked about. As contemporary memory studies acknowledges, memory is an activity that takes place in the present and serves the purposes of the present; we remember what is socially relevant for us and what the culturally available memorial forms steer us to remember (see e.g. Rigney, 2005: 17-18;Rothberg, 2009: 3-4). As Paul Ricoeur (2004Ricoeur ( [2000) puts it, 'it is always in historically limited cultural forms that the capacity to remember (faire mémoire) can be apprehended ' (p. 392). I take the notion of cultural memorial forms to signal how practices of meaning-giving are embedded in social and cultural webs of meaning that involve a process of negotiating a relationship to the past -relevant for how we orient ourselves to the present and future -through culturally mediated models of sense-making. I consider cultural narrative models to be particularly important cultural memorial forms, since such models shape our ways of drawing meaningful connections between experiences and events in time and are hence central for memory as a mode of sense-making. 1 Little explicit attention has been devoted to memory as a mode of sense-making that is mediated by cultural narrative models. In my contribution to this discussion, I would like to emphasise that practices of memory as resources for understanding are not merely cognitive but are entangled with affects and have ethical and political implications. Here, I use the notion of understanding in the broad phenomenological-hermeneutic sense, which includes the affective dimension: it is a mode of orienting ourselves in the world, a form of relationality that involves an affective relationship with the past. 2 A crucial question, both epistemologically and ethically, is how to evaluate when memory contributes to genuine understanding and when it prevents it. When does it serve ethically sustainable empathy and solidarity, and when does it lead to violent appropriation of the other?
In this article, I suggest that it is useful to place practices of memory on a continuum from subsumptive to non-subsumptive memory, depending on whether they foster subsumptive or nonsubsumptive understanding. I will thereby draw attention to an aspect of memory as a sense-making process that has not yet received attention in memory studies, and I will link it to a theory of nonsubsumptive empathy. I thereby address the need for conceptual tools that allow us to articulate the affordances and risks of different cultural memorial forms and to evaluate how they are used and abused in practices of narrating particular experiences and events. After briefly outlining my model of non-subsumptive (narrative) understanding (drawing on , I will expand on it by proposing a theory of non-subsumptive memory and narrative empathy. In the final part of this article, I will put the theory in dialogue with Jenny Erpenbeck's novel Gehen, ging, gegangen (2015, Go, Went, Gone), which can be seen to explore the workings of non-subsumptive memory and narrative empathy in a multifaceted way.

Non-subsumptive understanding
In the Western tradition of thought, philosophers have most often conceptualised understanding in terms of subsuming something singular under something general, such as a general concept, law, or model. René Descartes takes experience to conform to the innate ideas of the mind that regulate understanding, and Immanuel Kant's Kritik der reinen Vernunft (1781, Critique of Pure Reason) presents understanding as a process of organising sense-perceptions according to general, atemporal categories. 3 While the assimilatory dimension of understanding has been largely taken for granted in mainstream analytic philosophy, in the continental tradition Friedrich Nietzsche (2001Nietzsche ( [1873) has powerfully argued that knowledge as assimilation and appropriation is inherently violent because it masks the singularity of things by subsuming them under general concepts: 'Every concept comes into being by making equivalent that which is non-equivalent' (p. 145). He uses the leaves of a tree as an example: each one of them is different, but 'the concept "leaf"' homogenises them and makes us forget their differences (p. 145). The ethical problems inherent in this equalising tendency of language become more evident if we think of how the concept 'woman', for example, is used appropriatively to present 'women' as a homogenous group that differs crucially from 'men' and to impose a problematic gender binary on human beings. Drawing on the Nietzschean criticism, Emmanuel Levinas (1988: 170) equates 'comprehending' with 'englobing' and 'appropriating', and Jacques Derrida (1997[1967) writes about 'the originary violence of language' (or 'arche-violence') with reference to how language is based on classifying, naming, and inscribing 'the unique within the system' (p. 112). 4 However, instead of assuming that all understanding is necessarily and equally violent, we should explore the conditions of possibility of non-violent understanding. In the continental tradition, philosophical hermeneutics provides an alternative to the subsumption model. It presents understanding as a fundamentally temporal process that follows the structure of the hermeneutic circle: when we encounter something new in the world, we draw on our preunderstanding shaped by our earlier experiences; however, instead of the unfamiliar being simply subsumed under the familiar, the new experience can shape, modify, and transform our preconceptions. As Hans-Georg Gadamer (1997Gadamer ( [1960) writes, language usage does not consist of 'acts of subsumption, through which something particular is subordinated to a general concept'; instead, as we use general concepts in particular situations, these concepts are 'enriched by any given perception of a thing, so that what emerges is a new, more specific word formation which does more justice to the particularity of that act of perception ' (pp. 428-429).
The temporality of the use of language and of processes of understanding entails that they are always already infused with the unfamiliar, strange, and other: they are in a process of becoming. This hermeneutic conception emphasises the performative dimension of language: rather than merely representing what has happened, language gives reality structure and shape. This view allows us to see how understanding, mediated by language, neither necessarily perpetuates dominant sense-making practices nor is inevitably oppressive; it can also open up new possibilities, experiences, and realities. However, there are crucial differences in the extent to and ways in which concepts change in the process of understanding. Understanding in the strong hermeneutic sense is successful only when it has what Gadamer (1997Gadamer ( [1960: 353-361) calls an event-character and the structure of negativity: we properly understand only when we realise that things are not what we thought they were. Instead of subsuming the singular under general concepts, in such genuine understanding the singular has the power to transform the general. Gadamer does not adequately acknowledge, however, that often -in the absence of a non-subsumptive, dialogical ethos -lack of openness to otherness leads to violent appropriation (Meretoja, 2018: 110).
I argue that there is a continuum from subsumptive to non-subsumptive practices of understanding and that this continuum is relevant for theorising memory and empathy. I have earlier suggested that in the ethical evaluation of narratives, it is helpful to distinguish between subsumptive narrative practices that reinforce cultural stereotypes by subsuming singular experiences under culturally dominant narrative models and non-subsumptive narrative practices that challenge such categories of appropriation (Meretoja, 2018: 112-113). This is not meant as a clear-cut dichotomy, however, as there can be no purely subsumptive or non-subsumptive acts of (narrative) understanding. In all understanding, there is an element of the general and the particular, but their dialectical relationship can be closer to the subsumptive or non-subsumptive end of the continuum. I use the expressions 'subsumptive' and 'non-subsumptive' understanding/memory/empathy as a shorthand for cases that are close to those ends of the continuum. These concepts are hence heuristic tools for differentiating between various practices of understanding and relationality. While the use of cultural sense-making models is always dialogical in the minimal sense of involving the (re)interpretation of general models as they are applied to particular situations, it is sometimes dialogical in the stronger sense that involves openness to what is different and an active willingness to modify one's general models and categories. I will next expand on the model of non-subsumptive understanding to propose a theory of non-subsumptive memory, which provides tools for evaluating different practices of memory.

Non-subsumptive memory
While memory has been traditionally seen as a process of retrieving aspects of the past, recent memory studies has emphasised the dynamic, productive, and performative nature of memory.
Remembering is an activity: something that we do in the present. 5 It makes something present for particular purposes and thereby takes part in shaping current social reality. Like narrative, memory is selective and perspectival: it selects certain things from the past to be remembered and leaves others to oblivion; it displays things from a certain perspective and dismisses other perspectives. The selectivity and perspectivism inherent in memory affect how we interpret new experiences and situations on the basis of our past experiences and our ways of remembering them. Although memory is usually taken to concern the past, it is important to acknowledge that as a mode of sense-making, it is also directed at the present and future. Through practices of memory, we make sense of the past, but this activity takes place from the horizon of the present and serves as a way of orienting ourselves to the future. Paul Ricoeur (1999) emphasises this future-oriented aspect of memory when he writes about our duty to remember, which 'consists not only in having a deep concern for the past' but in reflection on how 'we may prevent the same events from recurring in the future' (pp. 9-10; see also 2004[2000]). Recent memory studies has discussed this issue in terms of collective responsibility (see e.g. Rothberg, 2019). In evaluating the ethical and political potential of memory for the present and future, it is useful to look at how practices of memory as processes of sense-making can be placed on a continuum from subsumptive to non-subsumptive.
There are different ways of using narratives of past experiences as models for making sense of new experiences. Memory follows the logic of subsumption when it is reified into a fixed memorial form under which new events and experiences are rigidly subsumed. In the act of subsumption, a particular phenomenon (experience, event, etc.) is subsumed under a memorial form that functions as a model of sense-making so that the phenomenon is taken to be the 'same' as something of which we have a memory or with which we are familiar, for example through cultural narratives. Culturally dominant narratives of Nazi Germany and East Germany, for example, can function as such sense-making models when developments in Trump's America are likened to the rise of Fascism in 1930s Europe or when the surveillance of dissidents in contemporary Turkey or Hungary is equated with Stasi terror. As Michael Rothberg (2017) puts it, comparison is unavoidable because 'we cannot not attempt to understand our local situation (whatever it is) without reference to global, historical developments in a variety of other national contexts', but it should do justice to both similarities and differences. If narratives of certain historical events are used as fixed templates under which very different historical phenomena are simply subsumed, we risk losing sight of their historical specificity.
In contrast, memory serves understanding non-subsumptively when we use our earlier experiences as a starting-point for understanding something new but without thereby subsuming the new under what we remember and understand from the past. Non-subsumptive memory functions as a resource for learning because it is dialogical: it is characterised by a willingness to change one's understanding both of past events in light of what one learns in the present and of current events in light of one's new understanding of the past. Non-subsumptive memory is open in both directions.
However, there can be no purely subsumptive or non-subsumptive memory, and in practice, processes of memory combine subsumptive and non-subsumptive elements in different ways. The meaning of the Holocaust and East Germany, for example, is shaped by a range of narratives and images that circulate across media; instead of having merely one general, shared meaning, they are sites of contestation, negotiation, and debate characterised by a dynamic of subsumptive and nonsubsumptive acts that may serve different and sometimes conflicting memorial aims. In the process of negotiating contested memory, attempts at subsumption can be challenged and reinterpreted in non-subsumptive ways, and the same practices of memory can have subsumptive and nonsubsumptive dimensions, depending on the perspective and context.
Mechanisms of subsumption and non-subsumption operate at the level of both individual and public memory. Here, I would like to emphasise the interplay between the individual and the cultural in all processes of memory. The conceptual dichotomy between individual and collective memory is still surprisingly pervasive in memory studies, despite all the criticism that has been levelled against it. For example, a recent Oxford handbook is called The Collective Memory Reader despite the editors' preferred term social memory studies (Olick et al., 2011: 40). The term collective memory seems to be more approachable and recognisable to a broad academic audience than cultural or social memory.
Although the concept of cultural memory has been developed partly as an attempt to overcome the dichotomy between the individual and the collective, its key theorists, Jan Assmann (1995) and Aleida Assmann (2011), saw it as separate from 'living' memory rooted in personal experience and used it with reference to memory embodied in and transmitted through 'material carriers such as memorials, monuments, museums, and archives' (p. 6). As Astrid Erll (2008) observes, however, the concept can be used to acknowledge that memory and culture intersect at not only collective but also individual level: 'we remember in socio-cultural contexts' (p. 5). Nevertheless, cultural memory is often perceived as a form of collective memory in contrast to the psychological memory that individuals have of their personal experiences. Ann Rigney (2005), for example, asserts that cultural memory 'is always "external" in Halbwachs's (1925) sense, in that it pertains by definition to other people's experiences as these have been relayed to us through various public media and multiple acts of communication ' (p. 15). Seeing cultural memory as a form of 'external' collective memory, however, risks dismissing the ways in which cultural mechanisms pervade individual processes of experiencing and remembering. The notion of cultural memorial forms, I argue, draws attention to how individual processes of meaning-giving are culturally mediated by (narrative) models of sense-making. In general, it has not been adequately acknowledged in memory studies that the practices and mechanisms of cultural memory -mediated by cultural memorial forms that function as (often narrative) models of sense-making -affect how things are experienced in the first place.
As an evaluative tool, the model of non-subsumptive memory can be compared to Rothberg's (2019) model, which has important heuristic value. Rothberg maps different forms of multidirectional memory by placing them at the intersection of an axis of political affect (a continuum from competition to solidarity) and an axis of comparison (a continuum from equation to differentiation). Equating discourses exemplify what I have called the logic of subsumption, whereas differentiating discourses function non-subsumptively. My model of non-subsumptive memory, however, differs from Rothberg's in the following ways.
First, non-subsumptive memory is not only about comparison but about the application of general sense-making models to particular cases. Although it often entails comparison, it is more broadly a mode of understanding, which involves giving meaning to one's own or others' experiences. While Rothberg's model is for analysing public discourses of memory, non-subsumptive memory can function at either public or individual and interpersonal level. When it comes to experiences of others, non-subsumptive understanding is a theory of relationality: how to encounter someone else in a non-appropriative, dialogical manner.
My next two critical points argue that a range of affects is relevant for the evaluation of memory practices. In terms of the object of memory (what is remembered), Rothberg's model privileges negative affect (suffering caused by violence and injustice), and in terms of the politically desired outcome, it privileges positive affect (solidarity and empathy). However, and this is my second point, it is equally important to remember positive affects (such as memories of solidarity). Memory studies has been dominated by 'a traumatic paradigm', privileging experiences of suffering as worth remembering, but a new movement is now emerging that focuses on the memory of joy, hope, and agency (see e.g. Reading and Katriel, 2015;Rigney, 2018;Sindbaek Andersen and Ortner, 2019). I see the notion of non-subsumptive memory as a way of drawing attention to the positive potential linked to agentic power and to the transformative potential of agency as each act of interpreting cultural memorial forms in a specific situation involves the possibility of interpreting otherwise. 6 Third, not only empathy and solidarity but also negative and ambivalent affects produced in the process of remembering can be highly valuable, such as affects of confusion, uncertainty, and unsettlement. Often it is when we realise that we do not understand that we are able to listen and learn and approach the other in an undogmatic manner. 7 Katja Garloff (2020: 215-218) goes even further in arguing that such negative affects as rage and resentment can be effective forms of communication for example in contesting normalisation in a post-genocide context. Fourth, the model of non-subsumptive memory makes the ethical evaluation of memory practices more explicit than Rothberg's model. Rothberg (2019) suggests that 'differentiated solidarity' offers the greatest 'political potential' (p. 124). Political potential, however, depends on one's political orientation. I would say that a non-subsumptive ethos (implying what Rothberg calls 'differentiation') is a necessary even if not an adequate condition for genuine understanding of the other and for an ethically sustainable relation of empathy and solidarity. Even if non-subsumptive memory practices can be linked to various political affects, insofar as they are animated by a non-subsumptive ethos of dialogue they are likely to foster receptivity to otherness that challenges one's beliefs and thereby provide a basis for the kind of other-oriented empathy that is respectful of the other and does not aim at appropriation.
Fifth, the affective and cognitive dimensions of memory are entangled in such complex ways that their neat separation is problematic. Although it may be analytically useful to differentiate between them in some contexts, their connections may be overlooked if they are thought of as separate axes. Empathy and solidarity are not mere affects and comparison and subsumption not mere cognitive operations. In my model, it is crucial that non-subsumptive memory has both an affective and a cognitive dimension. It is a certain -dialogical -way of relating to others, of orienting oneself towards them, coloured by a certain mood and attunement, characterised by sensitivity to difference and a willingness to listen and learn. Such non-subsumptive dialogicality is both a style of orientation/understanding and an affective mode of relationality.

Non-subsumptive narrative empathy
The theory of non-subsumptive memory that I have proposed here is integrally connected to a theory of non-subsumptive empathy. The concept of empathy has been used in different ways in various disciplines; here I will relate my approach particularly to discussions in cultural memory studies and in cognitive narratology, while also drawing on philosophical and psychological discussions on empathy. I argue that in both cognitive narratology and memory studies, a major problem with the concept of empathy is that it tends to be perceived in terms of feeling with the other -or even feeling what the other feels -as if we had access to the other's emotions with some kind of immediacy. 8 The idea of unmediated access has been fiercely criticised by twentieth-century hermeneutics (Gadamer, 1997(Gadamer, [1960: 218-250, 305), which emphasises the mediated nature of all understanding, but the idea of immediacy lives on in current approaches to empathy in the idea of direct sharing of emotion. The theory of non-subsumptive narrative empathy qualifies current discussions on empathy in four ways. First, it emphasises that empathy is not a mere affect but should be thought of as a mode of orientation that involves understanding and imagination. Second, it is ethically crucial to distinguish between empathy based on self-oriented and other-oriented perspective-taking and to reflect on how the latter, too, is culturally mediated. Third, empathy should be theorised as more firmly relational with more attention to how it takes place from the empathiser's historically and culturally constituted horizon of understanding. Fourth, narrative empathy can move us to action if it shapes our narrative agency and ethical identity.
First, both memory studies and narrative studies should acknowledge that empathy is not only an affective response but a mode of orientation that involves processes of understanding and imagination, in which narrative plays an important mediating and facilitating role. Currently, empathy is most often taken to be 'an affect' (Rothberg, 2019: 134) or 'other-oriented emotion' (Batson, 2011: 11). Cognitive narratologists believe that, equipped with mirror neurons, we are hardwired for 'automatically sharing feelings' and that in empathy 'I feel what you feel' (Keen, 2007: 4-5). Both imagining and understanding, however, can be seen as necessary requirements for affective engagement with the other as an other. 9 It is easier to imagine what the other feels when we have sufficient knowledge and understanding of the other's situation, and in order to properly understand we need to be able to imagine, in terms as complex as possible, what it is like to live in a certain historical and cultural world in that particular subject position. Hence, understanding and imagining are mutually dependent and feed into each other. Their inextricable link becomes particularly clear when we engage with the perspectives of others who live in social conditions that are markedly different from our own. Without sufficient knowledge and engagement with others, we are likely to project our own experiences and expectations onto their lives rather than imagining them as radically different from our own. Exchanging narratives about one another's lives can be an important vehicle of such understanding and imagination (see Goldie, 1999: 409). Nonsubsumptive narrative empathy, however, also requires a willingness and ability to learn from the other: imagining, empathising, and learning about the other go hand in hand. 10 Learning means not subsuming the other under one's ready-made categories or narrative templates but being open to what one does not yet have categories for and a willingness to modify one's narrative imagination. When this happens, we can talk about non-subsumptive narrative empathy.
Second, empathy as an orientation to the other shaped by openness and willingness to learn from the other crucially depends on acknowledging difference. Sameness and similarity are often seen as key concepts in theorising empathy (Assmann and Detmers, 2016: 8). From a hermeneutic perspective, however, ethical understanding of the other begins with acknowledging difference. Andreea Deciu Ritivoi (2016) emphasises, from such a perspective, the need 'to resist positing similarity between ourselves and others' (p. 63). In conceptually differentiating between various forms of empathetic understanding, it is useful to distinguish between self-oriented and other-oriented perspective-taking. Empirical research in psychology suggests that imagining what someone else experiences in a certain situation is a distinct cognitive and affective process compared to putting oneself in the other's shoes in the sense of imagining what oneself would feel in the other person's situation; the former does not involve a merging of the self and the other (Batson, 2009: 274). In psychology, these two modes are designated as 'imagine-other' and 'imagine-self' perspective-taking and in philosophy 'other-oriented' and 'self-oriented' perspective-taking (Coplan, 2011). Successful other-oriented perspective-taking involves the capacity to differentiate between oneself and the other. However, the discussion on other-oriented perspective-taking does not seem to take sufficiently into account that imagining the other's perspective is not something that we do by simply leaving behind, at will, our own values, understandings, and presuppositions. There is a need to rethink other-oriented perspective-taking to question the idea of immediacy and the voluntarist presuppositions inherent in it.
This brings me to the third point, which concerns the need to acknowledge the fundamentally relational and socially, culturally, and historically mediated nature of empathy. We always encounter the other in relation to our socially, culturally, and historically constituted horizon of experience and memory. This means that other-oriented empathy cannot be completely 'selfless' in the sense that we necessarily imagine the other's situation and perspective by drawing on our own life history and embodied memory, in a process mediated by cultural memorial forms. 11 Non-subsumptive empathy does not entail leaving behind one's own values and being transported to the other's perspective; instead, it involves putting one's values and commitments at play and at stake, letting them be tested and questioned but also allowing the encounter to clarify what in them is worth holding onto. In fact, a major advantage of a non-subsumptive, relational understanding of empathy is that it allows us to pay more attention to what we bring to the encounter with the other and to engage with it in a more self-aware and self-critical way. When we acknowledge that we encounter the other from our own horizon of understanding, experience, and memory, we are better able to reflect self-critically on our own presuppositions. Such awareness is hence ethically relevant and likely to promote other-oriented non-subsumptive engagement with the other.
Fourth, reflection on one's own role as an empathiser is particularly important if we want empathy not simply to remain a passing affect but to make a difference in terms of action. From this perspective, it is crucial to reflect on how various cultural memorial forms function in the present and what implications they have for our ethical agency. Narrative empathy is generally insufficient to promote moral action since it is based on stories of individual experience whereas large-scale social justice requires political efforts to overcome structural injustices. Nevertheless, some cultural memorial forms may succeed in using narrative strategies to elicit our empathy in ways that have potential to move us more fundamentally than merely moving us to feel with the other. This can happen if they move us to the point of shaping our narrative agency -the way in which we navigate our narrative environments and respond to the stories of others -and our narratively mediated ethical identity: our narrative sense of who we are and could be as moral agents (see Meretoja, 2018: 11-12, 102). Such a sense of agency affects how we act (including how we vote) and whether we end up acting in solidarity. Narrative empathy has the potential to move us to moral action if it succeeds in transforming our sense of what kind of moral agents we are and want to be. This can have a collective political impact if it engenders a transformation of cultural narrative imagination. Jenny Erpenbeck's (2015 novel Gehen, ging, gegangen (Go, Went, Gone) appeared at the height of the 2015 'migrant crisis' but is based on the 2012-2014 occupation of Oranienplatz, a protest by African refugees against German asylum politics. 12 Erpenbeck draws on interviews she conducted with the protesters. The novel's protagonist is a retired German professor of classics, Richard, who has an East German past and, in the narrative present, slowly gets to know African refugees who are looking for asylum in Berlin. The reception of the novel has emphasised its didactic aspects: how it sets a model for the reader on how to empathise with the plight of the refugees. 13 In contrast, I read the novel as an exploration of how memory and narrative can function as resources for non-subsumptive, dialogical engagement with the other.

Gehen, ging, gegangen: Becoming foreign as a shared experience
The novel presents narrative and memory as interlaced modes of sense-making that are integral to our attempts to orient ourselves in the world: 'They find their way by these stories? Yes. They find it by remembering? Yes' (Erpenbeck, 2017: 150-151). 14 In thematising the intimate link between narrative and memory, it shows how we remember what we receive in narratives, how narratives shape intersubjective reality by selecting certain things as worth remembering and neglecting others, and how the stories we have been raised on become part of us almost as if we personally remembered the recounted events. For Richard, an important formative memory is a story of how he was, as a toddler, in the aftermath of the Second World War, almost left behind on a train platform, but at the last moment a Russian soldier lifted him up and handed him through the train's window to his mother: 'This was a story his mother told him so many times that eventually it seemed to him he remembered it himself' (17). 15 For Richard, this memory has become a memorial form that functions as a narrative model of making sense -of exile and solidarity across difference -when he engages with the experiences of the refugees.
Erpenbeck's novel deals with the relationality of memory by showing how we relate to new situations on the basis of our memories of our past experiences, by relating the new to what we are familiar with and to narrative memories that function as memorial forms that provide narrative models of sense-making. Richard relates his new experiences first and foremost to his East German past -and to some extent to the Nazi past of Germany. Cultural narratives of the Holocaust and both personal and cultural memories of East Germany function as memorial forms that he uses in the attempt to relate to the situation of the refugees and to make sense of the German response to it. These templates both enable understanding and prevent it. In distinguishing between productive and problematic uses of memorial forms, it is crucial to evaluate whether they function subsumptively or non-subsumptively.
For Richard, a defining moment in his life has been the fall of the Wall, after which he lost his bearings and 'no longer knows his way around' (Erpenbeck, : 35/2017. In trying to engage with the refugees' experiences of exile, his East German past both helps him to see certain things and blinds him to others. Particularly in the beginning, the narrator repeatedly draws attention to what the protagonist does not perceive. For example, he does not hear the silence on Alexanderplatz, despite the magnitude of the silence produced by the asylum seekers who refuse to say their name: The silence of these men who would rather die than reveal their identity unites with the waiting of all these others who want their questions answered to produce a great silence in the middle of the square called Alexanderplatz in Berlin. . . . Why is it that Richard, walking past all these black and white people sitting and standing that afternoon, doesn't hear this silence? He's thinking of Rzeszów. (11) 16 Richard's archaeologist friend has told him about a system of cellars under the Town Hall of the Alexanderplatz, and he relates it to the labyrinth of tunnels under the Polish city of Rzeszów. Jews hid in both during the Nazi years. 'Just like in Rzeszów', he thinks (11/2015: 17). This act of subsumption, which involves seeing a sameness between the two, stops him from seeing what is going on in the square, out in the open, in plain sight. Here we can see how culturally mediated memory affects what we see and experience in the first place.
It is often the moment when Richard thinks that he understands -when he takes this or that to be exactly the same as something that he knows from another context -that he is led astray. In contrast, it is the humility linked to admitting his ignorance and the curiosity of wanting to know more that lead him to understanding something new. In these productive instances of understanding, however, he also draws on his past, particularly on those aspects of his life that involved an experience of being lost, disoriented, or estranged -an experience of becoming a stranger, not only to others but also to himself: 'In 1990 he suddenly found himself a citizen of a different country, from one day to the next, though the view out the window remained the same' (81). 17 A similar experience of disruption is crucial to the life-stories of the refugees: 'From one day to the next, our former life came to an end. . . . Our life was cut off from us that night, as if with a knife '. (90) 18 In response to retirement and the new perspectives that the dialogues with the refugees open up for him, Richard begins to feel even more foreign in his own life: he walks through his house 'as if strolling through a museum, as if he himself no longer belonged to it . . ., his own life, room after room, suddenly appeared to him utterly foreign, utterly unknown, as if from a far-off galaxy' (91). 19 A key question that the novel explores is how to find 'shared points of reference' (118) without subsuming the other under one's own expectations and preconceptions. In the narrative present, Richard's relationship with his lover has ended because he had too many expectations, to which she referred as 'happy-ending terrorism ' (116/2015: 125). With the refugees, too, he stumbles when he is bound by fixed expectations: 'Richard is now expecting to see the African at first incredulous, perhaps, then speechless with excitement, and finally overjoyed, jumping into the air with relief and pleasure, throwing his arms around Richard or at least bursting into tears. Nothing of the sort happens' (205-206). 20 One necessarily has a certain preunderstanding, but what is crucial is the ability to revise it. The novel draws attention to everything that Richard brings to the encounter with the other, instead of pretending that he has full access to the other. First, he makes sense of the refugees' life-stories through the categories he is most familiar with, mainly deriving from his own field, classics. He even renames the refugees after heroes of the Western mythological and literary canon (Apollo, Tristan), but eventually gets better at dialogical listening and is able to let his preconceptions change. The novel focuses on the conditions of possibility of a genuine encounter in which something ethically valuable happens -that is, on openness and willingness to learn as conditions for a dialogical encounter in which presuppositions become challenged and categories of understanding transformed.
The cultural background of the African refugees is so different from that of Richard that it would seem pretentious and dishonest to assume that he could simply feel with them what they feel. What he can do, however, is try to imagine what they are going through by drawing on his own memories of experiences of foreignness while at the same time acknowledging that he can only understand them partially and from his own perspective. For example, he remembers visiting the US and being 'beside himself with the foreignness' (185). One of the African refugees does not know how to act properly with the grandmother of one of Richard's friends. Here Richard's experience of empathy is coloured by his awareness of the cultural difference, which he does not pretend to fully understand but which makes him realise that he would be just as lost if he had to take care of the refugee's grandmother: 'Would he have any idea how to look after an African grandmother? Nana?' (185) 21 These processes of empathy are characterised by a salient awareness of the social and cultural differences, and this awareness often helps to counteract the subsumptive impulse and opens the way for a non-subsumptive engagement with the other, based on curiosity, humility, and wonder rather than an appropriative claim to feel what the other feels.
In turning attention to the conditions of possibility of transcultural empathy, Erpenbeck's novel reminds us that memory and empathy are socially embedded and take place in situations in which power and agency are unequally distributed. Richard is a highly privileged white man, although he has always felt like a second-class citizen in the united Germany, with a lower salary compared to his West German colleagues. Nevertheless, he is underprivileged in a very marginal way compared to the refugees. Over the course of the narrative, he grows increasingly aware of his own position of privilege: 'Richard knows he's one of very few people in this world who are in a position to take their pick of realities' (219). 22 Moreover, one of the important insights he draws from his East German past is that social conditions create our fortunes and misfortunes and our roles could easily be reversed: 'they were just the raw material for these political experiments. . . . Things might have turned out the other way around. . . . Sylvia says: I keep imagining that someday it'll be us having to flee, and no one will help us either' (95). 23 The possibility of role reversal is concretised in the episode in which Richard goes shopping but notices at the till that he has forgotten his wallet. The refugee Rufu pays for his shopping and refuses to be paid back in full (129/2015: 138).
As Richard reflects on the situation of the refugees by relating their situation -what they have to endure and how they are treated -to what happened in Nazi Germany, for him most crucial is the question of shared humanity. Underneath our clothes 'every one of us is naked and must surely, let's hope, have taken pleasure in sunshine and wind, . . . perhaps even have loved someone and been loved in return before dying one day': 'Enough grows and flows in this world to provide for all, and nonetheless . . . a struggle for survival is apparently taking place here' (210-211). 24 He asserts that how we now treat the refugees is the real test: 'only if they survived Germany now would Hitler truly have lost the war' (50). 25 Are we now able to count everyone in the sphere of humanity or do we still differentiate between those who merit to be treated as fully human and those who are second-class humans with no basic rights? Richard asks himself whether peacetime has hardened people, afflicting them with a 'poverty of experience', an 'emotional anemia', which can lead to such aggressive refusal to help those seeking refuge 'that it almost looks like war' (241/2015: 254).
At the same time, however, Richard grows more aware of how acknowledging shared humanity is not enough; it is also necessary to engage with the concrete, lived reality of each individual. The more he learns about the refugees, the better he is able to see the limits of his earlier categories of understanding, and the better he listens, the more he learns. Hence, openness is a precondition for new understanding, but the process of learning from others -often through narrative interactionalso helps one to become more open and undogmatic. Moreover, the novel shows that the plights of the refugees call for empathy that is more than a mere affective response -empathy that leads to acting in solidarity (Richard ends up helping his new friends the best he can, e.g. by teaching them German grammar and soliciting legal advice) and to treating the other as an equal. Richard suspects that one of his African friends, Osarobo, may have broken into his house when he was away, but he shies away from raising the issue with him. His friend Anne explains to him why he should 'make a scene': 'Because you have to take him seriously. If you make excuses for his betrayal, then you're basically just putting on airs, playing the morally superior European' (257). 26 In order to encounter the other in a non-condescending, non-appropriative manner, one should respond in the way that one responds to someone who is a moral agent with integrity and responsibility.

Towards explorative memory
Non-subsumptive memory and empathy are driven by curiosity and openness, by a desire to engage with the other in ways that entail exposing oneself to the other and a willingness to let go of one's own certainties. They display an orientation towards the other that is characterised by a mode of dialogic exploration.
Erpenbeck's novel as a whole conveys such a non-subsumptive ethos: its narrative style emphasises the open-ended, tentative, and preliminary nature of its narrative endeavour. The novel begins with the word 'perhaps' ('Perhaps many more years still lie before him, or perhaps only a few', 3/2015: 8), and it has no authoritative narrative voice. Readers only have access to how things appear from a limited perspective and some speculations on the possible reasons for the characters' actions and reactions. The air of uncertainty and open-endedness present throughout the novel invites readers to participate in the process of interpretation. In fact, Richard lacks transparency even to himself, which the narrator occasionally thematises, for example by saying that he attends a meeting 'for reasons unclear even to himself' (28/2015: 34).
Critics have observed that the 'emotional restraint' in Erpenbeck's narrative style works as a way of avoiding sentimentality or a moralising tone (Wood, 2017). Her non-subsumptive mode of narration, however, also has a direct thematic link to the novel's way of dealing with the conditions of possibility of non-subsumptive memory and empathy: it repeatedly shows how closed, fixed, subsumptive narratives tend to be linked to a form of memory that is less productive in terms of creating conditions for genuine understanding than narrative memorial forms that are less certain of themselves and invite dialogical, explorative engagement.
Non-subsumptive practices of memory function in the explorative mode in that rather than assert and explain, they open up new horizons for asking questions. All genuine understanding stems from asking genuine questions, ones that lack settled answers and lead us to an 'open state of indeterminacy' that allows 'testing of possibilities' (Gadamer, 1997(Gadamer, [1960: 363-364, 375). Erpenbeck's (2017) novel emphasises the primacy of questions and how they drive forward processes of dialogical understanding: 'the act of listening always contains the questions: What should you understand? What do you want to understand?' (p. 75) 27 Richard is convinced that 'everything depends on asking the right questions' (235), 28 and he begins to get to know the refugees by asking them questions about where they come from and what it was like at home. The memories they share through dialogic storytelling constitute a process of exploration in which what is at stake is who they are and how they exercise their narrative agency by telling their own stories (see Meretoja, 2018: 11-12). One of the refugees, Awad, would 'like to tell him his story . . .. Because if you want to arrive somewhere, you can't hide anything' (Erpenbeck, 2017: 57). 29 Their agency and sense of self is entangled in stories -ones they have started, ones into which they have been thrown, and ones they have taken up from the culturally available repertoire of narrative models -and their interaction means that they will become part of each other's stories, leaving a mark in one another's narrative memory: Richard asks himself, 'Will he too occupy some place in their stories?' (152) 30 Getting to know the refugees, in turn, transforms his narrative sense of self so that he wants to be someone who helps those in need.
When we approach memory as a resource for asking productive questions, the emphasis shifts from memory as knowledge to memory as a starting-point for not-knowing, which can be ethically more valuable than the presumption of knowledge. Here we should also acknowledge that a whole range of affects can be ethically valuable. Negative or ambivalent affects linked to experiences of perplexity, confusion, becoming aware of one's expectations and their ethically problematic implications, and fear and anxiety over the unknown can be ethically valuable resources for critical engagement with one's own horizon of memory and experience from which one encounters the other, in ways that allow one to overcome obstacles for genuine understanding and other-oriented empathy. Self-satisfied subjects who feel good about empathising with the plight of others are more likely to participate in subsumptive appropriation of others than those who engage with their own experiences of uneasiness, unsettlement, and confusion in the process of trying to imagine the others' experience.
In Gehen, ging, gegangen, the emphasis on the question-like, explorative mode of memory, narrative, and empathy means that a desire for narrative mastery is presented in a tensional relationship with what eludes narrative understanding. Richard looks for a narrative that would help him understand what it means to be a refugee: 'To investigate how one makes the transition from a full, readily comprehensible existence to the life of a refugee, which is open in all directions -drafty, as it were -he has to know what was at the beginning, what was in the middle, and what is now' (39). 31 The novel shows how that which cannot be told -in the Western story economy -cannot be remembered, either. The reporter asserts: 'If nothing special happens, I can't make a story out of it' (12)(13). 32 The everyday, slow processes in society cannot be easily narrated, and this is why structural violence tends to remain invisible. If the refugees jump off the roof of a high building, it would be more scandalous than the fact that the state lets them 'slowly expire under miserable conditions' (219-220/2015: 232), as only the first makes an easily tellable and sellable story. At first, Richard finds it exhausting to engage with the life-stories of the refugees because he cannot frame them and subsume them under a neat narrative: 'When an entire world you don't know crashes down on you, how do you start sorting it all out?' (48) 33 From the refugees' point of view, their stories are so overwhelming and contain so much pain that it is not easy for them to fight their way into 'the world of all that can be told' (59/2015: 65). The novel explores the tension between the need to narrate their own stories and the narrative templates of the prevailing story economy.
This aspect of the novel is linked to its critical engagement with the unequal distribution of narrative agency within society and across the globe (see Meretoja, 2018: 70-71). Some people have more power than others to determine which stories get told: 'the things you've experienced become baggage you can't get rid of, while others -people with the freedom to choose -get to decide which stories to hold on to' (Erpenbeck, 2017: 67). 34 This question of power is decisive as it is ultimately our narrative memories that define who we are: 'without memory, man is nothing more than a bit of flesh on the planet's surface' (151). 35 Erpenbeck contributes to cultural memorial forms that convey and encourage both a sense of shared humanity and a sense of the singularity of each individual life. Her novel suggests that we should draw both on memories of foreignness, marginalisation, neglect, and violence and on memories of solidarity and productive dialogue to keep ourselves emotionally open and alive to the plights of others, but at the same time we should accept that we can never fully understand another human being and all we can do is try to learn from one another, through curiosity, empathy, and a sense of solidarity. Her novel is an example of how aesthetic form itself can display the logic of non-subsumption that invites dialogic empathy. The whole novel is an open question, an exploration that is shot through with humility and modesty. One of the key questions it raises is this: could it be time now to move towards thinking of memory as a resource for dialogical understanding and imagination? This is a challenge that might be worth taking up in current memory studies and empathy studies in order to develop the idea of future-oriented memory and narrative empathy that function as resources for learning and asking productive questions that help us engage with others in non-appropriative, dialogical, ethically sustainable ways.
Although memorial forms are scarce in the sense that their cultural repertoire is limited (Rigney, 2005), their scarcity looks different if we acknowledge that those forms are in a constant process of change and transformation. Cultural memorial forms are not fixed and immutable; instead, they change over time as they are negotiated in new contexts. Memorial forms do not mean anything in themselves; they only elicit meanings when they are used and abused, sometimes in productive and sometimes in problematic ways. They gain new meanings as they are reinterpreted and applied to new situations, and literature can contribute to their transformation. In this article, dialogue with literature has elucidated the way in which cultural memorial forms as narrative sense-making models tend to be productive when they adapt and change as they are applied to new situations and harmful when they subsume new experiences under a fixed meaning template. From this analysis, non-subsumptive memory emerges as an important resource for learning and other-oriented empathy in processes of dialogical understanding.