Mimesis of Remembering

1 Mnestic Narration

The process of remembering, that is, the recapitulation or rather reconstruction of past events, can be counted as one of the fundamental subjects in literature. Narrative texts in particular not only often feature memory as a topic, but they also directly represent concrete processes of remembering by way of various narrative techniques. We call this ›mimesis of remembering‹: through a variety of textual strategies, literary texts can create ›memory-like‹ effects. How such ›mnestic narration‹ is achieved and what functions it might fulfil is the main concern of this article.^[1] While literature can entertain all kinds of relations with memory and memorialization, our perspective is thus indebted to the concept of ›memory in literature‹, i. e., the ways in which literary texts thematize and/or stage memory (cf. Erll/Nünning 2005). It is particularly the latter aspect that the term ›mnestic narration‹ describes.

In using the concept of ›mimesis‹, as part of the dyad ›mimesis/diegesis‹ (or showing vs. telling), we are evoking one of the oldest terms in literary theory, which has gone through many modifications and reinterpretations. Its basic meaning is that of ›imitation‹, which in the context of literature refers to the way in which texts represent extraliterary entities, actions, and events. By contrast, ›diegesis‹ refers to the mediation and commenting on the part of a speaker or narrator. In his Poetics, Aristotle salvaged the concept of mimesis from the negative connotations of a ›mere‹ and therefore deficient imitation of reality (as established by Plato) and famously emphasized its creative potential. In this understanding, literature does not simply imitate an extralinguistic reality, but that which is imitated is itself first created in the act of imitation (see Zapf 2001, 442). Further differentiations of the mimesis concept that are important to our conceptualization have been offered in Gérard Genette’s structuralist approach and in Paul Ricœur’s hermeneutics. Genette argues that »the very idea of showing, like that of imitation of narrative representation […] is completely illusory: in contrast to dramatic representation, no narrative can ›show‹ or ›imitate‹ the story it tells« (1980, 163 sq.). He insists that narrative, being created through language, can only provide an »illusion of mimesis«, since »language signifies without imitating« (ibid., 164, emphasis in the original). Meanwhile, Ricœur’s model of a three-fold mimesis is based on the idea of a dialectical relationship between literary texts and extraliterary reality, each exerting an effect on the other. Genette’s and Ricœur’s careful qualifications of the concept also inform our own understanding of a mimesis of remembering as we propose it in this article: we are not arguing that remembering is actually imitated, but that different processes of remembering, or maybe more precisely conceptualizations of such processes, can be staged in literary texts, which thereby create an illusion of mimesis.

In the following, we will describe and discuss the concrete techniques that are used in narrative texts in order to create mimesis in this sense. In so doing, we will focus on the representation of individuals’ recollections of their own past experiences, even if these recollections are always already embedded in and thus conditioned by social frameworks (cf. Halbwachs 1992). In the terms offered by Aleida and Jan Assmann, who distinguish between »individual«, »communicative« and »cultural« memory, what concerns us here is primarily the first level, remembering as »a matter of our neuro-mental system« (Assmann 2008, 109), and to some extent also the second, that is, memory as a socially constituted and communicated process. We are interested in how narrative fiction conjures up, to use Dorrit Cohn’s (1978) formula, »transparent minds«. What is thus outside the compass of our analysis are the exteriorized and institutionalized phenomena of cultural memory.

Where literary scholars have been interested in the representation of memory as a psychological process in this sense, they have often focused on the topic of trauma. Much of this work has been informed by Shoshana Felman’s and Cathy Caruth’s influential conceptualizations of trauma as a language crisis – in Caruth’s words, »the story of a wound that cries out, that addresses us in the attempt to tell us of a reality or truth that is not otherwise available« (Caruth 1996, 4). Studies of Holocaust literature in particular have advanced the understanding of how narrative writing can be read as a »textual performance of the trauma« (Kacandes 1999, 67), expressing on the one hand the »affront to human comprehension« (ibid.) and on the other, at least sometimes, a belief in the healing potential offered by processes of narrativization. Our own contribution is indebted to such readings of performances of remembering. As we see it, literary renderings of traumatic memories (or, more often as not, failures to remember) are often particularly creative in pushing the boundaries of narrative representation. However, our contribution does not only, or even primarily, aim to describe such special cases. Instead, we seek to offer a more general exploration of the techniques used in narrative texts to stage memory processes. What we want to provide is a systematic approach as well as a sense of established or naturalized forms, which may then be useful also for the analysis of trauma literature and other literary experiments that work with the subversion of these narrative conventions.^[2]

Within narrative theory, as the branch of literary studies that is particularly intensely concerned with the formal aspects of representation, an abundance of categories and concepts have been developed that can be used to analyse different components of narrative texts. Our contribution in this article is to introduce a selection of concepts we find especially fruitful for an analysis of what we call the mimesis of remembering and to sketch some suggestions of how they may be applied. Various aspects of the literary representation of memory and remembering have already been examined by others, and our own work rests on previous discussions on the relationship between memory and literature. Two early studies we have found especially useful are Aleida Assmann’s monograph Erinnerungsräume (1999) and Martin Löschnigg’s essay »The Prismatic Hues of Memory: Autobiographische Modellierung und die Rhetorik der Erinnerung in Dickens’ David Copperfield« (1999).^[3] Assmann shows – especially in the second and third part of her book – how literary texts from Shakespeare to Vonnegut have become media and repositories of memory. Löschnigg, in turn, is concerned with the ways in which Dickens’ David Copperfield models autobiographical memory. He uses this novel as a case study in order to elucidate by which means narrative texts can create the illusion of ›authentic‹ acts of remembering, or what he calls a »rhetoric of memory«. In Memory in Literature: From Rousseau to Neuroscience (2003), Suzanne Nalbantian shows how the representation of individual memory processes in the work of authors such as Rousseau, Wordsworth, Proust, Joyce and Woolf correlates with contemporary memory theories, but also prefigures neuroscientific findings. First approaches to a systematic narratological survey were offered by Marion Gymnich (2003) and Birgit Neumann (2005; 2010).

There are two fundamental points that we see as central to the mapping of the field we are offering here. First, in the following we will predominantly concentrate on two structural principles that are the basis for a narrative mimesis of remembering: such narratives need to feature a centre of subjective perception, a consciousness who performs the process of remembering (either on the level of the narrative mediation or the level of the characters), and they need to feature at least two different time levels. Second, however, obviously not all narratives that feature these very common two aspects are equally invested in representing processes of remembering. We therefore propose to think of the ›mnestic‹, i. e., memory-like, quality of texts as a scalar phenomenon: some narrative texts are more mnestic than others, that is, in these texts passages set in the narrative past are emphatically (and continuously) marked as rendering products or processes of remembering. In the following sections, we will not only introduce the various basic aspects of a mimesis of remembering – representation of time and space, narrative mediation and focalization, and unreliable narration –, but we will also discuss how these aspects are designed and combined in ways that serve to highlight a text’s mnestic qualities. Moreover, it will become clear that various combinations and framings of the same narrative techniques can serve to represent widely different conceptualizations of remembering.

2 Representation of Time: The Layered Temporality of Mnestic Narration

Narratological studies concerned with the representation of time have very often also been strongly interested in representations of remembering. Maybe the most prominent case in point is Gérard Genette’s seminal Discours du récit (1972; English translation Narrative Discourse, 1980), which offers a particularly fine-grained terminology for the analysis of the narrative representation of time. Interestingly, Genette develops his »essay in method« from the analysis of a single text: Marcel Proust’s A la recherche du temps perdu (1913–1927), arguably the most famous example of a novel (series) focussed on the process of remembering in the history of Western literature. This is hardly a coincidence; it seems obvious that the representation of memory processes in particular calls for a highly differentiated handling of time levels. As a consequence, Genette’s analyses of temporal structures already imply many crucial insights about the mimesis of remembering, even though this is not an aspect Genette himself pays much explicit attention to. Building on Günther Müller’s (1968) and Eberhard Lämmert’s (1955) studies on narrative representation of time, Genette defines three main categories that not only serve him to perform a precise description and analysis of how time is handled in the Recherche, but that he also applies to the analysis of narrative texts in general: order, duration, and frequency (see Genette 1980).

The category that is most obviously relevant to stagings of memory processes is that of order. Order is concerned with the relation between a chronology of events as they happened (story time) and the sequence in which these events are told in a story (discourse time). The basic technique for a staging of memory processes is the flashback or analepsis: events that happened earlier (according to the chronology of the story) are told later, and, importantly, it is made apparent that the content of the analepsis is a memory (the narrator’s or a character’s). Most longer narratives feature some analepses – but the mnestic quality of a text can be heightened by frequent switches between the time levels, so that readers are continuously reminded that the events represented in an analepsis are the contents of memory processes.

In classical fictional autobiographies, the analeptic passages themselves are often arranged in the chronological sequence of the story time, as for example in Dickens’ David Copperfield, in which David remembers sequences from his life, from childhood to adulthood. In the case of Dickens’ novel (as in many other works), such a linear arrangement of the analepses lends itself to representing recollection as a process that yields a coherent and meaningful life story. In turn, a jumbling of the linear sequence can serve to foreground the process by which a self recollects and deals with individual memories. For example, in Kazuo Ishiguro’s When We Were Orphans (2000), the analepses are arranged in such a way that the process of remembering appears as a working through layers, a gradual approach towards events that are meaningful to the rememberer. In this way, the life story appears less as a congruent chain of events and their consequences, and more like a puzzle in the process of being put together by the person who remembers. This is one technique that lends itself to representing the process of remembering not just as the retrieval of a sequence of past events, but a more intricate process in which present and past level are interlaced (see also Nünning 2002, 403).

Another point that is significant for the analysis of the relation between first narrative and analepses is the question which of the levels is the dominant one. In some cases, the first narrative is dominant, with only few and short analepses – the narrated memories – interrupting the otherwise chronological flow of the narration, so that the acts of remembering are just a small part of the whole content. In other cases, the first narrative is only a brief sequence that recedes behind the analepsis, as is, for example, the case in Umberto Eco’s The Name of the Rose. In this novel, the homodiegetic narrator, the monk Adso of Melk, is introduced in the first narrative as setting out, at the end of a long life, to recount events that happened in his youth. It is only at the end of the novel, in the epilogue, that the narration returns to the level of the first narrative, after the events in question have been laid out in detail. Thus, the analepsis forms the main part of The Name of the Rose. It is this strong dominance of the time level featuring that which Adso remembers that makes it likely that readers will soon lose sight of the fact that the narrated events are recollected by old Adso. Interestingly, then, a novel like Eco’s has a comparatively low mnestic quality, precisely because the level of analepsis is so dominant and almost the whole text describes one long memory – which means that the process of remembering, for the reader, soon recedes into the background.

What this example suggests is that neither a dominance of the first narrative nor of the analepsis (or analepses) is best suited to foregrounding the mimesis of remembering. Rather, such a foregrounding is more easily effected by texts which prominently feature both time levels and alternate between them. Frequent shifts between the levels can increase the mnestic quality of a text; so do explicit references from one level to the other. The latter even feature (to a small extent) in the analeptic main part of The Name of the Rose, when the narrator inserts metanarrative comments or refers to insights which the past self featured in this main part could not possess.

In some cases, such comments become a major technique for the mimesis of remembering. David Copperfield can again serve as a case in point: on the level of the first narrative, the narrator keeps foregrounding the fact that the content of his narration in the analepses is remembered content: »Looking back […] into the blank of my infancy, the first objects I can remember as standing out by themselves from a confusion of things, are my mother and Peggotty. What else do I remember? Let me see.« (Dickens 1966, 61) Martin Löschnigg (1999) has described such passages as »rhetoric of memory«. They serve as explicit reminders that the described events are to be seen as recollections, and they draw attention to the processual character of remembering. Another option is the fusing together of time levels that became a staple in modernist writing, as exemplified in Virginia Woolf’s Mrs Dalloway: »And then, thought Clarissa Dalloway, what a morning – fresh as if issued to children on a beach. What a lark! What a plunge! For so it had always seemed to her when, with a little squeak of the hinges, which she could hear now, she had burst open the French windows and plunged at Bourton into the open air.« (Woolf 2000, 3) While in David Copperfield, the narrative present is little more than a springboard from which the rememberer is flung back into the past, in Mrs Dalloway remembering is represented as a process that suffuses the present, so that the squeaking of the hinges in the present both triggers a memory of past squeaking and is itself affectively and cognitively loaded with past experience.

An aspect that has an important function in the connection or separation of time levels is that of tense. An easy and frequently used way of distinguishing first narrative and analepsis is the use of the present tense in the former and in the latter, the simple past (the »usual« narrative tense, as Franz K. Stanzel [1984, 99] calls it). In novels that on the whole follow this classical pattern, a deviation from this use of tense can be employed to great effect. We can see this in David Copperfield, where the present tense highlights recollections that have made a special impression on the narrator – for example in the passage where David remembers the day he received the news that his mother had died:

How well I recollect the kind of day it was! I smell the fog that hung about the place; I see the hoar frost, ghostly, through it; I feel my rimy hair fall clammy on my cheek; I look along the dim perspective of the schoolroom, with a sputtering candle here and there to light up the foggy morning, and the breath of the boys wreathing and smoking in the raw cold as they blow upon their fingers, and tap their feet upon the floor. (Dickens 1966, 175)

The present tense foregrounds the connection between two time levels: it suggests that sensations that belong to the level of the past are so vivid to the remembering self that it feels as if he were experiencing the situation again.^[4] In this way, two time levels are merged (see also Löschnigg 1999, 189). In Mrs Dalloway, Woolf achieves a similar effect with the use of free indirect discourse: »What a lark! What a plunge!« The twenty-first century trend towards novels written entirely in the present tense represents a further turn of the screw: in Ali Smith’s How to Be Both (2014), for example, the present tense more strongly highlights notions of remembering as an experience (and process of construction) happening in the present.

Genette’s second main category for the analysis of narrative time is duration. Duration deals with the speed of the narration, that is, the relation between story time and discourse time, or the time that is taken up by the events that are described and the number of pages used to narrate these events. Genette distinguishes four types of narrative speed, namely summary, pause, ellipses and scene. In analepses representing content of memories, shifts between summary and scene in particular are used to indicate which parts of memories are especially significant to the remembering individual. If events are related in detail (scenes), this usually suggests that they are of greater significance to the rememberer, whereas summaries often signal that there are no events of particular significance to look back to (»The rest of the half-year is a jumble in my recollection of the daily strife and struggle of our lives; of the waning summer and the changing season«, Dickens 1966, 159). The same goes for ellipses (»I pass over all that happened at school, until the anniversary of my birthday came round in March«, Dickens 1966, 175).

Genette’s third category, frequency, foregrounds the question of the relation between the frequency of an event and the frequency of its narrative representation. The default is singulative narration, which means that something that happened once is told once; in repetitive narration, something that happened once is told several times, and in iterative narration, something that happened several times is told only once. The case of iterative narration may be particularly well-suited to showing the relevance of frequency in a mimesis of remembering. Genette introduces this type of narration as a technique that is of relatively low significance in classical literary narration and only starts to be fully developed in the work of Flaubert and especially Proust, but also remarks that »Philippe Lejeune points out very correctly that from the time of Rousseau and Chateaubriand, autobiographical narrative has had recourse to the iterative more than fictional narrative has, especially (and very naturally) for evoking childhood memories.« (Genette 1988, 39) Indeed, iterative narration seems especially well-suited to staging an aspect of remembering that Genette here seems to be hinting at (»naturally«), and that is also well-documented in newer psychological research on memory: that people who remember tend to conflate events that have happened several or many times in similar ways (say, family Sunday lunch) into one memory (see Schacter 2001, 90).

It seems that the representation of time as it can be found in novels like the ones by Dickens and Proust serves to render an impression of memory as at times elusive, but overall, fairly reliable connection between past and present. By contrast, in many contemporary novels (for example Kazuo Ishiguro’s works) such techniques often serve to radically challenge the idea of memory as reliable and to explore the impact of trauma on memory. In terms of order, such techniques can, for example, include the representation of events that are impossible to place chronologically (Genette calls these »achronies«; 1980, 79). In terms of frequency, an often-used technique is the repetition of the ›same‹ memory in a variation that casts doubt on the facticity of the remembered event. Ursula Heise (1997), in her study on the representation of time in postmodern fiction, has argued that in postmodernism, narrative techniques represent a new understanding of time itself (not just of memory). In section 4, we will come back to the connections between the development of narrative techniques and changing notions about how memory works. First, however, we are going to examine the role that narrative representation of space plays for the mimesis of remembering, as it is closely connected with the representation of time.

3 Space as a Trigger of Memory, Remembered Space, and the Spatialization of Time

Space as a phenomenon in the novel is even harder to delineate systematically than time.^[5] From concrete setting to abstract ideas, for example the connections between an inner life and outer circumstances, the term space can be used to refer to a broad range of different phenomena. In the following, we will therefore only discuss some of the most important functions of the representation of space for the mimesis of remembering.

There are mainly two aspects of space that are fundamental to the narrative mimesis of remembering. First, narrative spaces can be said to offer a mimetic representation of real-life spaces, as Erich Auerbach describes it in his seminal work Mimesis (1946). Secondly, space in literature has (very influentially e. g. by Ernst Cassirer) long been understood as symbolic, as charged with additional significance. For the analysis of representations of memory processes, one particularly productive concept is connected with the latter understanding of literary space: Mikhail Bakhtin’s concept of the ›chronotope‹. This idea highlights the reciprocal interconnection between space and time in narrative literature, a ›spatialization of time‹ and a ›temporalization of space‹ (see Nünning 1995, 157). It is in works focussing on the topic of remembering that space and time are often particularly closely connected. The representation of specific places/settings often functions as a hinge between two time levels, for example when a character returns to a significant place and is then prompted to remember events that transpired there. Space, then, becomes a trigger for memory; spatial experience morphs into temporal experience.

A concrete spatial constellation that can be found in many narrative texts is a character’s return to their childhood home. Often, these spaces relate to time in an ambivalent manner. On the one hand, the childhood homes often serve to foreground the continuity of spatial parameters through time, effecting a feeling of familiarity: nothing has changed. The symbolic function of such a space is that of a repository of time, as Gaston Bachelard describes it in his Poetics of Space: »In its countless alveoli space contains compressed time. That is what space is for. […] The finest specimens of fossilized duration concretized as a result of long sojourn, are to be found in and through space.« (Bachelard 1994, 8 sq.) However, on the other hand the function as a repository of time is closely connected with another function, one that at first sight appears to be directly contradicting the first one: places also manifest change. They do so in two different ways. For one thing, places themselves change. Secondly, maybe even more importantly, the continuity of place can also serve to highlight the development undergone by the subject perceiving the place. It is not just the place that has become strange to the subject, but the subject has also become strange to the place.

The journey is a classical motif often involving the temporalization of space, in cases where the movement through space correlates with a movement through time (the return to the childhood home, as in the passage above, is just one of many possible examples). There is a whole range of different roles that such a journey into the past can play for the remembering person: it can, for example, be represented as fostering a process of identity formation, as a confrontation with (sometimes insurmountable) trauma, or as a chase after phantoms that ultimately stay elusive (as is the case with many of Ishiguro’s novels).

One phenomenon that can be understood as a ›spatialization of time‹ in a broad sense involves the representation of processes by which experience – especially traumatic experience – is inscribed onto the body of a rememberer. A prototypical example of this is the representation of scars and their origin. In such cases, the body itself is the space that becomes the point of intersection between past and present. In Toni Morrison’s novel Beloved (1987), for example, the tree-shaped scar that marks the back of Sethe, a former slave, is a leitmotif which time and again connects the level of the present with that of the past. The scar visualizes the suffering of the past as continually present, thereby also suggesting the impossibility of understanding past and present as completely separate entities. Sethe’s scar is traumatic memory made flesh, a perpetual companion (see also Paul 1997).^[6]

In the spatialization of time and temporalization of space, then, as these examples show, concrete places that are represented in detail become triggers of memory, while the temporal shift at the same time often means that they are part of the memory content. Such remembered space, in turn, is frequently charged with special symbolic significance. Their symbolic power can be semanticized in spatial relations of the type that Jurij Lotman, in his influential study The Structure of the Artistic Text (1977), has described as characteristic of literary texts in general. As Lotman shows, »the language of spatial relations turns out to be one of the basic means for comprehending reality« (1977, 218). While »the structure of the space of a text becomes a model of the structure of the space of the universe, and the internal syntagmatics of the elements within a text becomes the language of spatial modeling« (ibid., 217), spatial oppositions such as ›up–down‹ and ›close–far‹ are charged with non-spatial meaning such as ›familiar–strange‹ or ›good–evil‹.

An illustrative example of the functions that such semanticizations of space can have for the representation of memory processes can be found in Ishiguro’s already-mentioned novel When We Were Orphans. The first-person narrator Christopher Banks remembers his childhood in Shanghai, where his father, an English businessman, owned a large house. Christopher’s memories revolve around two places in particular: his childhood home in the international district, where he spent a protected childhood playing in the garden with his friend Akira, and the Chinese part of Shanghai, which was forbidden territory for the two children and for Christopher is connected with the greatest trauma of his childhood, his parents’ disappearance. In the semantic logic of the novel, the spatial opposition between international district and Chinese ghetto is clearly marked as representing a familiar sanctuary vs. representing the danger (and also fascination) of the unknown. It is also clear that the ascription of such a significance to the places is performed not just on the level of the organisation of the novel as a whole, but that at the same time is characteristic of the protagonist and his memory process.

The use of literary space as a reflection of subjective states of mind is of course not limited to cases that concentrate on the representation of memory processes. In his comprehensive study Raum, Situation, erzählte Wirklichkeit (»Space, situation, narrated reality«, 1978), Gerhard Hoffmann promotes the concept of ›mood-invested space‹ (gestimmter Raum), which describes such an externalisation of inner processes (see Hoffmann 1978, 55).^[7] He sees mood-invested space as one of the basic types of spatial representation in narrative fiction in general. Hoffmann also emphasizes that the ›objectivity‹ of mood-invested space is retained insofar as the qualities are not ascribed directly to an experiencing subject, but to the space surrounding them. In the case of the doubly semanticized remembered space, the balance of this ›double nature‹ (Doppelnatur) – objective as well as subjective – further tilts towards subjectivity (see Erickson/Gymnich 1999).

The principle of spatialization is also a central aspect of most prevalent memory metaphors (see Butzer 2005), which in turn often play important roles in narrative texts. In her study Memory in Literature (2003, 117–128), Suzanne Nalbantian shows how Anaïs Nin and Jorge Luis Borges employ the labyrinth – and the movement through the labyrinth – as a fundamental metaphor for memory. Moreover, many narrative texts feature »store metaphors« (Magazinmetaphern, Weinrich 1976) to represent memory, most prominently the library, but also, for example, the cave and the treasure chamber. Remembering as a process, in turn, is often depicted as a character’s movement through the metaphorical space. Weinrich’s second big group of memory metaphors, the tablet metaphors (Wachstafelmetaphern), is spatial insofar as these metaphors are often spatially located objects onto which memories are inscribed (such as a book, a tablet or the human body itself, as discussed above). A story like Anthony Doerr’s »Memory Wall« exploits the cultural pervasiveness of such metaphors by turning them into a reality in the storyworld: here, memory becomes a process of technological reproducibility, as particular individual memories are stored and organized in the form of cartridges, which then become desired commodities in a society in which forgetting, for various reasons, constitutes an increasing problem.

All components of the mimesis of remembering by way of spatial representation, as we have described them, have in common that they can function as hinges between time levels and as such potentially increase the mnestic quality of a text. However, beside the concatenation of time levels, there is also a further aspect that fundamentally contributes to the mnestic quality: the degree to which this concatenation is represented as being tied to a subject and their perception. The following section will thus discuss some of the most important implications and functions of narrative mediation and focalization for the mimesis of remembering.

4 Narrative Transmission and the Perspective of Memory

The ›classic‹ narrative form of conveying memories in narrative texts is that of first-person narration (e. g. in autobiography). ›Somebody‹ recollects and tells what they have experienced at certain times, typically providing some context for the experience and explaining what consequences have arisen from it. This ›somebody‹ simultaneously acts as the narrator and character of their story. According to Stanzel’s (1984, 60) »typological circle of narrative situations«, in which he differentiates between authorial, figural and first-person narrative situations, the narrator’s belonging to the world of characters constitutes the main characteristic of the first-person narrative situation. This distinction, however, still leaves central questions unanswered, such as: from whose perspective is the event depicted, from that of the narrating/remembering-I or from the perspective of the remembered/experiencing-I?^[8] The distinction between narrating-I and experiencing-I is by no means trivial, as Genette (1980, 194) points out: the identity between narrator and character (or ›hero‹) »must not conceal the difference in function and, particularly, the difference in information. The narrator almost always ›knows‹ more than the hero, and therefore for the narrator focalization trough the hero is a restriction of field just as artificial in the first person as in the third.« While the subjective process of remembering can also be staged by means of the figurative narrative situation, from a narratological point of view this difference makes the first-person situation particularly interesting (and challenging) to describe, which is why the following section will focus on this type of narration.

In texts with first-person narrators the reader is typically confronted with the different perspectives of the narrating-I and the experiencing-I. Usually, the narrating-I has an informational, and often also moral and psychological, advantage over the character as hero and comments and evaluates the latter’s perceptions accordingly. The narrating/remembering-I is more or less strongly juxtaposed with the remembered self, looking back on the latter in the sense of an »object I« (Fauconneau Dufresne 1985, 22, our transl.). Thus, the past is clearly recognizable as a remembered past, and the character can be clearly recognized as a remembered character; the story is presented as a retrospective production of meaning. Classic examples of this include Laurence Sterne’s Tristram Shandy (1759–1767), Charles Dickens’ David Copperfield, Joseph Conrad’s »Youth: A Narrative« (1902) or Daniel Defoe’s Moll Flanders (1722). In the latter, the title character and autodiegetic narrator describes her life story clearly and dominantly from the perspective of the narrative present, and is also clearly aware of this: »so you may give me leave to speak of myself under that name [of Moll Flanders] till I dare own who I have been, as well as who I am« (Defoe 1997, 13). The protagonist’s memoirs are always guided by the perspective of the ›morally purified‹ narrator, and even in the passages that render Moll’s perception as a character, the narrator’s horizon of experience resonates. The character’s ›who I have been‹ can only be defined in relation to the narrator’s ›who I am‹. Interestingly, there is also a parallel here to identity research, according to which people construct a biographical or narrative identity from the interpretation of their own memories and intentions (see Quante/Straub 2001, 269), and based on the awareness that these »experiences and actions are stored as past (memory)« (ibid., 268, our transl.). Accordingly, Moll’s descriptions of the past (like David Copperfield’s) are primarily linked to the perspective of the narrating-I. The story of the protagonist is already arranged causally and chronologically, the past is clearly recognizable as such, for example in the meta-narrative comments (»Thus far I have had a smooth story to tell of myself …«, Defoe (1722, 24), or the subsequent attribution of meaning typical of identity-creating memories and the marking of so-called turning points (»From this time my head ran upon strange things, and I may truly say I was not myself …« ibid., 27).

However, remembering can also mean, at least emotionally, an immersion into the past for the one who remembers. This phenomenon of reliving or reexperiencing can be said to be staged, for example, when a first-person narrator (such as David Copperfield in the excerpt quoted above in section 2), actually relates the past from the perceptual horizon of the experiencing-I (i. e. the character as focalizer). As Martin Löschnigg (1999, 188–189) already points out, such passages represent an intense realization or »presentification« (Vergegenwärtigung) of the content of these memories, in which the temporal levels of the experiencing-I and narrating-I seem to coincide. In this case, the change in the focalization from narrator to character goes hand in hand with a change from first narrative to analepsis and thus, as described in section 2, contributes to an increase in the text’s mnestic quality.

This is even more clearly the case, in terms of a ›rhetoric of remembering‹, in Kazuo Ishiguro’s The Remains of the Day (1989). Ishiguro’s novel, which constantly both thematizes and stages memory, uses narrative means to create the illusion of what happens with every act of remembering, as described by one of the pioneers of memory research, the psychologist Endel Tulving, in particularly vivid terms. For Tulving, remembering resembles a kind of mental time travel, a distinctly human »ability to experience again now, in a different situation and perhaps in a different form, happenings from the past, and know that the experience refers to an event that occurred in another time and another place« (1983, 1). On his journey through Cornwall, the protagonist of Ishiguro’s novel, the elderly butler Stevens, also undertakes a mental journey that takes him into his own past. This journey, as the reader learns little by little, has some inconsistencies and painful memories in store for the exemplary butler. While in large parts of the novel, similar to Moll Flanders, the horizon of experience of the protagonist as narrator overlaps with that of the protagonist as character, one finds time and again extensive text passages in which the experiencing/remembered-I clearly serves to function as the center of perception, which in the extra-literary reality would be impossible. This becomes clear in the passages of unmediated direct speech, in which the narrator recedes completely in ›a kind of self-forgetfulness‹ (cf. Martínez/Scheffel 1999, 50, our transl.), and the reader is instead directly involved in the scene. A case in point would be the dispute between Stevens and his secret love Miss Kenton, in which Stevens seems to reexperience the events all over again (Ishiguro 1989, 58).

Postmodern novels such as The Remains of the Day or Graham Swift’s Last Orders (1996) differ from classical realist novels in the functions that can be assigned to such blurrings of the lines between remembered-I and remembering-I: while in Dickens’ novel, as Löschnigg (1999, 184) remarks, »memory is affirmed as an integrative element of subject constitution« and serves to »authenticate the narrative discourse« (ibid., 196, our transl.) as well as to create narrative suspense by holding back information, contemporary literary texts arguably foreground the complexities of its modes of operation as well as the restrictions and shortcomings of human memory, as shown, for example, in psychology and neurobiology (cf. Schmidt 1991; Schacter 2001, and Markowitsch 2002). The focus tends to be on thematizing and staging the selectivity, subjectivity and the general susceptibility of memory to various disruptive factors as well as the constitutive role of forgetting.

Such understandings of the workings of memory can also be staged by means of an elision of the distinction between remembering on the one hand – that is, the reconstruction of the past as guided by present necessities for action and therefore grounded in the here and now – and the actual past as the previous unfolding of events on the other. Some texts feature a simultaneous and equivalent juxtaposition of several time layers in which the different levels of information, the perspective and the psychological constitution of the character and/or the narrator can alternate or overlap. In such a case it can sometimes be very difficult to distinguish between the remembering and the remembered or experiencing self and thus to answer the question of what is a memory of the narrated events (within the framework of fiction) and what is the actual past.

This partially indissoluble superimposition of the narrator’s and the character’s perspectives can perhaps best be described along the lines of Phelan’s (2001) notion of »dual focalization«, i. e., in our case the juxtaposition of the perspectives of the remembering and the remembered/experiencing-I.^[9] Another example from The Remains of the Day may serve to illustrate this further. Following the seemingly sudden dialogue between Stevens and Miss Kenton, the narrator comments:

[N]ow that I think further about it, I am not sure Miss Kenton spoke quite so boldly that day. We did, of course, over the years of working closely together come to have some very frank exchanges, but the afternoon I am recalling was still early in our relationship and I cannot see even Miss Kenton having been so forward. (Ishiguro 1989, 60)

Stevens, in this case clearly recognizable in his role as a narrating-I, admits certain problems with remembering and does not rule out that the events never took place in the way he describes them. Admitting that he may have inferred from the general communication style between himself and Miss Kenton to this specific conversation not only questions the ›authenticity‹ of the conversation, but also illustrates the continuous superimposition of remembering-I and remembered/experiencing-I and thus of present and past, even if one of these temporal levels is always in the foreground. This side-by-side is also, according to newer research in memory studies, what characterizes the act of remembering: »Instead, a feeling of remembering emerges from the comparison of two images: one in the present and one in the past.« (Schacter 1996, 28)^[10] Or, to quote Genette (1980, 168 sq.): »Extreme mediation [in the retrospective story], and at the same time utmost immediacy [in the feeling of reliving]. That too is perhaps symbolized by the rapture of reminiscence.«

Against this background, it also makes sense to us to introduce a further distinction, namely between the remembered and the experiencing self, terms which have been used synonymously so far. If one assumes that the remembering-I is most likely to be identified with the protagonist as the narrator, the experiencing-I corresponds to the protagonist as a character. Conversely, the ›remembered I‹ is characterized by the dual focalization or, viewed the other way round, the simultaneous presence on different time levels, and thus less easy to assign. According to the different sensations that one can experience while remembering, a distinction could be made between an unmediated immersion into the past (as triggered by Marcel’s sweet-smelling madeleines, for example) and a conscious, targeted reconstruction of one’s own past (e. g. for the purpose of constructing identity as in David Copperfield’s ›This is how I was then, this is how I am today‹). The first case would therefore be more related to the experiencing-I, while in the second case one would have to speak of the remembered-I, since the realization (Vergegenwärtigung) of one’s own identity at a certain point in time always means an artificial restriction of one’s field of vision. Frequently occurring changes between the time levels or increasingly occurring comments that can be clearly assigned to the narrator would be, for example, textual signals that indicate a ›remembered-I‹. A high degree of self-effacement on the part of the narrator as well as the tendency towards the dramatic mode, on the other hand, would rather suggest an experiencing-I in the sense of an ›ecstatic ascending memory‹.

The dramatic mode, i. e., the predominance of unmediated character dialogues, as illustrated by the example of The Remains of the Day, also creates a ›reality effect‹ in the sense of Roland Barthes (1994). It suggests not just an actual reality, i. e., a ›real‹ penetration into the past, but also the ›felt reality‹, i. e., the illusion that what is remembered represents a true-to-original image of the past and is accordingly perceived vividly. Genette (1988, 46 sq.) mentions the absence or self-effacement of the narrator and the richness of detail in the narrative as characteristics of mimetic illusion; it is especially insignificant details that drive this illusion forward. In the above-mentioned dialogue from The Remains of the Day, it is above all the exact reproduction of the conversation in direct speech (and thus the absence of the narrator) that favors such a mimesis effect. Whether the dialogue actually took place this way, or whether it represents Stevens’ memory, is certainly not clear on the basis of textual information.

Under the catchphrase »memory’s point of view« (Schacter 1996, 18), psychological research takes up two forms of memory, the distinction of which goes back to Freud: field memories and observer memories. In the ›mode‹ of field memories, which according to empirical studies (cf. ibid.) dominates most of human memory, one remembers past experiences from a perspective that comes relatively close to the actual perceptual perspective. In observer memories, by contrast, an ›outside perspective‹ prevails; the remembering person ›sees‹ himself or herself in the remembered scene from an observer’s point of view. Apart from the fact that the latter perspective dominates mainly in childhood memories and is characterized by the inevitable modification of the event, the observer perspective is mainly adopted when it comes to the reconstruction of certain ›objective circumstances‹. Field memories, in turn, according to Freud, dominate our immediate memories and then guide our perspective when we specifically want to revive emotions from a certain situation.

This differentiation between field and observer memories has a structural parallel to narrative technique (cf. also Gymnich 2003, 44) and can also be connected with the distinction between remembering and remembered or experiencing-I. If the focus in an autodiegetic narrative is dominated by the ›figure as character‹ (experiencing-I), this could be interpreted as a kind of staged field memory. If, conversely, the ›figure as narrator‹ (remembering-I) is in the foreground, the dominance of the observer’s perspective can be determined by analogy. Ian McEwan’s Enduring Love (1997), in which the narrator-character Joe describes his memories of a balloon accident, is a case in point. While the dominance of the narrating-I as a focalizing instance with all its knowledge advantage already comes to light in the very first sentence (»The beginning is simple to mark.« McEwan 1997, 1), this retrospective gesture is reinforced by likening this memory perspective to the perspective of a buzzard circling above the scene: »I see us from three hundred feet up, through the eyes of the buzzard we had watched earlier, soaring, circling and dipping in the tumult of currents« (ibid.). This rather extreme observer perspective, which implies the need for an emotionally distant memory that is as objective and comprehensive as possible, can be associated with the recollection of the narrating-I.

The first chapter of McEwan’s novel is almost exclusively devoted to the description of the novel’s core event: exact descriptions of external details, the breaking of the chronology as well as the insertion of subsequently acquired information about the other characters suggest the external perspective of the remembering-I, as do statements like »I’m holding back, delaying the information« (ibid., 2) and the metaphorical attribution of the act of memory as an »aftermath«, which literally means the second, usually more nutrient-rich cut of a meadow. Between these ›objective‹ descriptions, however, there are always passages that indicate a field memory and thus the dominance of the experiencing-I; textual markers, for instance, consist of many verbs of perception and sensation that clearly relate to Joe as the acting character. Thus, it is not only the perspectives of the narrating-I and the narrated-I that overlap in the scene, but also different perspectives of memory, which can be described with the vocabulary of psychology as field and observer memories.

5 Mnestic Narration and the Problem of Unreliability

As already hinted at in the previous section, the selectivity and subjectivity of memory as well as its proneness to error and fabrication, summed up by Schacter (2001) as »memory’s seven sins«, have been a particularly productive theme in fictional narrative especially in the last few decades. This raises the question of how, in the context of mnestic narration, one might conceptualize the concept of the ›unreliable narrator‹ (Booth 1983, 339) – a narrator whose rendering of a story readers are prompted to question, either in terms of the facticity of the events he or she is rendering, or in terms of his or her judgment of such events (Nünning 1998).

At first sight, the potential of unreliable narration for the mimesis of memory seems obvious: it makes immediate sense that unreliable narration could be used to stage the selectivity and subjectivity of the memory process, for example when a narrator like Stevens renders two different versions of the same memory. However, one might also argue – as in fact we did in the original version of this article – that in a way, a narrator whose discourse pays tribute to the fundamental unreliability of memory should therefore be seen as more, not less, reliable. This line of argument to some extent takes up a line of thought put forward by Kathleen Wall in a seminal contribution on Ishiguro’s narrators, which argues that Ishiguro’s narrators challenge a »fixation with an authoritative version of events« (1994, 37) prevalent in theories of unreliability so far.

In this updated version of our article, we would like to offer a modification of both our own and Wall’s thoughts (see also in more depth Birke 2008). It now seems to us that an important factor in the diagnosis of a narrator’s unreliability is narrative convention: while this convention demands that homodiegetic narrators display some ›life-like‹ limitations, for example in that they should not be able to ›remember‹ events they did not themselves experience, narrative convention does offer a range of acceptable options that might be unusual in a real-life person. These conventions are subject to variation between genres and to diachronic change: we would argue, for example, that within the conventions of nineteenth-century realism (which are still in operation in many ›traditional‹ novels written today), David Copperfield’s apparent ability to remember whole conversations verbatim is in line with readerly expectations and does not in itself prompt a questioning of David’s reliability as a narrator.

To reject the label of ›unreliability‹ for narrators like Ishiguro’s to us seems to be throwing out the baby with the bathwater. Rather, we would propose to retain the concept as describing the prevalence of contradictions within the text that prompt readers to start constructing a different, more convincing or more authoritative version of events. The point, then, would not be that a case like Stevens’ should not be classified as narrative unreliability – we would argue that it should be to the extent that the gaps and contradictions in his narrative prompt readers to start trying to put together a fuller, more convincing and thereby satisfying account of Stevens’ past than Stevens himself is capable of rendering.

What we would like to propose is differentiating between a type of unreliable narration that prompts an authoritative reading of a narrator’s past ›behind their back‹ and a type that does not elicit such a reading. Patrick McGrath’s neogothic novel Spider (1990) would be a good example of the former type: it invites the reader to follow the unravelling of its narrator Dennis Cleg’s narrative of the past and to reconstruct the truth about his past that Cleg himself has been trying to evade. Conversely, novels such as Ishiguro’s When We Were Orphans (2000) or Julian Barnes’ The Sense of an Ending (2011) also prompt their readers to question the narrators’ recollections and explanations, but they do not provide hints that add up to a satisfying sense of solving the puzzle. Instead of just calling attention to rememberers’ difficulties, even delusions in reconstructing their past, the latter novels also offer a fundamental exploration of memory’s role as a ›way of worldmaking‹ (Goodman 1978) for the individual, prompting readers to ponder the extent to which our constructions of the past may be just as, possibly even more important, to us than the question of ›what really happened‹.

Our revised understanding of narrative unreliability with regard to mnestic narration ties in with an emerging consensus in recent narratological debates that unreliability cannot be understood as a purely textual phenomenon, but that it also centrally involves »contextual frames of reference« (Zerweck 2019, 218). Drawing on Monika Fludernik’s (1996) argument about the way in which readers ›naturalize‹ certain narrative techniques in relation to real-world references (i. e., thereby turning them into conventions), Bruno Zerweck (ibid., 223) has argued that by »mediating between ›real‹ contextual frames (on both sides, author as well as reader) and the ›imaginary‹ world, narrative unreliability functions in a given period as a means to foreground cultural discourses – e. g. epistemological criticism – or to question ethical norms«. Unreliable narration as a literary technique, we would argue, serves to prompt readers to reflect on memory’s functions as well as its imperfections, thus calling for a »functional-historical approach to narrative unreliability« (Zerweck 2019, 215) that is sensitive to the historically variable understanding of memory processes as well as changes in the conventions and primary preoccupations of fictional narratives.

6 Concluding Thoughts

The possibilities of a narrative staging of individual memory are manifold and could of course only be briefly touched upon in our contribution. However, even this selective ensemble of literary representation processes has shown in what subtle and multilayered fashion narrative texts approach the phenomenon of remembering and how they can depict it. In the original 2005 version of this article, we already considered how closely the functionalization of the forms of representation in various texts is related to the prevailing contemporary memory and identity theories. We juxtaposed, for example, a case like Dickens’ David Copperfield, where human memory is viewed critically in that the storage qualities of memory appear to be limited, but in principle not as dubious in the long term, with the novels by Ishiguro, which stage a strong uncertainty about the reliability of memory and the possibility of an ›objectively‹ comprehensible merging of memories into meaningful life stories.

When translating the article, we realized that despite our own emphasis of the idea that ›mimesis of memory‹ might not mean the same in, say, 1850 as in the year 2000, we had at some points in the article inadvertently oscillated towards presupposing one universal standard about a ›life-like‹ evocation of remembering. In our translation we have now clarified our position. We have also rewritten the section on unreliable narration to reflect the development in our thoughts about the necessity to uphold a clearer separation between unreliable narration as a technique in fictional narrative and the potential unreliability of memory processes – a separation that, we think, also allows for a clearer discussion of how these two concepts might fruitfully be set into relation with each other.

In conceptualizing the ›mimesis of remembering‹ as tied in with developments in narrative conventions as well as with changing views on how memory processes function, we see ourselves as aligned with calls towards a diachronization and a contextualization of narratological practices (see e. g. Nünning 2009, Sommer 2021; Birke/Kukkonen/von Contzen 2022). For a full understanding of the representation of memory processes in literature, we contend, it is not enough to apply the narratological tool box to describe the techniques and forms involved – one must also consider how those techniques and forms resonate in a particular time in literary history, and how they relate to extraliterary evolutions of thought about the workings and function of remembering.