Fictional Minds suggests that readers understand novels primarily by following the functioning of the minds of characters in the novel storyworlds. Despite the importance of this aspect of the reading process, traditional narrative theory does not include a complete and coherent theory of fictional minds. Readers create a continuing consciousness out of scattered references to a particular character and read this consciousness as an "embedded narrative" within the whole narrative of the novel. The combination of these embedded narratives forms the plot. This perspective on narrative enables us to explore hitherto neglected aspects of fictional minds such as dispositions, emotions, and action. It also highlights the social, public, and dialogic mind and the "mind beyond the skin." For example, much of our thought is "intermental," or joint, group, or shared; even our identity is, to an extent, socially distributed. Written in a clear and accessible style, Fictional Minds analyzes constructions of characters' minds in the fictional texts of a wide range of authors, from Aphra Behn and Henry Fielding to Evelyn Waugh and Thomas Pynchon. In its innovative and groundbreaking explorations, this interdisciplinary project also makes substantial use of "real-mind" disciplines such as philosophy, psychology, psycholinguistics, and cognitive science.

Analyzes constructions of characters' minds in the fictional texts of a wide range of authors

About the Author
Alan Palmer is an independent scholar living in London, England. He has a PhD from the University of East London.

Reviews
"Though critics and cognitive scientists may define reading differently, the two disciplines can be of use to one another-as Fictional Minds by Alan Palmer clearly demonstrates"-Times Literary Supplement

Narrative

THOMAS ALBRECHT AND CE LINE SURPRENANT

This chapter reviews books and journal articles published in 2004 in the field of narrative theory.

The 2003 YWCCT review of works published in the field of narrative theory during that year (YWCCT 13[2005]) noted that the former was marked by a 'distinct shift towards the cognitive sciences'. The reviewer spoke of the 'narrative turn' taken by cognitive sciences, psychology and psychiatry, who have come to recognize the importance of narratives in their respective fields, and thus have taken 'narrative beyond the confines of the literary critic and the linguist' (YWCCT 13[2005] 99). The works here reviewed, especially Alan Palmer's Fictional Minds, seek to demonstrate the extent to which narratology in turn can benefit by adopting conceptual and methodological elements imported from these disciplines (Palmer received the 2005 Modern Language Association Prize for Independent Scholars). A common idea runs through works of narrative criticism published in 2004, namely that an interdisciplinary approach to narratives is better suited to the 'inter-medial matter' that the latter are. It is noteworthy that the call for interdisciplinarity in narrative theory goes together with that for giving up 'the privacy of consciousness' as a dominant and limiting paradigm.

The 'turn to narrative' in cognitive sciences, psychology and psychiatry follows from the idea that narratives are constitutives of our Western lives, and from the discovery of their potential therapeutic and heuristic power (note that Freudian psychoanalysis never had the need to effect a 'narrative turn', as Roy Schafer discusses in his essay 'Narrating, Attending, and Empathizing', reviewed below). As for the complementary 'shift' by narrative theory towards cognitive and other sciences, it claims to be motivated by the need to move beyond the perceived limitations of 'classical narratology', specifically the latter's alleged reliance on linguistic models and its scientistic taxonomic aims. Palmer, for example, locates his contribution in 'post-classical narrative theory' (p. 152), in the vein of David Herman, the editor of the collection Narratologies: New Perspectives on Narrative Analysis (OhioStateUP [1999]), who introduced the term in order to designate the proliferation of new methodologies and hypotheses concerning the form and function of storytelling. Ge rard Genette (whose book Metalepse is reviewed below) would probably disclaim the idea that 'narratology' constitutes only a 'moment' in the history of narrative study, having himself welcomed the insights of analytic philosophy (in L'Oeuvre d'art 1. Immanence et transcendance E d. du Seuil [1994] and 2. La relation esthetique [1997]) while pursuing his work on the relation between poetics and fiction. Similarly, the

2004 re-edition of Brian Rogers' important work on the narrative techniques of A ` la recherche du temps perdu gives us an indication that we can still fruitfully engage with earlier phases of narratology, which have benefited from the advances of intertextual and genetic criticism.

That said, Jonathan Culler's call, at the end of his article 'Omniscience'-'It's time to remove the blinders and explore alternative vocabularies better attuned to the strange effects of literature' (Narrative 12:i[2004] 32)-would seem to exemplify the shared, but by then already well-established, concern in works published in 2004 for transforming classical narratology. The descriptions of the paradigms to be overcome vary in precision, but there emerges in many of the works under review a picture of 'classical narratology' as a set of flawed conceptualizations, represented as (or, perhaps more accurately, simply called) 'structuralist' and 'formalist' narratological methodologies characterized by an exaggerated degree of abstraction, 'purity' and too great of a reliance on linguistics. 'Structuralism' denotes less the very precise series of methodological procedures which we find in classical structuralist works than any excess of theorization more or less directly linked with the historical moment of structuralism. In addition to the polemic against structuralism or 'structuralism', studies published in 2004 continue to question the nature of narratives and seek to expand the condition of

'narrativity' beyond 'verbal media'. However, their efforts do not dispel the ambiguity that surrounds the object of narrative studies; instead, their interpretative hypotheses and methods sometimes appear to make of narratives themselves a secondary matter.

This sidestepping of the question of narrative as such is decidedly not the case in the anthology Narrative across Media: The Languages of Storytelling, edited by Marie-Laure Ryan, one of the most important publications of 2004 in the area of narrative studies. Comprised largely of essays by scholars who teach narrative theory and media studies at universities in Scandinavia and the United States, it is significant among other reasons for the potentially useful contribution it makes to the ongoing paradigmatic debate about what a narrative actually is and about what it means to say that a particular medium is narrating (rather than doing something else). Broadly generalized, its explicit and implicit definition of narrative has two main components, from which it then draws two important implications for the fields of narratology and media studies.

In her Introduction to Narrative across Media, Ryan finds in the canonical structuralist and language-based narrative theories a shared inability to provide a tenable, satisfactory definition of narrative itself. Following the well-documented inclination of much recent work in narrative studies already described above, she proposes to shift the definition away from the traditional linguistic paradigms and towards more cognitive approaches.

Narrative, she argues, might be defined not as a linguistic object, but as a kind of meaning that is produced in the interaction between a semiotic object (in whatever medium) and its recipient. To qualify minimally as narrative, this meaning must somehow entail the dimensions of space (a world, characters, objects), time (changes of state) and plot (causal relations, motivations). In some cases, especially in verbal artefacts, narrative meaning is directly encoded into the object by means of signs. But in other cases, in particular in non-verbal artefacts, it is something that might be decoded by the recipient in response to the object. According to Ryan's expanded definition, ten, the term narrative designates less an object and more a 'mental image' (p. 11), a cognitive construct built by a recipient interacting with the formal properties of an object. Hence, for Ryan, narrative entails not only those signifying objects that already 'are' narratives, but those that 'possess narrativity', which is to say objects that may potentially be decoded by a recipient as narrative

(p. 9). We find one example of this paradigmatic shift from the object to the recipient in David Bordwell's contribution to the volume, 'Neo-Structuralist Narratology and the Functions of Filmic Storytelling'. Elaborating his wellrehearsed

polemic against what he calls neo-structuralism, Bordwell maintains that narrative films must be analysed not in terms of a set of statically conceived formal features, but functionally, in terms of 'effects registered by a perceiver prepared to grasp a narrative' (p. 204).

What follows from this first component of Ryan's definition is a 'mediumfree' or 'trans-medial' (as opposed to linguistic) understanding of narrative. Even though Ryan concedes that narrative does largely assume and lend itself to verbal forms of representation, she maintains that works in any number of different media can potentially be called narrative, assuming they possess or elicit what she calls narrativity. This would be true even for primarily or exclusively non-verbal media, as long as they allow a recipient to infer some kind of narrative meaning. Narrative across Media explores this possibility in essays by Wendy Steiner on early Renaissance oil painting, Eero Tarasti on the narrativity of instrumental music and Peter Lunenfeld on video art.

Complementary (and sometimes contradictory) to their view of narrative as a quality intuited by a recipient, the contributors to Narrative across Media also define narrative in formalist terms, as a representation that encodes narrative meaning by means of its media-specific signs. What distinguishes their approaches from earlier formalist approaches, however, is that they locate such signs of narrative meaning not only in linguistic texts, but in a variety of verbal and non-verbal media. Their underlying point is that various media can possess and elicit narrativity, and that this narrativity is itself always media-specific. According to this view, any narrative meaning (and any meaning more generally) an object has or elicits is dependent on its media-specific encoding and transmission of that meaning. For Ryan, this second aspect of narrative's redefinition demonstrates 'how the intrinsic properties of the medium shape the form of narrative and affect the [recipient's] narrative experience' (p. 1). According to her theory of narrative, the medium (understood as a set of formal constraints and as a set of unique possibilities) is itself the decisive category in determining the narrative quality of the given object and its reception as narrative.

David Herman's essay 'Towards a Transmedial Narratology' offers a particularly compact example of this argument that narrative meaning is inextricable from the 'medial support' by means of which it reaches a recipient. Herman distinguishes between the medium-specific narrativity of oral and literary stories, for instance at the level of their respective representation of spatial and temporal relationships, in order to make the larger case for the medium-dependency of all storytelling. In advocating for 'a general theory about the links between stories and their media' (p. 67), he writes:

the story logic of a tale . . . told conversationally is bound by different constraints than those bearing on a literary tale . . . even though the two narratives [may] focus on similar experiences. Only barely initiated here, a project for future research is to determine just what sorts of constraints shape the communicative properties of each storytelling medium. (p. 68)

In making this point, Herman is primarily interested in calling attention to the uniqueness of each narrative medium (and in validating the accompanying traditional narratological distinction between story and discourse), but he also implies that not all media, given the specific constraints of their communicative properties, may actually qualify as narrative. This implication is explored at length in Espen Aarseth's contribution to Narrative across Media, 'Quest Games as Post-Narrative Discourse'. Aarseth maintains that computer games are decidedly not narratives and should therefore not be analysed according to the schemes of traditional narratology. This is because of what he calls an 'ontological difference' between games and stories: 'this difference is probably best described with the word choice' (p. 366). Aarseth goes on to elaborate the ontological rift he finds between narratives and games as a distinction, among other things, between representation and simulation, retrospection and action, and constative and performative language. In choosing the order of the selections in Narrative across Media, Ryan has suggestively juxtaposed Aarseth's essay with her own article, 'Will New Media Produce New Narratives?', in which she makes a case for the narrative potential of certain forms of digital media, including some computer games.

Wendy Steiner's fine contribution to Narrative across Media, an essay entitled 'Pictorial Narrativity', provides perhaps the best example of both aspects of Ryan's unifying editorial argument: the potential narrativity of nonverbal media and the medium-specific character of any medium's narrativity. Steiner's analysis focuses on narrative paintings of the early and middle quattrocento, a transitional genre located between the non-realist narrativity of medieval art and the non-narrativity of high Renaissance realism. Through a close reading of Benozzo Gozzoli's The Dance of Salome and Beheading of St John the Baptist, a painting in which three discrete episodes from Salome's story are non-consecutively juxtaposed, Steiner shows how Gozzoli produces a complex narrative effect through the directionality of the figures' gazes, through the use of spatial recession (suggesting temporal recession), and through metaphoric and metonymic echoes in the figures' clothes, colouring, poses and facial expressions. Her compelling and often counterintuitive

discussion of the painting, a work whose narrativity seems at first glance to be limited and crude at best, makes a strong case for the argument that a given medium's greatest narrative potential lies in its own medium-specific constraints and possibilities (rather than in its more or less successful approximations of linguistic models).

In its double definition of narrative (as a cognitive construct and as a medium-dependent set of signs), Ryan's anthology has two important implications for the fields of narrative theory and media studies: a transmedial approach to narrative as such and, complementarily, a media-specific analysis of any given medium's narrative potential. Accordingly, the collection is divided into five sections, each organized around a particular medium: face-to-face narration (which is to say oral storytelling), still pictures, moving pictures, music and digital media. The section on music is particularly valuable, in consideration of most readers' presumed unfamiliarity with this area of research, and contains a useful overview by Emma Kafalenos of recent scholarship on the relation between music and narrative.

Not unlike Ryan's discussion of the limitations of exclusively linguistic definitions of narrative, Alan Palmer's Fictional Minds provides a detailed account of the limitations of some of the most important narratological concepts, through a study of what it calls 'fictional minds' (rather than of 'characters'). Under this heading, we are not simply presented with an 'alternative' term, such as the ones demanded by Culler to replace inadequate isolated concepts, but with the outline of a new approach to what is shown to be an essential aspect of the process of reading fiction, namely the narrator's and reader's construction of characters through the ascription of consciousness to them, that is, through the attribution of a variety of thought processes to them. 'How precisely do these groups of words [of which characters are made] become the recognizable fictional minds that are clearly contained in fictional texts?' (p. 12). The presentation of the 'interrelations between different types of thought in fiction' is not merely one component of narratives among many equally important other ones. Palmer's claim is stronger: 'narrative fiction is, in essence, the presentation of fictional mental functioning' (pp. 5, 185). The reading process eminently includes the ability to 'decode other minds' (p. 131). The book stems from the observation that what one recognizes easily as 'one of the pleasures of reading novels', namely 'being told what a variety of fictional people are thinking' (p. 10), is in fact very odd. By what narrative processes (or occult intervention) are we given direct access to someone else's mind? Many of the concepts developed in narratology are an attempt to reduce or to engage with that strangeness (or in deconstructive ways, to 'exaggerate' it, as in Nicholas Royle, 'The Telepathic Effect', reviewed in Chapter 7 of YWCCT 13[2005]).

For Palmer, the strangeness comes partly from the fact that consciousness and thought are viewed as private and individualistic domains, which can be accessed solely through introspection, the narrative form of which is the 'flow', the 'stream-of-consciousness'. Narrative theorists have up until now overemphasized the role of that device in transformations of the novel because they failed to perceive the solipsistic view of the mind on which it rests.

If we turn to 'real-mind discourses' (cognitive science, psychology and psycholinguistics), we discover an alternative picture of the mind, notably the one developed in philosophy of mind in the vein of Gilbert Ryle and others,

whereby 'to talk of a person's mind is . . . to talk of the person's abilities, liabilities, and inclinations to do and undergo certain sorts of things' (p. 11), and access to someone else's thoughts becomes a contextual and social matter.

The term 'mind' is used instead of consciousness and thought because it 'embraces all aspects of our inner life: not just cognition and perception, but also dispositions, feelings, beliefs, and emotions' (p. 19). Moreover, the paradigm of the mind which discourses of cognitive science, psychology and the philosophy of mind propose is 'a mode of action in which mental language is not privileged' (p. 171). Attached to this view of the mind is Wittgenstein's claim that 'meaning is not inner, mysterious, private, and psychological, but outer, evident, public, and behavioral' (p. 142). Novels contain, according to Palmer, 'a set of instructions that relate to mental functioning', which cannot be deciphered without holding a 'holistic view of the whole social mind in action'. The book elaborates a functional and teleological method on how best to read these instructions.

After a programmatic Introduction, Chapters 2 and 3 explore existing and classical narratological approaches toward fictional minds. The centrality of fictional minds has been in the best of cases merely implicit, or has been ignored when narratives have been considered through the categories story analysis, possible worlds, characterization, frames and focalization. Some of the most striking arguments of the book are found in Chapter 3, which discusses the problems surrounding the use of 'Speech Categories' for the presentation of fictional thought. Classical narratology, in adopting a bias towards 'speech categories', has unduly focused on 'inner speech', that is, on self-conscious thought. Being in the 'grip of the verbal norm', narratology has excluded areas of the mind that cannot be analysed through the speech category: beliefs and latent states of mind, intentions, purposes, judgements, skills, knowledge, imagination, intellect, volition, habits of thoughts and dispositions. To use speech categories to analyse characters' minds 'gives the wrong impression that [they] really only consist of a private passive flow of consciousness' (p. 59), which would explain 'the high regard in which free indirect discourse is held' (p. 64). The mind is active, social and contextual, and the agent of action, rather than 'simply the object of a discourse' (p. 53). Palmer calls for 'an explicit recognition that much of the thought that takes place in novels is . . . purposeful, engaged, social interaction' (p. 59). His argument against the 'verbal bias' is matched by Ryan and the contributors of her anthology, who defend the idea that narratives are trans-medial and not tied to the linguistic form, even if the latter is the best known of its forms. To make of characters' mind a private matter has implications for how one characterizes the relation between characters, narrators and readers, and explains the neglect of 'thought report' in existing narratology. The narrator appears to be obtrusive and distorting (p. 77). Palmer proposes 'an approach towards narrators' presentations of the whole mind that focuses on both states of mind and inner speech and that acknowledges the indispensable and pivotal role of thought report in linking individual mental functioning to its social context' (p. 76).

Chapters 4 and 5 aim to widen and deepen the concept of fictional mind beyond 'inner speech' (p. 87) by turning to conceptual tools offered by the real-mind discourse. For example, functionalism, which considers the mind as an 'information-processing device' (p. 88), offers a concept of the mind in action, which stands in stark contrast to the passivity of 'inner speech'. From that point onwards, as Palmer reviews various areas of the mind (from action,

non-consciousness, dispositions, to emotions), the book will not cease to reinforce the notion that the 'whole mind' is above all a social matter, so much so that the importance of real-mind discourses for narrative theory seems

entirely to lie, on the one hand, in their affirmation of the mind as a social phenomena, as the main source of the 'intersubjective approach' which the book adopts, and, on the other hand, in their oppositions to the 'verbal bias' in conceptions of the mind (the key to the whole book is stated in Chapter 5, in terms of the need 'to rediscover the social character of mind' (J.R. Searle, The Rediscovery of the Mind. MIT. [1992] p. 248]). Russian thought following the

1917 revolution provides a means to do so, among others, with its 'functional emphasis on the social nature of thought and on the public nature of apparently private mental life' (p. 147). Together with Bakhtin's notion of the utterance, Palmer introduces, among others, the idea of 'intermental thinking', which designates shared processes of thought qualified by cognitivists as 'non-conscious', the notion of 'distributed cognition', and emphasizes the usefulness of Marie-Laure Ryan's notion of 'embedded narratives'-'any storylike representation produced in the mind of a character and reproduced in the mind of the reader' (Style 20[1986] 320).

Chapters 6 and 7 outline the approach and provide concrete illustrations of the kind of narrative comprehension which it allows. Palmer illustrates his theses throughout the book by means of a small number of examples, which are extracted mostly from works by Thomas Pynchon, Gustave Flaubert, Jane Austen and Charles Dickens (Pip, Emma Woodhouse, Olympia and Emma Bovary somewhat become the protagonists of Palmer's theses). The latter are taken up again in the last chapters (the book should be an aid to the teaching of practical criticism (p. 6)) to demonstrate that passages which classical narratology would have classified under the heading of 'inner thought' are in fact 'presentations of mental functioning and of dispositions to behave in a certain way' (p. 171). Consider the following passage, divided into its component actions:

(a) Not that Emma was gay and thoughtless from any real felicity; it was rather because she felt less happy than she had expected.

(b) she laughed

(c) Because she was disappointed;

(d) and though she liked [Frank Churchill] for his attentions, and thought them all, whether in friendship, admiration, or playfulness, extremely judicious, they were not winning back her heart.

(e) She still intended him for her friend. (p. 170)

The passage, Palmer argues, presents 'thought report of a dense, complex layering of a wide variety of states of mind that comprises a causal network of reasons and motives for actions'. Emma's mental functioning relates to the 'management of her relationships with others' (p. 172). The passage is a matter of behaviour, too, and not solely of consciousness. It illustrates 'the dialogic nature of consciousness': the approach outlined in the book allows us to perceive that 'Emma's thoughts are in a fundamentally misconceived dialogue with what she imagines to be Churchill's thoughts' (p. 174).

As an unavoidable result of Palmer's introduction of 'some unfamiliar ideas into narrative theory' (p. 152), parts of the book read as a review of literature (for example Chapter 6) in which one is presented with a repertoire of 'cognitive science techniques relating to narrative comprehension' by authors such as David Herman, the editor of the 'Frontiers of Narrative' series in which Palmer's book is included, Manfred Jahn, Mark Turner, Gilles Fauconnier, Monika Fludernik and Catherine Emmott (pp. 174-80). Although the concepts which the book imports are sometimes rebarbative (for example when part of the mental functioning of characters in the novel is described in terms of 'impression management' which readers should learn to identify), Palmer's approach has interesting implications for the history of the novel. For example, it is worth reflecting on the suggestion that narratology's reliance on speech categories has given a picture of English Literature as consisting 'to a surprisingly large extent of thoughtful characters in scenes of lonely self-communion' (p. 58). Certain authors may have been excluded in so far as their mode of presentation of characters does not lend itself to demonstrating the perceived 'verbal bias'. Adopting Palmer's methodology could also allow us to re-evaluate the modernist emphasis on 'stream-ofconsciousness' as being supposedly more realist than careful plotting (p. 91).

The literary representation of consciousness, and specifically of interpersonal forms of consciousness, is also taken up in George Butte's monograph I Know That You Know That I Know: Narrating Subjects from Moll Flanders to Marnie, albeit in a phenomenological idiom, rather than the cognitivist one. Butte's work frames its contribution to narrative studies as a reconsideration of modern narrative's arguably most privileged object, human subjectivity and consciousness. There are two aspects to its argument, one historical and one theoretical. The historical argument identifies an important paradigm shift that occurs in the English novel around 1800: 'a sea change in the representation of consciousnesses in narratives in English becomes visible in the time of Jane Austen' (p. 25). The change is that human subjectivity comes no longer to be depicted as a state of interiority, autonomy and fundamental isolation, but as 'deeply intersubjective', as constitutively entangled in a 'web of partially interpenetrating consciousnesses' (p. 28). According to Butte, subjectivity in novels comes to take the form of a three-tiered consciousness: a consciousness of one's own consciousness and gestures, a consciousness of the other's consciousness of one's consciousness and gestures, and a consciousness of the other's mirroring, representation, and appropriation of one's consciousness and gestures. These three levels of consciousness, schematically put, are the three levels of knowledge indicated in the title of Butte's book: I know that you know that I know. Butte is not interested in the early nineteenth-century emergence of this tripartite schema as a historical phenomenon and he admits that he can offer no reliable explanation for why it appears at the moment he says it does. Instead, he regards the schema as a fundamental reconceptualization of the self and of community as intersubjective. The subject, he maintains, comes to be depicted in literary narrative as a consciousness of something beyond itself, and as a tangible and visible body, which is to say as a series of perceptions, speech acts and physical gestures that are directed outward at another person and witnessed by that person. As a form of consciousness, therefore, it is shown to be a registering and appropriating consciousness of the other, of the other's consciousness, body, speech acts, gestures and appropriations.

'To conceptualize credibly this sea change in narrative practice', Butte maintains, 'requires a different kind of reading of consciousness, a re-formed phenomenological reading, by means of a poststructuralist phenomenology' (p. 4). Butte finds a model for such a reading in the work of Maurice Merleau-Ponty, and his discussion of Merleau-Ponty makes up the second important aspect of his book, its theoretical polemic. 'For Merleau-Ponty', he writes:

the [intersubjective] process begins when a self perceives the gestures, either in body or word, of another consciousness, and it continues when the self can perceive in those gestures an awareness of her or his own gestures. Subsequently the self, upon revealing a consciousness of the other's response, perceives yet another gesture responding to its response, so that out of this conversation of symbolic behaviors emerges a web woven from elements of mutually exchanged consciousnesses. (p. 28)

This kind of intersubjective web (the metaphor is Butte's, not Merleau-Ponty's), built out of reciprocal gestures and perceptions, is what Butte also finds depicted in the eighteenth- and nineteenth-century English novel (as well as in classical Hollywood cinema), which to him confirms the relevance of phenomenological criticism to the analysis of realist narrative. Realist narratives depict human consciousness as intersubjective, he argues, and intersubjectivity in turn takes the form of a narrative. Butte locates the essential link between the two in the shared element of time, since any intersecting and engaging between subjects necessarily occurs in temporally discrete instances (I know, you know, I know, and so on), which is to say as a story.

A key modification of alternate theories of intersubjectivity that Butte finds in Merleau-Ponty's writings is that for Merleau-Ponty the other never becomes wholly transparent to the self, but is also never wholly opaque. A mutual perception and acknowledgement may take place between the self and the other via language and physical gestures, which for the self is a first step towards an awareness of something beyond its own consciousness. For Butte, this awareness is neither a full self-transcendence by the self nor a full merging between self and other, but only a partial merging, a reciprocal contact built out of mutual perceptions of the other's perceptions. One theoretical significance Butte finds in the intersubjective model he derives from Merleau-Ponty is that it allows him to circumnavigate what he sees as two opposing absolutes governing twentieth-century discourse on subjectivity: phenomenological literary criticism, on the one hand, and Lacanian psychoanalysis, on the other. According to Butte, traditional phenomenological criticism (associated by him with Georges Poulet and the early work of J. Hillis Miller) 'idealistically' conceives of the subject as autonomous, transcendent (of history and social contexts) and entirely transparent (to other subjects), while Lacanian criticism (associated by him with the work of Kaja Silverman, Slavoj Z? iz? ek and Lacan himself, among others) 'nihilistically' conceives of it as opaque, fragmented, and as entirely constituted in the images and discourses of the other. The modified phenomenological criticism advocated by Butte would allow him to avoid the limitations he associates with each of these two approaches: the essentialism and universalism of traditional phenomenology (by locating the subject in a specified place and time), and what he sees as the Lacanian disregard for the actual body (since for Merleau-Ponty, intersubjectivity takes the form of physical gestures as much as of language). At the same time, Butte acknowledges his intellectual debt to both approaches: to the traditional phenomenological emphasis on consciousness, intentionality and the body, for instance, and to the Lacanian emphasis on the subject's constitutive mirroring of others' bodies, gestures and discourses.

Following its extended theoretical Introduction, I Know That You Know That I Know is divided into two sections. The first outlines the historical shift in literary subjectivity that Butte dates to around 1800 by means of comparative analyses of a series of paired eighteenth- and nineteenth-century English novels: Moll Flanders and Great Expectations, Pamela and The Turn of the Screw, and Tom Jones and Middlemarch. The second section is organized generically. It examines complex intersubjectivity in comedy and masquerade, focusing on Emma, The Awkward Age, Jane Eyre and films that include Howard Hawks' His Girl Friday, Hitchcock's Marnie and Woody Allen's Broadway Danny Rose. While the book's individual sections all give the reader much to think about, the relation between the opening theoretical discussion of phenomenology and the subsequent readings of the novels remains ambiguously bilateral. The two would appear to be circularly related, as mutual confirmations, since phenomenological theory and literary interpretation each come to serve as evidence for the other. For instance, Butte maintains that the complex intersubjectivity he finds in nineteenthcentury literature 'posits Merleau-Ponty's separated but initiating and embodied subjects' (p. 27), and he conversely claims that the historical 'change in the way stories narrate human consciousness . . . would not be visible without the newly ground, or at least newly reground, lens of phenomenological reading' (p. 236).

Like many of the works reviewed in this chapter, Fotis Jannidis' Figur und Person: Beitrag zu einer historischen Narratologie-which translates as Character and Person: A Contribution to a Historical Narratology-frames its polemic as a movement beyond what it perceives as the stasis of traditional structuralist and textually-based analyses of narrative. Jannidis attempts to provide a trans-historical, cross-cultural theory of character in literary narrative. Character has been a largely overlooked topic in narrative theory, he maintains, in part because structuralism has not been able to provide the theoretical means with which to account for it. And while it has been the object of various individual studies from various approaches (for instance, analyses of particular archetypes), it has thus far not been subjected to a unifying, coherent narratological theory. This alleged lack is what Jannidis' book sets out to fill.

Closely related to Palmer's concern and approach, the main aspects of Jannidis' problematic include the question of how a literary character becomes intelligible as such, the question of how its various characteristics constitute a unified entity distinct from other entities, the Aristotelian question of the relation between character and plot, the potential usefulness of character typologies, the unique contribution characters make to a literary work's overall meaning, and the role of characters as points of identification and orientation for the reader. In addressing the first of these aspects, Jannidis takes as point of departure the question of how readers reliably know that a given set of verbal cues in a text refers to a particular character rather than to an object or to another character, even though those cues may in themselves be inadequate for exclusively specifying the character (since they could potentially be referring to another character) and even though the character could potentially be referred to by an entirely different set of cues. This question (which is also Palmer's) leads Jannidis to conclude that literary characters are never reducible to sets of linguistic signs and symbols in the text, but are rather something a reader must intuit out of the signs he or she is given. Character is for him an aggregate phenomenon, partially constituted out of textual signs and partially a mental construct built by a reader in response to those signs: 'character is a model reader's mental model, which is formed incrementally over the course of the text's progression' (p. 240, our translation). In its emphasis on the reader's interaction with the text, Jannidis' definition of character is not unlike the redefinition of narrative provided by Ryan and other contributors in Narrative across Media. Like Ryan, Jannidis advocates moving narrative theory in the direction of cognitive science, thereby wanting to add new complexity and depth to what he considers the rigorously theoretical but pragmatically limited models provided by traditional structuralist narratology.

The definition of character as a 'model reader's text-based mental model' effectively joins together the book's two organizing terms, Figur and Person, in so far as Figur is indicative of text and Person is indicative of the reader's mental intuitions of characters, intuitions that are somehow like his or her apprehensions of actual persons. For Jannidis, the joining together of Figur and Person is at bottom a synthesis of structuralism and cognitive science, and ultimately a transcendence of each of their limits:

the essential reference point [for my project] is two approaches to character which can schematically be summarized as (post)-structuralist and cognitive. According to the first approach, character is a rhetorical phenomenon constituted out of semantic markers clustered around a proper name. According to the second, it is a mental representation by an empirical reader, formed on the basis of cognitive schemas. An approach contrary to both these approaches will be developed here step by step, on the basis of the analysis of individual textual passages. (p. 109) Jannidis' approach is less a contradiction of the two approaches he mentions than a fusion of them: 'the theoretically ambitious reaction [to the stasis of structuralist analysis] consists in taking cognitive research results and applying them to the analysis of texts' (p. 9).

In the claim that character is a mental concept formed or shaped (gebildet) by a reader, Jannidis attributes a role to the reader that goes beyond merely intuiting or inferring a meaning (in this case, character) from the verbal cues provided by a text. The reader, he suggests, adds something constitutive to what is in the text. So character, in so far as it becomes intelligible as such, must be thought of as a combination of text and readerly addition. What the reader adds is what Jannidis calls Weltwissen, his or her knowledge of both the representation's given fictional world and the actual world, which includes empirical experience and various historically and culturally determined forms of knowledge (for example, the prevalent folk psychologies of the day). By relying on these external sources of knowledge for guidance, the reader

is able to fill textual gaps, and to resolve inconsistencies and ambiguities in the textual signs he or she reads, thereby allowing those signs to cohere into an intelligible person who is clearly distinct from other persons and from objects. So while Jannidis insists that literary characters and real people are categorically dissimilar entities, he also maintains that our ability to apprehend characters necessarily depends on our knowledge of actual persons and the actual world, knowledge that is always to some degree historically determined. For him, literary characters are neither textual signs nor living people, but combine aspects of both in so far as they are at least partially constituted out of the reader's intuitive sense of what living people are like. Jannidis maintains that his definition of character (as an aggregate of textual signs and readerly intuitions) is a synthesis of both these aspects. Addressing the paradoxical relation between, on the one hand, the categorical difference between literary characters and real people, and, on the other hand, the similarity between our apprehension of characters and our apprehension of people, he writes:

Purist approaches, which want to see character as simply a textual phenomenon, are as useless for convincingly accounting for this relationship as are those approaches which presume too close a proximity between characters and real life. Only the concept of [character as] a model reader's text-based mental model can properly apprehend the role played by Weltwissen in the constitution of such a model, as well as apprehending the principle difference between a character and a living person. (p. 243)

By taking culturally and historically determined forms of knowledge and experience into account as an additional ingredient in the construction of character, Jannidis sees his project as combining structuralism and cognitive approaches with the insights of historicism and cultural studies. Presumably it is this additional perspective that prompts him to call his project (in his book's subtitle) a historical narratology.

Jannidis' book is divided into five chapters. The first chapter is focused on narrative communication and advocates the turn to an 'inference-based' (rather than 'code-based') understanding of it. The second chapter addresses the apprehension of character as a distinct problem for literary criticism; it asks very seriously the question of how we know that a character is a character and not another kind of entity. The third chapter attempts to answer this question by outlining specifically how an entity in a represented fictional world is designated a character (rather than something else), and how it may be referred to by various codes and yet reliably distinguished from other characters and objects. The fourth chapter then proposes a theoretical prototype of character, based on Jannidis' understanding of character as a reader's mental model: very briefly summarized, character is defined as an entity having a visible and tangible exteriority (accessible to other characters, narrators and readers) as well as a presumed interiority (that is potentially accessible). Both interiority and exteriority have identifiable long-term and short-term properties. In addition, character implies a capacity for action, motives, intentions and being subjected to external influences. This prototype is then developed in the fifth chapter, which discusses how character complexity and depth is achieved through different forms of accessory information and characterization, the introduction into the plot of various kinds of motives, and the prompting of reader identification by means of textual codes. In taking these carefully orchestrated steps in the construction of character and of a theory of character, Jannidis supports his claims with a series of quotations that-despite the stated cross-cultural and transnational ambitions of the argument-is drawn almost exclusively from German sources. Some of his examples refer to widely known texts, for instance Thomas Mann's Buddenbrooks, E.T.A. Hoffmann's 'The Sandman', Murasaki Shikibu's Tale of Genji and the fairy tales of the Brothers Grimm, but most of his references will likely be unfamiliar to non-German and non-Germanist readers.

Among the classical narratological concepts at issue in Alan Palmer's Fictional Minds is that of focalization, which was first introduced by Genette in his extended analysis of A ` la recherche du temps perdu in Figures III (Seuil [1972]), but which was subsequently both enlarged and decried by a number of narratologists. Focalization refers to the narrative process which attributes to one or more characters what was previously called a perspective, an angle of vision, or a point of view on the events which are related. In Palmer's review and vocabulary, one of the limitations of the concept is that it was well suited for acts of perception (focalization answers to the question 'who sees?'), that is, to only one aspect of mental functioning, but much less so for cognition, emotion, or any other mental activity in which fictional minds are engaged.

Like most other classical narratological concepts, focalization is built on the primacy of the individual and of consciousness, whereas more recent concepts, such as, for example, 'embedded narratives', 'doubly embedded narratives', or 'situated identity' (pp. 230-9), are better able to account for the socially situated and dialogic nature of consciousness.

It is not a coincidence, however, if in Fictional Minds Palmer cites Marcel Proust's narrator, in support of his arguments about the 'socially situated' nature of consciousness, first with respect to the way in which 'it is only with the passions of others that we are ever really familiar, and what we come to discover about our own can only be learned from them' (Swann's Way, trans. C.K. Scott-Moncrieff and T. Kilmartin. Vintage [1996] 154, quoted in Palmer, p. 134). For, as the second, augmented edition of Brian Rogers' The Narrative Techniques of A ` la recherche du temps perdu shows, narrative devices in Proust's oeuvre have greatly contributed to an interrogation of the primacy of consciousness, whether they be analysed in terms of 'double vision', as in the 1965 first edition, or, following Genette, in terms of 'dual internal focalization' in the 2004 one. Turning to Rogers' rich analyses of Proust's narrative techniques casts some doubts on the grievances, current in 'cognitivists' narratology, against the predominance of the paradigm of individual consciousness in classical narratology.

Proust's novel continues to stand as one the most influential works for twentieth-century narrative theory (the revision of Rogers' book is justified by the advances in narratological and genetic approaches to Proust, which have culminated in the 1987-89 Ple iade edition of the work, and recently in the Dictionnaire Proust by Annick Bouillaguet and Brian Rogers, eds. (Honore Champion [2004]). Whereas Palmer's prime concern would appear to lie in the theory of fictional minds, which is graspable almost independently from any references to works of fiction (the aim is to bridge an unwanted gap between fictional and 'real-mind' discourses), in Rogers' book, as in many works of Proustian scholarship, theorizations are subordinated to the novel, if only because the latter unusually comes with its own considerable share of 'theories'. Just exactly which voice (the narrator, the characters, the author) is responsible for them and which status one should attribute to them, are some of the questions Rogers' book helps to explore.

The premise of the study is that A ` la Recherche 'is a demonstration of a philosophy and an esthetic' (p. 144), and that narrative techniques are comparable to the role of arguments in reasoning. That Proust's novel leaves so little space to theorization might be partly due to its plot, which integrates 'experiments in fiction present in all the works preceding it' (p. 12). The novel presents 'the narrator as a character of fiction' whose 'journey from childhood to approaching death is the subject of his novel' (p. 12). The conceptions of reality and of art which the narrator successively maintains are narrated retrospectively 'as stages in the journey through life', in the form of memories, which have a dreamlike character. The perspective from which the memories are told, through images and colours, correspond to conceptions of realities adopted and given up by the successive 'moi', which are combined in the denouement, whereby the story of the fictional character (the narrator) is placed in 'a framework of poetry' (p. 12).

Part One explores the 'development of montages combining different genres and focalisations', found in Les Plaisirs et les jours, among others. Rogers shows that in these texts, as well as in Jean Santeuil and in the sketches of Contre Sainte-Beuve (all Bibliothe` que de la Ple iade [1971]), Proust experiments with chronology and focalization, and with the functions of philosopher, historian, literary critic, journalist, chronicler, novelist and poet, each of which corresponds to a style of writing. Recent intertextual studies have shown the extent to which Proust imitates the techniques of a wide range of works by seventeenth- to nineteenth-century novelists, such as Balzac, Dostoevsky, Barby d'Aurevilly and Me rime e, among many others. In the later novel, among the 'stages in the journey through life' are the experiments in fiction (mostly with the genre of the confession, the pastiche and imitation), which characterize Proust's early work and which form part of A ` la recherche, most importantly the experiment with focalization.

Part Two analyses the structure of A ` la recherche and identifies the principal narrative techniques employed to achieve the synthesis of poetry and fiction, who, according to Rogers, the novel accomplishes. 'Dual focalisation' is the technique which Proust uses to juxtapose the perspective of the narrator and the artist:

One perspective reflects the viewpoint of the artist who relives the episodes of the drama and recognises them as stages leading to the composition of his work; the second presents the same episodes in the perspective of the man unaware of his vocation or unwilling to investigate the significance of sensation and involuntary memory. (p. 121)

Rogers enumerates and analyses the 'variations of perspectives', which are used to emphasize the dialectic between the philosophical and aesthetic conclusions attributed to the artist and the former selves (devices such as the use of letters, the delegation of perspectives in 'chain of perspectives', interrogations and enquiries, etc.). The 'fictional drama which shows

an artist's attempt to compose a novel' is made of the successive phase of evolution of a man through Time and Space.The 'kaleidoscope of perspectives' are those of the man which evolves in Time, but they are reflected through the 'double internal focalization' of the artist, whose perspective is that of dream and memory which opens the novel, and which corresponds to the aesthetic perspective of the final revelation. The successive perspectives of the 'I' and the 'artist' go together with narrative 'frameworks combining a variety of genres', such as the promenades, the leitmotifs, the Salons and Conversations, among others. That the novel culminates in a synthesis, a reconciliation of fragmentary perspectives, and of 'fiction and poetry', as Rogers claims, is still a controversial issue in Proustian narratological and genetic scholarship. Nevertheless, by linking the combination of genres and experiments in fiction, characteristic of the early work, to the device of 'double internal focalization' and the variation of perspectives, Rogers shows that narrative models of what would undoubtedly be one instance of 'classical narratology' need not be those of an individual consciousness. The 'consciousness' which emerges from Proust's work is one dispersed across a variety of genres, voices and surmounted intellectual perspectives, and not contained by an 'individual'.

In his article entitled 'Omniscience' (Narrative 12:i[2004] 22-34), Jonathan Culler suggests that his titular concept is not a very useful one since it 'obfuscates the various phenomena that provoke us to posit the idea' (p. 22). Its most problematic aspect is that it is based on an analogy between God and the author, without there being a theological ground to the analogy, since theological debates link the problem of omniscience to that of free will. Culler considers, among others, Meir Sternberg's defence of the concept in Expositional Modes and Temporal Ordering in Fiction (IndianaUP [1978]), who argues that the expression 'the omniscient author' is a 'self-implicative attribution'. The author knows everything there is to know about the world he or she has created and the variations in the extensiveness of what is communicated provide the principle for a typology of narrators. Palmer discusses the omission of informations by the narrator, such as for example, in Henry Fielding, under the heading of paralipses (Fictional Minds OhioStateUP [2004] pp. 240-4) that issue under the heading of paralepsis to refer to the narrator in Fielding, who tells very little). However, what are the limits of that knowledge? How are we to conceive of it? Sternberg is probably referring to what is in fact 'a product of conventional performative power of language, or at best, omnipotence' (p. 24). The problem with the idea of the all-knowing narrator is that it confuses the artistic choices made by the author with 'the decisions of an imagined narrator' (p. 25).

Theorists who endorse the concept of omniscience treat narrators as persons, which leaves them with only two possible models: mortal and divine persons (p. 25). The two models provide a limited view of the kind of knowledge which can be attributed to narrators: it is either full or partial, according to unclear lines of separation.

The extension or limitation of knowledge is therefore not a useful point of entry into the problem of the narrator, and the article examines four sorts of phenomena, which theorists have described under the heading of 'omniscience', in order to find alternative descriptions of them. The range of phenomena comes very close to some of the concerns of the studies reviewed in this chapter: (1) the performative authoritativeness of narrative declarations; (2) the reporting of innermost thought and feelings, which are usually inacessible; (3) authorial narration (which seems to be a synonym of metalepsis, since it refers precisely to the moments 'where the narrator flaunts her godlike ability to determine how things turn out' and is exemplified by Diderot's Jacques le fataliste, a classic case of such flaunting); and (4) the synoptic impersonal narration of the realist tradition (p. 26).

Culler shows how the effects otherwise attributed to knowledge are in fact not a matter of knowledge, but, in the case of performative authoritativeness, for example, bespeak of the conventions through which we attribute truthfulness to declarative sentences within narrative fiction. The recourse to an omniscient narrator can also be explained by the propensity with which, when faced with the reporting of a character's thought to which we cannot have access, we invent a consciousness as the source of this knowledge. The access to a character's thought need not be conceived in terms of

knowledge, but may be imaginatively framed in other ways, such as, for example, telepathy, which entails the relay, the transposition and the translating of thought into free indirect speech away from the unique source of narrative, which the narrator only apparently constitutes. Culler comes close to Rogers, who argues that the narrator in A ` la recherche is a 'fictional character', when he illustrates the imaginative use of alternative narrative means with the story of Un Amour de Swann (in A ` la recherche du temps perdu. J.-Y. Tadie , ed. Gallimard. Bibliothe` que de la Ple iade. Vol. 1 [1987-89] 4 vol.) in Du cote de chez Swann (in A ` la recherche du temps perdu. Gallimard. Bibliothe` que de la Ple iade. Vol. 1 [1987-89] 4 vol.), since the latter 'depends on Marcel's imaginative, telepathic retailing of information that the novelist invented (just as he invented Marcel and Marcel's account of himself)' (p. 30). To consider telepathy as a useful concept is part of the broader suggestion of Culler's article and of the works reviewed in this chapter, that we should find other means of description and understanding of narrative agents (ones that could be 'a recorder, a presenter of signs, a transmission device', for example). Like Rogers, Culler interestingly comes back to his earlier narratological work, by reflecting on and responding to criticism addressed to his use of the concept of omniscience.

Ge rard Genette's Figures III (Seuil [1972], trans. Narrative Discourse:An Essay in Method. CornUP [1980]) is perhaps one of the most influential texts in contemporary narratology, judging by its ubiquity, long after its publication in French and in English, in the pages of the books here reviewed. In Metalepse: De la figure a` la fiction-which translates as Metalepsis: From Figure to Fiction-Genette himself sends us back to his 1972 work, in order to situate his annexation of the rhetorical notion of metalepsis to the field of narratology. Metalepse is an expanded version of Genette's contribution to a 2002 conference on metalepsis, a conference which bespeaks of the term's topicality (La Metalepse aujourd'hui, the papers of which are published in Jean-Marie Schaeffer and John Pier eds. Metalepses: suspension volontaire de la credulite. Centre National de la Recherche Scientifique [2004]).

Genette gets his definition from the eighteenth- and early nineteenth-century rhetoriticians Dumarsais (Des Tropes [1730]) and Fontanier (Les Figures du Discours [1821-7]), for whom metalepsis is a synonym of metonymy, with

some variations in their respective understanding of the causal relations between terms which the latter implies and with some divergences as to whether metonymy can be a noun or a proposition. Dumarsais and Fontanier, however, do agree on a particular case, namely the metalepsis of the author, which is a figure whereby the author seemingly intervenes in the action he is describing: 'instead of simply narrating something which is being done, or which has been done, one asks, orders that it be done' (our translation, p. 13).

Genette pushes aside the problem of the proximity between metalepsis, metonymy and hypotyposis, which is at issue in classical rhetoric and limits the use of the concept of metalepsis to 'a manipulation . . . of the particular causal relation which unites, in one way or another, the author to his/her work, or, more widely, the producer of a representation to the representation itself'. Metalepsis will come to extend to 'many other modes of figural and fictional transgressions of the frames of narratives' (p. 14), which the book discusses.

'Un Amour de Swann' in A ` la recherche du temps perdu presents a prime example of a metalepsis of the author. The narrator pretends to have heard all the details of Swann's life, while he suppresses all traces of the source of his knowledge of the events. The narrative is metaleptic, since the narrator 'invents' what he narrates (p. 15). What is identified as a localized figure within a larger whole can be 'enlarged' in fiction, and it is these 'aggravated' cases towards which Genette turns.

Metaleptic fictions engage the two levels at which fiction is played out: the diegetic and metadiegetic levels. The metaleptic relation almost always functions as a relation between a so-called real level and one assumed to be fictional: for example, the level at which Scheherazade distracts the king and the level of the tales she tells him. Another example is that of the short story 'The Continuity of Parks' by Julio Cortazar (Armes secretes [1959]), where a fictional character reads a novel and is eventually killed by one of the characters in the novel he is reading. The transgression between levels can occur in two directions-when 'real' life enters the fictional world, or vice versa. Classical rhetoric has considered only the first kind and Genette's Metalepse is concerned with metaleptic fictions whereby fiction inserts itself in 'real life'. For Genette, the latter could not have been possible before Romanticism, since it rests on a modern idea of literary creation, which grants autonomy to products of the imagination (p. 27).

The rather brief theoretical opening discussion is followed by a great number of examples intended to specify further the notion of metalepsis, from John Fowles' The French Lieutenant's Woman [1969], in which the narrator sits in a train compartment with his hero and wonders what to do with him (p. 35), to the relation between character and actor in the television series Dallas, to Diderot's discussion of Greuze's paintings in the Salons (Garnier [1959]). Given the variety of examples and the prominence of cinematographic ones, it is striking that the framework of Genette's discussion remains quite imprecise (even if the notion is said to fall within the competence of the analysis of narratives, of rhetorical figures and of the theory of fiction (p. 7)). The fictional 'dramatic form' (the play within the play) is singled out as one of the richest metaleptic devices. In The ophile Gautier's novel Capitaine Fracasse ([1861] Garnier [1961]), the narrator reports the thoughts of the characters of the fictional play, rather than those of the novel, for whom the play is 'real'. Through Genette's discussion of Capitaine Fracasse, the trope of metalepsis turns out to be relevant to problems raised in Fictional Minds, concerning the narrator's access to the subjectivity of others, and the importance of our understanding of 'thought report'. What is 'fictionalized' is here the narrator's access to fictional minds. Gautier's novel reminds us, following the narratologist Ka te Hamburger, that it is only fictionally that we can have access to someone else's subjectivity. For the analyst of metaleptic fictions, if the narrator has access to the subjectivity of the fictional actors, it is because he invents it (p. 47). The category of 'invention' bypasses the problem of access to 'inner thought' so decried by Palmer. Genette's approach to the problem of characters through the category of metalepsis might begin to weaken Palmer's convincing argument against the 'verbal bias' of classical narratology.

This is because the metaleptic operations which Genette enumerates are far from being only 'verbal' and are in fact in a great number of cases cinematographic ones (although cinema too is 'verbal'). For example, metalepsis can take the form of a 'metaphysical joke', such as in Howard Hawks' The Girl Friday [1940], wherein a character is considered to look like the actor who plays him (p. 58). In Woody Allen's The Purple Rose of Cairo [1985], the movements by the characters from the screen to the real world and vice versa are metaleptic actions. Another metaleptic possibility consists in making an actor play more than one role, such as in Vertigo [1958].

Under the heading of metalepsis, the book takes up again elements of analysis scattered throughout Genette's earlier work. For example, it comes back to Genette's essay entitled 'Frontie` res du re cit' (Figures II. E d. du Seuil [1969]), which discussed the distinction criticized by Plato, concerning a passage from the Iliad, where Homer assumes the voice of Chryses rather than reporting his speech in indirect discourse. In the context of the analyses devoted to metalepsis, the passage from one mode to the other now becomes a matter of the crossing of an impassable threshold.

However, in so far as Genette's book consists in a long list of examples, it does not elaborate sufficiently on the reconceptualization that takes place. The notion of metalepsis indeed goes together with other 'classical narratological' notions, that between 'je-narrant' and 'je-narre', notably around autobiography, where the 'I' is necessarily divided. In fact, not only is the pronoun 'I' itself a 'metaleptic operator', but all self-referential utterances are metaleptic (a suggestion that comes close to Marie-Laure Ryan's analysis of metalepsis in mathematics and logic in her essay 'Metaleptic Machines', which is reviewed below).

Genette ultimately extends the notion to cover all fiction: 'This perpetual and reciprocal transfusion of the real diegesis to the fictional one, and from one fiction to the other, is the very soul of fiction in general, and of all fiction in particular. Every fiction is traversed by metalepsis' (p. 131). With this generalization, one is reminded of Genette's illuminating essay on 'La rhe torique restreinte' from Figures III (E d. du Sevil [1972]), which presents the history of rhetoric as a gradual reduction from many figures to two, metaphor and metonymy. Even though fiction represents for Genette an

'enlargement' of the figure of metalepsis, a 'reinforcement' of it, one wonders whether one figure, however protean it appears to be, can indeed become the matter of all fiction, without annulling itself.

Marie-Laure Ryan's essay 'Metaleptic Machines' (Semiotica 150:i-iv[2004] 439-69) also explores the resourcefulness of the notion of metalepsis for the analysis of narratives and makes it a central operation of literature. The article extends further the concerns of narratology beyond the printed book, which Ryan's anthology Narrative across Media also sets out to do, by exploring the 'special affinities' that metalepsis has with computers and its 'manifestations in digital culture' (p. 439). Ryan relates her interest in metalepsis back to, among other sources, Genette's introduction of the idea of

diegetic levels 'to describe the phenomenon of the proliferation of stories within stories' (p. 439). Every story involves at least two levels (a 'real-world level, on which an author communicates with a reader, and a primary fictional level, on which a narrator communicates with a narratee within an imaginary world' (p. 439)). Metalepsis denotes possible kinds of transgressions of the borders between these distinct narrative levels, and Ryan argues that it is possible to offer a typology of these transgressions.

In some instances, jumps between narrative levels have no serious consequences. For example, in rhetorical metalepsis, such as one finds in Diderot's Jacques le fataliste [1796], where an authorial figure wonders what he should do with his character, the intervention is only a 'temporary breach of illusion'. Here, 'the operation ends up reasserting the existence of boundaries' (p. 441). The ontological type of metalepsis is more disturbing, since 'it opens a passage between levels that result in their interpenetration, or mutual contamination' (p. 442). Woody Allen's 'The Kugelmass Episode' is a classical example of the ontological transgression of narrative levels. In that short story, the protagonist is transported into the world of Flaubert's Madame Bovary [1857]. The academic study of the book in universities is henceforth perturbed, since Kugelmass replaces Le on and Rodolphe as Emma's lover. The metaleptic operation brings about the entanglement

between the 'imaginary world of Madame Bovary' and the 'fictional ''real'' world of Kugelmass': 'should we say that Emma become real when she visits New York in the primary fictional world, or that Kugelmass becomes fictional when he visits Emma in the France of Flaubert's novel?' (p. 443).

It is a premise of Ryan's article that the 'playful subversion of narrative logic' in literary manifestations of metalepsis is a 'harmless' game compared with metaleptic activity in the arts, in logic, mathematics, language and science: 'when it affects these disciplines, metalepsis carries far more disturbing epistemological consequences . . . for even if the metaleptic processes . . . do not invalidate logic, mathematics, and the scientific method, they certainly raise troubling questions about the limits of these modes of thinking' (p. 445).

Ryan initially presents the literary manifestations of metalepsis in terms of the metaphor of the computer stack (to illustrate narrative levels), but as we move to the more 'disturbing' occurrences of metalepsis in non-literary disciplines, the models employed to describe metalepsis are diversified (for instance, the model of the tangled hierarchies, or of the strange loop); they themselves become occurrences of metaleptic phenomena (Ryan discusses Kurt Go del's 1931 paper 'On Formally Undecidable Propositions' and Alan Turing's theorem about the 'halting problem'), and eventually, when computers, computer games, digital texts and virtual reality are at issue, they stand as the very medium of the most troubling kind of metaleptic activity. For 'the stratified architecture of the computer . . . creates opportunities for metaleptic operations of a technological nature (viruses, programs that operate on themselves), as well as for manipulations that lead to ludic and artistic effects (p. 451). Ryan's article does contain interesting examples of digital texts produced, among others, by the school of poet-programmers such as John Caley, examples which transgress the border between surface texts and text-codes. It is odd that the metaleptic possibilities in the 'artistic domain' should be exemplified by the theatrical work of Jean Genet's Les Bonnes, where the characters of the play themselves stage a play and commit a fictional murder that has 'real' consequences in the first narrative level. For the notion

of metalepsis elaborated by Genette, to which Ryan is admittedly indebted, has a lot to do with the self-referential tricks, the 'metaleptic behaviour' which the theatre allows, which gives it more importance than that of a mere example.

The extension of metalepsis which Ryan proposes is suggestive, notably as a means of reframing the classical problem of scientific objectivity (of the relation between the observer and the phenomenon under observation) in terms of unavoidable 'metaleptic contamination' (p. 448). The narratological model is, however, sometimes overshadowed by the pathological one: metalepsis is a 'benign or invasive growth', a cancer that destroys the ambitions of mathematical systems (p. 446)). Ryan opposes the playful literary manifestations (which provide a typology of metalepsis) with the more disturbing variations associated with the technological possibilities of the computer, and the mathematical and logical systems presiding over their development. Technological metaleptic activities may well be radically threatening our sense of reality; nevertheless, no amount of transumption can 'shake our conviction that we inhabit the only world that exists ''for real'', because this world is the one that we inhabit corporeally' (p. 464). It is 'our bodies', Ryan curiously concludes, that ultimately bring us back, no matter how metaleptically we have been led astray from it, to the 'ground level of reality'.

While Ryan and many of the authors reviewed in this chapter share an interest in expanding narrative studies into research areas opened by cognitive science and cultural studies, the interests of Margaret Reid's Cultural Secrets as Narrative Form: Storytelling in Nineteenth-Century America are less in narrative or in elements of narrative as such, but in the workings of a specific narrative genre, the historical romance. In her readings of three nineteenthcentury American historical romances (James Fenimore Cooper's The Spy, Nathaniel Hawthorne's The Scarlet Letter and Owen Wister's The Virginian), Reid finds a shared narrative poetics: literary representations of American culture that interweave in a single text two distinct levels of narrative. The first level is what she calls 'narrative design', an explicit and well-ordered narrative that is to some degree about the formation or elaboration of an American national identity. The second is what she calls 'secret histories', a seriesof interpolated fragments and images referring to historical moments, experiences and traumas that have largely been forgotten or repressed, and that re-emerge only vaguely and intermittently. While Reid's project defines itself primarily in historical and national terms (as a study of nineteenthcentury American literary nationalism, modelled on Benedict Anderson's Imagined Communities (Verso [1983]), it also holds interest for narrative studies in its investigation of the double narrative structure of historical romance and in its attention to the important question of what narratives do (schematically summarized, her book investigates how they work to define communities and simultaneously to encrypt memories and histories those definitions and those communities would repress).

For Reid, the first level of narrative is what she calls the official story. In the nineteenth century, she argues, American writers adopt the historical romance as a vehicle for imagining national identity, in particular at moments of significant historical or cultural transformation, for instance during the antebellum 1850s or the nationalist 1810s and 1820s. As narratives, she maintains, historical romances construct beginnings and endings, order and symmetry, and patterns of conflict and resolution. They also provide symbolic forms through which American nationalist abstractions like independence' or 'manifest destiny' can be represented. Thus, they become a convenient ideological medium through which American national self-definitions are shaped. Through their transformation of the contingency of historical events into well-ordered plots, Reid adds, romances inherently work to erase or obscure what she calls the secrets and forgotten moments of the past, specifically those moments (or experiences or memories) that do not fit into the narrative design that has been imposed. Thus, she partially reads the genre as a defensive impulse protecting against the potential return of something from the past that has not been successfully assimilated.

If the first level of narrative is that of story, the second level is that of history. For Reid, it takes the form of fragments, images, voices and material artefacts that do not fit easily or fully into the plots constructed by the narratives in which they appear. For example, Reid proposes that Hawthorne's The Scarlet Letter, written in the context of antebellum uncertainty about the future, is motivated by an anxiety about historiography itself, about being able to tell the nation's story in such a way as to be able somehow to defend against the pressures of growing civil unrest. She argues that the circulation of the scarlet letter as a material artefact throughout the novel implies a less reassuring and less predictable view of history than do the cyclical and linear theories to which Hawthorne's novel ostensibly turns as narrative solutions for thinking about America's past and future. To give one other example: Reid's chapter on Cooper's The Spy, a novel about American national origins in the War of Independence, focuses on encoded voices in the narrative that recall the revolution not as a consensual collective action on the part of an emerging nation, but as a complicated nexus of separate, often conflicting motives and energies.

According to Reid, the relation between the two levels of narrative is complex and interactive. Paradoxically, the first level is itself what articulates and encodes the second level, even though it does not acknowledge the latter

and even though it directly or indirectly works to erase the remnants of the past that become intermittently visible or audible at the second level. The second level, by Reid's definition of it, is difficult to summarize: among other things, it includes the secrets, repressions and unarticulated alternatives of the first level narrative, the traces of past experiences and of collective or individual trauma, and remote historical moments that linger in the present consciousness. Broadly summarized, its voices and images provide Reid with a critique of, or at least an alternative to, the resolutions and designs of the first level narrative. The interaction between the two levels, Reid argues, should significantly complicate any theory of nineteenth-century narrative definitions of American national identity. Those definitions should not be reduced simply

to first level plots and designs, she argues, but must be analysed as a dynamic, conflicted interaction between what she calls 'narrative designs' and what she calls 'secret histories'.

Reid's theory of a second level of expression within a dominant narrative is avowedly indebted, among other sources, to Fredric Jameson's idea of the political unconscious and to Walter Benjamin's notions of 'story' and 'storyteller' (as distinct from narrative and narrator). The book is divided into three long chapters, each devoted to one of Reid's primary texts, and concludes by examining its nineteenth-century model's legacy in the twentieth century through discussions of Faulkner's Absalom, Absalom [1936] and Thomas Pynchon's The Crying of Lot 49 [1965]. Although Reid is foremost interested in history and in the second-level narrative as a would-be authentic repository of history, the model she constructs also has relevance to the study of narrative poetics. For one thing, it gives an example of what Marie-Laure

Ryan has called a 'One Channel' narrative medium telling more than a single story at a time. Several of the essay contributions to her anthology Narrative across Media examine how a 'Multiple Channels' medium (for instance, film, dance or installation art) can tell two competing stories at once. In her essay 'Art Spiegelman's Maus and the Graphic Narrative', for instance, Jeanne Ewert analyses how graphic novels may tell concurrent stories by means of visual images and verbal text, and how the images may provide counternarratives to the dominant verbal narrative provided in the headings and speech balloons. Similarly, Peter J. Rabinowitz's essay 'Music, Genre, and Narrative Theory' argues that the conventionally melodramatic ending staged in the Broadway musical Showboat is ironically commented on by the accompanying musical score, which Rabinowitz classifies as a narrative in its own right. In both essays, the two narratives allegedly being told are each attributed to a different medial channel. Reid, in contrast, attributes the two narratives she finds in historical romances to a single channel, the literary text. So for the interests of narrative studies, there are two important levels of narrative at work in Reid's own text, not unlike in the novels she discusses. One level tells us about the significance of what nineteenth-century historical romances tell: secret histories, embedded within dominant narratives of national identity. The second tells us about the significance of that form of telling as such.

In his Introduction to the 1999 anthology Narratologies, David Herman remarks how the theoretical models of structuralist narratology had been called into questions by Artificial Intelligence, hypertext, psychoanalysis, film studies and linguistics. When turning to Andrew Hock-Soon Ng's monograph Dimensions of Monstrosity in Contemporary Narratives: Theory, Psychoanalysis, Postmodernism, it is difficult to imagine under which disciplinary heading the items on Herman's list could be reconciled. For Ng's psychoanalytic approach to monster narratives in twentieth-century literature could not stand further apart from cognitivist approaches. While Ng, like Reid, turns to the study of a specific genre, his aim is also theoretical. Dimensions of Monstrosity proposes a theory of the monster elaborated through the Lacanian model of desire, and its reworking by Julia Kristeva and Slavoj Z? iz? ek. As a way to counteract the limits of psychoanalytic discourse for the interpretation of literature-one can always be merely carrying out the psychoanalysis of fictional characters-the theoretical framework is extended to include 'post-modern geopolitics, gender, and cultural studies' (p. 17). The psychoanalytic framework is a familiar one to readers of literary theorizations of the Gothic and the fantastic, since the latter incorporate a dimension of the monstrous in the very elaboration of their theories of the subject (notably in their inflexion of Freudian notions of the 'death drive').

Unlike previous monsters narratives, such as fin de sie`cle Gothic ones, which situate monsters in a geographical and physical 'elsewhere', contemporary monster narratives locate the monster 'within us'; the monster both inhabits

and undermines the human system. Hence, in Ng's analytic terms, it problematizes the complex 'ideological construction' of the self by society, culture and language, or the 'Symbolic order'.

Contemporary monster narratives tell us about the way in which human reality is constructed through fantasies, which disguise a Void. Monster narratives disclose the 'signifying fantasies' as fantasies and it is the effect of that traumatic disclosure on the subject which the former present. Given the Lacanian framework of the book, and how the latter has appealed to discourses on femininity, it is not surprising to read that monstrosity is intimately linked with the maternal body as a site of abjection defined as 'what has become emptied of reality' (p. 11).

The book analyses different aspects of monstrosity as they variously take form in nine narratives. Peter Ackroyd's Hawkmoor [1985] and J.G. Ballard's High Rise [1975] discuss the spatial dimension of monstrosity in Chapter 1, and Bret Easton Ellis' American Psycho [1991] and Joyce Carol Oates' Zombie [1996] provide examples of 'horror's new monsters' in Chapter 2, the serial killers. Monstrosity can also be located in the family, through the figure of the child as monster, which destroys the symbolic system in Iain Banks' The Wasp Factory [1984] and Doris Lessing's The Fifth Child [1988]. The body is obviously the site of monstrous fabrication, and Ng deals with this aspect through Patrick McGrath's Angel [1991] and Angela Carter's Night at the Circus [1984], by considering the destabilizing figure of the angel through the theme of disguise.

A great deal of emphasis is placed upon the relation between language and the monster, for one of the interesting suggestions of the book is that language turns monstrous when articulating monstrosity (the language of monstrosity is the failure to name (p. 185)). This is developed more fully in the last chapter of the book, through an analysis of Janice Galloway's 'Blood', where Ng argues that the monster is tied to language precisely while it also undermines it.

Volume 23, number 2 of the journal Literature and Medicine is devoted to the question 'Why do stories matter?' in order to reassess the role of empathy and narration in the therapeutic relation. Psychoanalysis is the obvious candidate to call for such an enquiry (the issue contains papers from a conference held at the University of Florida at Gainesville in February 2004 on 'narrative medicine and psychoanalysis'). For psychoanalysis, the editors Maura Spiegel and Rita Charon remind us in their Preface, 'telling a story and listening to a story are the most serious of human enterprises', a suggestion which is fully explored in psychoanalyst Roy Schafer's contribution, 'Narrating, Attending, and Empathizing' (L&M 23:ii[2004] 241-51).

Schafer begins by situating and reviewing his contributions to the study of psychoanalytic theorizing and modes of practices, in an important number of books and articles published between 1983 and 1997. Drawing from philosophy of mind, Schafer proposes to reconsider the psychoanalytic sessions 'as sites of thought actions, speech actions, and nonverbal actions performed by patients and analyst' (p. 241). For actions are not only diverse (it can be a matter of 'reasoning, dreaming, remembering, inhibiting, repressing, deliberately being passive, and in certain key respect having feelings' (p. 242)), but there are also many ways in which to describe what one is doing (or, as Palmer does, to describe what one is thinking). To conceive of the therapeutic relation in terms of 'action language' in this way displaces the primacy of 'Freud's model of language' for psychoanalysis and has allowed Schafer to develop a 'narratological account of psychoanalysis', in so far as to describe an action can be 'redescribed as an act of narration' (p. 242). The redescription includes the therapeutic, analytic and theorizing act, which becomes 'one form of narrative action'. Schools of analytic thought each defend their 'story lines' or 'plot lines', which are the 'metanarratives' that guide the therapeutic relation. Neo-Kleinian concepts and practices have contributed further to reinforce Schafer's views on the narrative act.

What follows from Schafer's summary presentation of his contributions to the theorization of the analytic practice also remains cursory. The rest of the article redescribes aspects of the therapeutic relation as narrative actions. 'Paying therapeutic attention' and 'empathizing' are narrative actions; they consist in forms of storytelling. The idea of narrative that emerges from Schafer's redefinition of therapeutic acts as storytelling is imprecise: it appears to denote the simple fact of 'retelling [more or less elaborately] of the analytic material at hand', which amounts to an interpretation. There is a narrative because the material has been 'transformed' in order to increase its 'relevance and intelligibility'. Storytelling, however, 'is not fictional in the fairy tale or daydream sense'. Rather, the reformulation of the material is a way of 'saying what, at that time, you believe to be true or at least a useful step toward defining what is true in the patient's conscious and unconscious psychical worlds'. In linking narratives to truth, the definition of narratives is further generalized to include 'all accounts of what is the case' which may be considered 'versions, or ways of telling and retelling' (p. 243).

Narrative transformations also occur in the act of empathizing, in so far as empathy implies the blurring of the boundaries between the self and the other. The processes through which the latter takes place (incorporation, projection, identification) harbour potential narrative transformations.

A third therapeutic act also entails potential narrative acts. The analyst rearranges, recontextualizes the material which the patient arranges according to his or her psychical worlds. Recontextualization complements what occurs in empathy. It consists in 'constructing appropriate contexts' by taking into account how 'the patient's experienced life-historical background ' is dealt with during the session. Schafer defends a 'constructivist' approach, which allows for a variety of ways in which to recontextualize, even if 'each event or action has limited narrative potential' (p. 244).

Even though the concern with narration is intimately linked with interpretation, the article touches only briefly on how the 'narrational conception of psychoanalysis' may be insufficiently scientific and empirical for some, and too hermeneutical for others. Given the complexity of the problem, it refers much too rapidly to Paul Ricoeur's controversial attempt at reconciling the two so-called poles of psychoanalysis, the scientific and the hermeneutic. As a result, the 'narrational' conception of psychoanalysis presented by Schafer would seem to denote above all the constructivist and pluralist approaches to the problem of interpretation and truth, rather than to involve precise features of narratives. It is 'narration' which would appear to take on the form of some of the components of the therapeutic relation (attending, empathizing, recontextualizing), rather than the reverse, so much so that it might be more accurate to speak of a therapeutic conception of narratives. It is nevertheless interesting to compare the clinical psychoanalytical point of view on narrative to the transformation of psychoanalytical concepts in psychoanalytic literary criticism, which Ng's monograph exemplifies.

To conclude the review of works published in narrative theory in 2004, it is worth calling attention to one more transformation of the paradigms of classical narratology, this time in the area of research on visual narrativity.

Mieke Bal, one of the most prominent scholars in the field, describes this paradigm shift in a brief article entitled 'The Changing Profession: Figuration' (PMLA 119:v[2004] 1289-91). Having become interested in the visual status of the narrative focalizer expounded in film studies, she defends the latter as a key concept to 'pictorial narratology that acknowledges the gap between literary and visual imagery' (p. 1290). Suspicious of the grievances against the

predominance of visuality in contemporary culture, she provides examples of situations where images are doing the work of narratives, notably in the political struggles of Zaire, or where it is impossible to separate clearly literary and visual media, such as in films, wherein the montage 'becomes a locus of intermediality' between the verbal and the visual media. The two modes of producing narratives can tell us about forms of sociability and subjectivity.

They are sometimes combined and 'practiced', as in the field of 'cultural memory' associated with the work of Marianne Hirsch. The article seems to be a plea against unsophisticated views of the many media in which and by means of which we live (the defence of pictorial narrative is incompatible with the simplistic rejection of television as a destroyer of the novel, for example, and-presumably too-of narratives and of narratology). The article is too programmatic to have elaborated substantially on Lyotard's notion of figuration, which it mentions, but it usefully provides theoretical and bibliographical paths for further work.

Like the work of Bal, Palmer, Ryan and of other authors whose work is reviewed in this chapter, Zenonas Norkus' essay 'Historical Narratives as Pictures: On Elective Affinities between Verbal and Pictorial Representation' (Journal of Narrative Theory 34:ii[2004] 173-206) is concerned with the predominance of verbalization, or of the category of speech, for the understanding of narratives. The unjustified emphasis on the category of speech is not contested head on, however, but rather through the reversal of a prevalent commonplace in Art History and in 'structuralist narratology' concerning the textual nature of pictures, whereby a picture is 'considered to be a more ''natural'' mode of representation than text' (p. 173) and, because of that, something that requires to be 'read', to be deciphered on the model of textual decipherment. The author proposes to reverse the terms of the commonplace, by recalling debates about the predominance of visual metaphors in narratology and to consider that 'texts are pictures'. What the inversion of the leitmetapher allows us to see, however, is not how misguidedly 'pictures' have traditionally been assimilated to texts, but rather the extent to which 'visual metaphors' (the world of 'pictures') have played a constitutive role in narratological theory (p. 175), especially as far as the metaphors related to linear perspective are concerned. The debate on the relation between the textual and the visual is narrowed down to conceptions of narratives within historiographical narratology, and to the analysis of historical narratives.

By turning to that field, Norkus' article seeks to expand a suggestion made by the historian F.R. Ankersmit that 'the study of history is more a depiction than a verbalization of the past', in order to show that one must tackle jointly 'the problem of coherent depiction of objects in space and that of the coherent narration of events in time' by comparing the methods which are used 'to construct space in pictures, and those to construct historical narratives' (p. 179). The article's central thesis is that there is a structural homology between the conventions guiding the representation of the past from the rise of

historism [sic] and the rule of linear perspective, which has been the correct way of representing space from the Renaissance onwards, until its contestation by modernism. The article provides an interesting counterpart to the

cognitivist contestation of the 'verbal bias'.

Books Reviewed

Butte, George. I Know That You Know That I Know: Narrating Subjects from Moll Flanders to Marnie. Theory and Interpretation of Narrative. OhioStateUP. [2004] pp. 272. GBP29.96 ISBN 0 8142 0945 9.

Genette, Ge rard. Metalepse: De la figure a` la fiction. E ditions du Seuil. [2004] pp. 132. pb E20 ISBN 2 0206 0130 3.

Jannidis, Fotis. Figur und Person: Beitrag zu einer historischen Narratologie. Narratologia. De Gruyter. [2004] pp. 294. E88 ISBN 3 1101 7807 0.

Ng, Andrew Hock-Soon. Dimensions of Monstrosity in Contemporary Narratives: Theory, Psychoanalysis, Postmodernism. Palgrave. [2004] pp. 240. GBP47.50 ISBN 1 4039 4446 6.

Palmer, Alan. Fictional Minds. Frontiers of Narrative. UNebP. [2004] pp. vith275. GBP34.50 ISBN 0 8032 3743 X.

Reid, Margaret. Cultural Secrets as Narrative Form: Storytelling in Nineteenth-Century America. OhioStateUP. [2004] pp. 304. hb GBP47.89 ISBN 0 8142 0947 5, pb GBP13.61 ISBN 0 8142 5118 8.

Rogers, Brian. Proust's Narrative Techniques, 2nd edn. Recherches Proustiennes. Honore Champion. [1965, revised and augmented 2004] pp. 208. E40 ISBN 2 7453 1091 7.

Ryan, Marie-Laure, ed. Narrative across Media: The Languages of Storytelling. Frontiers of Narrative. UNebP. [2004] pp. 422. hb $75 ISBN 0 8032 3944 0, pb $35 ISBN 0 8032 8993 6.

"Palmer's book develops a radically new and powerfully integrative approach to the minds presented in fictional texts. Fictional Minds promises to be a pathbreaking book that will have a major impact on several fields of study."-David Herman, author of Story Logic: Problems and Possibilities of Narrative



Book Information
ISBN 9780803218352
Author Alan Palmer
Format Paperback
Page Count 276
Imprint University of Nebraska Press
Publisher University of Nebraska Press
Weight(grams) 437g