The Language of Stories
How do we read stories? How do they engage our minds and create meaning? Are they a mental con...
160 downloads
2443 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
The Language of Stories
How do we read stories? How do they engage our minds and create meaning? Are they a mental construct, a linguistic one, or a cultural one? What is the difference between real stories and fictional ones? This book addresses such questions by describing the conceptual and linguistic underpinnings of narrative interpretation. Barbara Dancygier discusses literary texts as linguistic artifacts, describing the processes which drive the emergence of literary meaning. If a text means something to someone, she argues, there have to be linguistic phenomena that make it possible. Drawing on blending theory and construction grammar, the book focuses its linguistic lens on the concepts of the narrator and the story, and defines narrative viewpoint in a new way. Textual use of pronouns and other referential expressions also comes under investigation, including their occurrence in constructions of speech and thought representation. The examples come from a wide spectrum of texts, primarily novels and drama, by authors such as William Shakespeare, Margaret Atwood, Philip Roth, Dave Eggers, Jan Potocki, and Mikhail Bulgakov. Barbara Dancygier is Professor in the Department of English at the University of British Columbia, Canada. Her publications include Conditionals and Prediction (Cambridge, 1998) and Mental Spaces in Grammar (with Eve Sweetser, Cambridge, 2005).
The Language of Stories A Cognitive Approach Barbara Dancygier
cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Tokyo, Mexico City Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9781107005822 © Barbara Dancygier 2012 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2012 Printed in the United Kingdom at the University Press, Cambridge A catalog record for this publication is available from the British Library Library of Congress Cataloging in Publication data Dancygier, Barbara. The language of stories : a cognitive approach / Barbara Dancygier. p. cm. Includes bibliographical references and index. ISBN 978-1-107-00582-2 1. Language and languages – Style – Psychological aspects. 2. Cognitive grammar. 3. Discourse analysis, Literary. I. Title. P301.5.P75D36 2011 8080 .036–dc22 2011015550 ISBN 978-1-107-00582-2 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
for Jacek and Szymek
Contents
List of figures Acknowledgements
page ix x
Introduction
1
1
4
Language and literary narratives 1.1 1.2 1.3 1.4 1.5 1.6 1.7
2
Blending, narrative spaces, and the emergent story 2.1 2.2 2.3 2.4
3
4
Where does narrative meaning come from? Literary analysis and linguistic analysis Literature, language, and human nature Literary texts and communication Why is fiction special? Narrative and grounding Approaching narratives
Applying blending to fictional narratives Narrative spaces as mental spaces Narrative spaces – an example Emergent story
6 8 11 16 21 23 29
31 32 35 40 53
Stories and their tellers
58
3.1 3.2 3.3 3.4 3.5 3.6 3.7
Narrators, narrative spaces, and viewpoint Types of teller and epistemic viewpoint Second-person narratives The teller, the author, and the character Multiple tellers Narrative space embedding Narrative viewpoint and narrative spaces
60 64 76 79 80 84 86
Viewpoint: representation and compression
87
4.1 4.2 4.3 4.4 4.5 4.6 4.7
91 96 100 102 106 108 112
Viewpoint and representation Viewpoint compression Decompression for viewpoint Fictive vision, causation, and change The micro level, the macro level, and viewpoint compression Speech, thought, and multiple levels of representation Narrative thought and intersubjectivity
viii
5
Contents
Referential expressions and narrative spaces 5.1 5.2 5.3 5.4 5.5 5.6
6
Fictional minds and embodiment in drama and fiction 6.1 6.2 6.3 6.4 6.5
7
Deictic ground in literary discourse Mental spaces, physical spaces, and dramatic narratives Materiality of the stage and fictional minds From dramatic narratives to novelistic narratives Fictional minds, bodies, and brains
Speech and thought in the narrative 7.1 7.2 7.3 7.4
8
Compression, decompression, and cross-space mappings Proper names, frame metonymy, and the status of a character Role-value mappings as cross-space connectors Common nouns Personal pronouns, viewpoint, and the narrator Deictic I and the construal of subjectivity
Types of discourse spaces in the narrative Speaking for thinking Levels of embedding in thought representation Viewpoint compression and constructional compositionality
Stories in the mind 8.1 8.2 8.3 8.4
The linguistics of literature The storyworld reality Blending and narrative analysis A bridge to the truth
Notes to the text References Literary works cited Index
117 118 119 121 128 129 136
139 139 141 146 164 168
171 172 174 178 183
195 195 200 201 203
205 210 223 225
Figures
2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3.6 4.1 4.2 4.3 5.1 6.1 6.2 6.3 7.1 7.2 7.3 7.4
Narrative spaces (example 2) Narrative spaces in The Blind Assassin Referential connections across spaces in The Blind Assassin Intrusive narrator (example 2) Ego-viewpoints Present time and present tense (example 7) Ego-viewpoint in the present tense (example 8) Second-person narrative (examples 11 and 12) Second-person narrative (examples 13–15) Reported voices and viewpoint (example 1) Representation and compression 1 (example 11) Representation and compression 2 (example 11) Narrative spaces and roles in American Pastoral The mental space schema of a theatrical performance Acting/Speaking blend Feeling/Speaking blend Discourse blend (examples 4 and 5) The ‘crow’ blend (examples 9, 10, and 11) Third-person report (example 15) First-person report (example 18)
page 39 41 52 65 68 71 72 77 78 91 98 99 127 145 156 157 176 180 184 188
ix
Acknowledgements
A project like this one requires that the author finds friends in unexpected places, and I was lucky to meet many, and usually when I needed them most. I was writing this book feeling cautious about what various groups of people might say about it, and they invariably gave me their support and encouragement. I thus want to thank all of those who could have dampened my enthusiasm and didn’t, and who gave me good advice instead. The list is too long to quote here. I owe special thanks to colleagues who read major fragments of earlier drafts and helped with very specific responses: Uri Margolin, Eve Sweetser, Sean McAlister, Patsy Badir, Lieven Vandelanotte, Margery Fee, Suzanne Fagel, Maarten van Leeuwen, and my Cambridge University Press reviewers. Their careful consideration of my ideas and structural choices has led me through many hours of doubt. I owe much to the scholarship and inspiration of Mark Turner, Gilles Fauconnier, and Eve Sweetser. I also want to thank David Herman, who encouraged me and trusted in the final outcome before I had the courage to call this project a book, and Alan Palmer, who listened and shared thoughts when I needed it badly. I am grateful to Elena Semino and Paul Simpson for their sustained support of my various projects and their interest in my work. One of the chapters of this book could never have come into existence if not for the inspiration and friendship of Michael Booth and Amy Cook. I have also been lucky to know Lieven Vandelanotte, whose work, guidance, and support were with me all along. I have been given much support by the literature and linguistics colleagues in my own department – the Department of English at the University of British Columbia. This book was conceived against the background of their vast knowledge and wonderful open-mindedness, and I am very grateful to them all for their encouragement, friendship, and sense of humor. My graduate students have been my day-to-day audience, and have sat through many hours of workshops in which we discussed their ideas and mine. They have asked excellent questions, and have helped me focus when I was slipping. I can only mention some of them here: Jennifer Schnepf, Sean McAlister, Mike x
Acknowledgements
xi
Borkent, Slade Stolar, Tiffany Johnstone, Janet Lermitte, Robin Steen, Eve Preus, and Mark Deggan. Last, but not least, my family has helped me in more ways than I can count. My husband, Jacek, supported me with his impeccable literary taste and his boundless computer expertise – both were indispensable in the completion of this book. I also want to thank him for the motto to Chapter 8. My son, Szymek, infected me with his admiration for Melville and Poe, and also his passion for traditional Japanese theatre. Many ideas in this book owe much to his careful reading and willingness to share thoughts. This project has been supported financially with a grant from the Social Sciences and Humanities Research Council of Canada, number 410-20051644. I am grateful to SSHRC officers for making the funding available and providing prompt answers to all my questions. Last but not least, I want to thank the editorial staff at Cambridge University Press, and especially my editor, Andrew Winnard. His professional insight helped me see the big picture and guided me through various rapids. Finally, I owe much to the writers of all the brilliant stories I got a chance to read again and again. Thanks to them, writing this book was an exciting experience. It would not be fair to say that all these great novels were just excellent materials to work on, as on occasion I had to leave my analyst persona aside, and let the text speak to me. The stories just wouldn’t let me do anything but read. I am truly grateful for these moments. I have no doubt that the book has many faults, which my critics will be quick to point out. But I am sure I would have made many more mistakes if not for all the people who were willing to listen, read, and talk. I am deeply grateful to all of them.
Excerpts from American Pastoral by Philip Roth. Copyright © 1997 by Philip Roth. Used by permission of Houghton Mifflin Harcourt Publishing Company. All rights reserved. Excerpts from The Human Stain by Philip Roth. Copyright © 2000 by Philip Roth. Used by permission of Houghton Mifflin Harcourt Publishing Company. All rights reserved. Excerpts from The Blind Assassin by Margaret Atwood are reproduced with permission of Curtis Brown Group Ltd., London on behalf of Margaret Atwood. Copyright © Margaret Atwood 2000 Details of all the literary texts referred to are given in the ‘Literary works cited’ section of this book.
Introduction
The relationship between language and literature is a contentious issue. On the one hand, it may simply be described as a relationship between raw material and a finished product – language provides the basis on which creative and unique works of literature emerge. On the other hand, once we look at meaning, the dividing lines begin to fade – it is difficult to define a sharp boundary separating the meaning of literary works and the meaning of other texts. One way of downplaying the obvious links is to claim that fiction engages knowledge in a much broader and much more culturally specific way than everyday use of language does. But that would be an exaggeration. One could not follow an ordinary discussion of, say, climate change if one did not have any prior knowledge of the issue. Terms such as global warming or carbon neutral mean something only in the context of the debate and the science underlying it. It is also not true that literature is different simply because it is not intended to describe facts which have actually occurred. Linguistic constructions such as conditionals do that too – they set up imaginary scenarios and explore their consequences. The dividing line is either too fine to see, or shifts too much, depending on the assumptions. In this book I attempt not to debate those differences, but instead to talk about literary texts as linguistic artifacts, focusing on the description of language processes which participate in the emergence of literary meaning. My point of departure is the belief that if a text means something to someone, there have to be linguistic phenomena that make that possible. I found much inspiration and help in two cognitive theories of language: construction grammar, with its interest in the interaction between lexis and grammar and its trust in the meaningfulness of syntactic choices, and conceptual integration, with its focus on the mechanism of the emergence of meaning, regardless of the means that lead to the initiation of the process. Both theories have been applied to various works of fiction. However, the goal of this book is not to discuss specific interpretations of specific texts in terms of blending and construction grammar, but rather to explain how such interpretations might be arrived at. In other words, this book explores the processes of meaning construction which yield (all the varied) interpretations 1
2
Introduction
of literary texts. Crucially, I attempt to start with sentence-level phenomena and study the ways in which they contribute to text-wide meanings, relying on blending through all the levels. This project is thus an attempt to propose a blending interpretation of fictional storytelling, along with the linguistic strategies it involves. In order to talk about linguistics and literature, I had to accept the necessity of zooming in and out all the time. Some of the fragments in this book are closer to linguistics, especially when the analysis focuses on specific sentence-level choices, while other parts discuss the structure of entire novels, with limited attention to the sentence-level language they rely on. Such shifts of focus may seem somewhat distracting, but I also hope they will keep various kinds of readers engaged. This book is addressed to cognitive linguists, stylisticians, narratologists and literary critics alike, though these groups will surely find different parts of it relevant to a different degree. It also attempts to facilitate a conversation across these divisions. The structure of the book also reflects its multiple foci. Chapter 1 discusses recent work which sheds some light on the role of narratives in our understanding of language, literature, and culture; it contextualizes the ensuing discussion against a broader range of research. Chapter 2 briefly introduces blending theory and constructs some of the main tools of my methodology. Among others, the chapter discusses primary mechanisms of story construction and explains my understanding of the story as a conceptual construct. In Chapter 3, I propose the analysis of the concept of a ‘narrator’ in terms of epistemic viewpoint. Going against the tradition which assumes that every story needs a teller, I discuss ‘narratorship’ as a cluster of epistemic and linguistic choices which organize the text-wide viewpoint structure of the story. Chapter 4 investigates viewpoint configurations at lower levels of narrative structure, and argues for an understanding of narrative discourse as directed primarily at prompting conceptualizations of various kinds. In this chapter I also introduce the primary mechanism of story emergence – viewpoint compression. Then, Chapter 5 discusses the story-constructing role of referential expressions, especially pronouns and role descriptors. Chapter 6 briefly introduces narrative underpinnings of the discourse of theatre, arguing for tight conceptual connections between the text of the play, the space of the theatre, and the bodies of actors. I also show how the cognitive specificity of the theatre set-up affects linguistic choices, especially the emergence of specialized constructions. In the remainder of the chapter I attempt to trace the route which leads from on-stage narratives, where all the modalities contribute to the emergence of the story, to novelistic, or on-page, narratives, relying exclusively on linguistic forms. In this context, I also discuss the primary importance of the forms through which fictional minds are represented in narratives. In Chapter 7, I continue the discussion of the ways in which
Introduction
3
representing characters’ thoughts and emotions (not just words) lies at the core of narrative discourse. I give a brief overview of the constructions used to represent conceptualizations held by various story participants, to conclude that such textual choices are subordinate to the overall mechanisms of viewpoint compression. Finally, Chapter 8 provides some closure to the main themes of the book. While there may be only limited explicit attention given to the issue, the ensuing text also argues for a mode of analysis which treats the goals of literary, linguistic, and cognitive studies as possibly overlapping or even shared. There was much uncharted territory and a vast volume of research to contend with, but some of the central questions have emerged clearly. How does the mind negotiate the divide between everyday discourse and fiction? How does our creativity work in these two contexts? What assumptions about the nature of mind and language yield results potentially relevant to the humanities? What exactly can literary work contribute to the discussion? There are clearly more such questions than any single book can raise. This book attempts to identify and discuss the few that seemed important.
1
Language and literary narratives
There is nothing so strange it cannot be true, and no story so unlikely it cannot be told. No story is a lie, for a tale is a bridge that leads to the truth. (The Arabian Nights, retold by Neil Philip)
This book is about how the centuries-old thoughts quoted above are indeed true. Why are stories not lies, even though they don’t tell the truth? How do they help us to learn from our experience and the experience of others? And how does language support the meaning of stories? The structure of that “bridge … to the truth” is what I will try to understand. The human ability, or even desire, to tell, understand, watch, and create stories has engaged a number of disciplines, each of which poses a different set of questions about the core of the phenomenon. Why do we enjoy stories? What’s in it for us as a species? Could our culture exist without stories? Are they a mental construct, a linguistic construct, or a cultural construct? Is there a difference between real stories and fictional stories? These are just some of the questions of interest to anthropologists, psychologists, narratologists, philosophers, linguists, and literary scholars. The answers have been, by necessity, partial, and directed at the interests of the disciplines they emerge from, but it is becoming increasingly clear to all concerned that some cross-disciplinary dialogue is necessary. This book is an attempt to bring together at least some of the questions out there while focusing on one central aspect of storytelling: how do stories construct meaning? The answer to a broad question like this requires making some assumptions about the general mechanisms of meaning construction and about the correlations between formal choices and the meanings prompted. This approach would naturally define the project as a basically linguistic or stylistic one, if not for the fact that while meaning construction is a process prompted by forms (linguistic, visual, artistic, and otherwise), it is in fact a cognitive process, involving various faculties of the mind. Elucidating the emergence of meaning in various storytelling contexts thus requires adopting a methodology which is sensitive enough to fine-grained linguistic details, while opening itself up to questions about cultural context, our cognitive capacities, and 4
Language and literary narratives
5
the simple fact of there being various ways to interpret any storytelling artifact. The very term ‘meaning construction’ requires a brief explanation, so as to be clearly distinguished from literary and anthropological work reliant on ‘social constructivism.’ Throughout this book I will be using ‘meaning construction’ to refer to the processes which yield meanings of language expressions based on the frames evoked, constructional meaning patterns, and, among others, blending. Meaning construction in this sense relies on the specific word and grammar choices. Even in the simplest case such as a choice of I will do it, which commits the speaker, versus I would do it, which does not commit her at all, the construal of the former expression as a promise and of the latter as advice relies on meaning-construction processes crucially dependent on the local language choices. The construction depends on all the linguistic forms used, from the choice of the first person I (rather than third person he, which would not commit the speaker in any way), and of the modal will (rather than might, which would not involve a commitment), to the choice of tense. The present tense form will prompts a more committed meaning than the past form would. Also, while a social constructivist reading of a text is a matter of a critic uncovering frames possibly (but not necessarily) existing in the writer’s mind, the framing of past tense is a different phenomenon. Past tense forms are commonly used to mark epistemic or emotional distance, and thus are often used for politeness (so that Could you call me tomorrow? is more polite than Can you call me tomorrow?). The framing of the past tense such that it marks the speaker as less committed is a matter of standard language use, independent of any particular language user.1 The methodological framework to be used in this book relies primarily on two theories – mental spaces theory and its extension known as blending, or conceptual integration.2 Both theories assume that meaning is not a set of discrete conceptual packets neatly correlated with linguistic forms, especially lexical forms, but that in each case it emerges as a result of one’s use of formal signals (verbal or not). It also relies on processes which economically use the available complexity of possible aspects of meaning to yield an interpretation which optimally fits the needs of ongoing discourse. In actual discourse, though, interpretations are often negotiated rather than simply communicated, and thus any account of meaning needs also to consider signals built into the form being communicated which address the argumentative and inferential aspects of meaning construction. The fact that such mechanisms take the form of specialized grammatical constructions (cf. Verhagen 2005, 2008; Dancygier 2008c) should further convince us of the centrality of meaning negotiation to any meaning-emergence processes. As this book argues, understanding narratives relies on very similar processes: emergence, construction, and negotiation of meaning through specific language choices. The choices may be found at the lowest level of linguistic structure (such as using the first or the third person in
6
Language and literary narratives
the representation of narrative voice) or they can pertain to text-wide decisions on narrative structure (whether the story is presented chronologically or in a fragmented manner), but they are language choices readers rely on in the construction of narrative meaning. 1.1
Where does narrative meaning come from?
Meaning is often talked about in terms which seek to divorce it from the situation in which it emerges, and to propose formal constraints on what counts as a meaningful utterance. In the context of stories, this approach would not yield any understanding of their meaning, simply because stories do not open themselves to this kind of analysis. The same can be said about attempts, especially in structuralist narratology, to arrive at clear taxonomies and terminological distinctions. These taxonomies range over a very broad spectrum of phenomena, starting from representations of time flow to a typology of ‘narrators.’ While this kind of work has alerted linguists and stylisticians to phenomena which can now conveniently be talked about as ‘flashbacks’ or ‘intrusive narrators,’ it has not explained adequately why these forms would be used and how they are understood. Trivializing things a little, one could ask why flashbacks are so common in novels or movies (and, yes, also in natural conversations and oral stories), if most narratologists claim at the same time that the story is (roughly) a temporal sequence of linked events? And is a narratorial voice which (who?) addresses the reader directly just a frill, a naïve stage in the development of narrative form, or a charming attempt to make written fiction feel like a conversation? Admittedly, recent work in cognitive narratology (cf. Herman 2002, 2003a), Fludernik’s natural narratology (1997, 2003) and stylistics3 marks a move away from taxonomies and in the direction of cognitive explanations, but there is still much to say about the connection between the lowest-level language choices and the narrative, and, primarily, about how this connection is mediated by cognitive mechanisms involved in the emergence of meaning. It seems to be an accepted view in many disciplines (including some linguistic and cognitive theories) that language is a separate module of the mind and thus cannot be considered in the context of art, literature, or culture. The fallacy is compounded by a common belief that linguistic analyses are too formal and too focused on representing structural patterns to be useful in an enquiry into meaning in any context broader than a sentence. While this is true in many cases, it is not always true. The linguistic models to be relied on in this book, within the broadly understood discipline of cognitive linguistics, assume that meaning construction is a process which affects numerous levels, and that the overall meaning does not arise as a result of simple addition or compounding of the lower-level constructs. On the contrary, meaning-construction processes
1.1 Where does narrative meaning come from?
7
invariably involve selection and narrowing on the one hand, and emergence of new configurations on the other. There is a significant range of linguistic work using mental spaces and blending to account for meanings of nominal compounds and sentential or phrasal-level constructions,4 and such analyses typically involve uncovering the composite constructs at the lower levels. Research has also been done on the emergence of meaningful blends in discourse and conversation (cf. papers in Oakley and Hougaard 2008; Pascual 2002, 2006a, 2006b, 2008). What the analysis proposed in this book aims at is describing the consequences of lower-level linguistic choices at the higher levels of narrative discourse. While there are some limits to treating the whole text as a linguistic artifact, the focus on processes rather than forms opens longer and more complex stretches of discourse to linguistic analysis. Thus the intended contribution of the analyses proposed in this book is to offer a view of meaningconstruction processes in multidimensional artifacts such as novels or plays. The processes of selection, narrowing, and emergence of new meanings are complex and depend on more than the understanding of specific language items. As Deacon argues (2006), blending is an interpretive process which relies on our generic symbol-processing and referential capacities. That is, the ability to use linguistic symbols in new combinations, to juxtapose them or compose them into new configurations opens our capacity for coining new meanings to nearly limitless possibilities. One might read this to mean that the potential for our linguistic abilities to be used in the construction of new meanings is not inherently limited and relies on the fact that manipulation of symbolic concepts is the daily bread of our cognition. Indeed, even a very simple expression can illustrate that potential. For example, the expression emotional roller-coaster cannot be explained without an analysis of the meaning of the component concepts, especially roller-coaster. It is a fully conventionalized expression which refers to a machine found in amusement parks which takes people on a seemingly dangerous ride involving abrupt ups and downs, lack of control of one’s movements, as well as high speed. While any such experience in real life would be unwelcome, it is treated as ‘fun’ in the context of an amusement park. The addition of the adjective emotional changes the meaning significantly. There is no longer any physical object involved, there is nothing suggesting entertainment, or even an experiencing person’s conscious decision to be engaged in the activity. An emotional roller-coaster evokes negative (rather than amused) emotional response parallel to what roller-coasters do (jerking people up and down at a breakneck speed); it applies to a person’s psychological state, unpleasantly and abruptly altering between elation and depression. The composition process of this blend creates a concept not available in the component expressions, and, even more important, evokes an emotional response opposite to the original. This is also one of the points Deacon makes when he stresses the fact that emergent reference may involve conflicting emotions.
8
Language and literary narratives
The brief discussion above does not do justice to the complexity of the blend described, even though it remains at a relatively low level of complexity, which is the lexicon. And yet, the analyses throughout this book will attempt to look at narrative structure in its entirety through the lens of such low-level choices. While longer texts will be the target, sentence-level phenomena such as pronominal choice, the use of tense, the impact of referential expressions, or representation of discourse will be considered as the lowest level of meaning construction. In general, the assumption of this book (and other cognitive linguistic texts) is that grammar is as meaningful as lexis, and thus grammatical choices impact the meaning of texts. That is, a text’s reliance on interior monologue or on direct discourse is a stylistic choice which, via the interpretation of the specific constructions, impacts the way the narrative is understood. This analysis will also go a step further in arguing that a use of a specific form, associated with a meaningful construction, may ‘infect’ other aspects of the expression and carry constructional meaning while only one bit of the construction is present. The linguistic process whereby chunks of constructions spread constructional meaning has been described as the mechanism of constructional compositionality (Dancygier and Sweetser 2005; Dancygier 2005b). Examples are common. The simplest instance I can recall is an advertisement for Schweppes tonic in the London underground, many years ago. It consisted primarily of the letters Sch . . ., running on the wall along the tracks. The combination of letters, unusual in English, brought up the brand name, and its flagship drink. It seems a long way from this to a novel, but I will argue that strategic (rather than random) linguistic choices, such as the choice of firstperson or third-person narration, do propagate through narrative structure to orchestrate narrative viewpoint throughout the levels. In other words, anaphoric and cataphoric use of pronouns is a concept which explains a host of referential meanings affecting sentences, but in a longer text these concepts do not suffice (cf. Emmott 1997). Similarly, a consistent choice of past tense as the narrative form affects not only the meaning of every sentence used as ‘narration,’ but also the viewpoint construction in any sentence elsewhere in the narrative which could be described as representing speech or thought. For these aspects of meaning to be accounted for, we need to assume that grammatical choices can single-handedly construct (not just participate in) meaning configurations that they are ordinarily a part of. 1.2
Literary analysis and linguistic analysis
To be able to make its point, this book will engage with a number of storytelling artifacts. I will be relying on material mainly from novels, but also from travel narratives and plays; there will also be some short examples from poetry. Texts will be chosen to represent particular phenomena, though some questions
1.2 Literary analysis and linguistic analysis
9
regarding the historical emergence of various narrative strategies will also be addressed. Crucially, none of the literary periods will be specifically foregrounded. This evokes the question often raised in the discussions of cognitively motivated attempts to talk about literary texts. As Weber 2004 (among others) points out, cognitive methodology often downplays or disregards the historical and cultural embedding of the texts, or differences among readers. If it is indeed the case, then does it mean that such analyses are lacking something essential? Certainly not, and there are many reasons for this. First, it does not seem to be the case that everything that can be said about a historically situated text is period specific. If anything at all is indeed wrong, it is the suggestion that only period studies can reveal frames and cognitive patterns. What cognitive methodology intends to capture is patterns of thought, and not exclusively patterns of Victorian thought or modernist thought. Still, this approach does not preclude cognitively informative work looking at a specific period. In fact, quite a few Renaissance scholars (Hart and McConachie 2006; Cook 2007, 2010; Moschovakis 2006) are doing cognitively informed work themselves, and the conversation between the generic and the specific is most fruitful for both sides. Second, it is not true that ‘cognitive’ equals ‘culture neutral.’ Some of the most thorough cognitive analyses of literary texts (cf. Sweetser 2004; Turner 2004; Canning 2008) are also deeply immersed in the beliefs of their respective periods. Still, there are probably many cognitive analyses where writers ignore literary/cultural/historical expertise, either because it is not relevant to their topic or because they have different goals, and it would seem unreasonable to assume that all cognitive work on literary texts has to be supported with indepth considerations of their historical context. Also, there is the question of bridging the gap across different ways in which language is talked about in linguistics and in literary studies. Current approaches to language in the work of philosophers such as Derrida or Lacan are not easy to reconcile with routine linguistic analysis. However, cognitive approaches to meaning are less formal than most and also more open to questions of cultural or emotional framing, so it seems possible that the linguistic and literary scholars might find shared interests here. Linguistic work on literature is often seen as restricting the freedom to offer a new interpretation (cf. Jackson 2005). Yet, while meaning is not entirely indeterminate, it is also not determinate. It is a perfectly natural reaction on the part of those engaged with various sources of interpretation to feel that someone arguing for a single interpretation is missing the point. And it is true that many stylistic and cognitive analyses can be easily read to claim that what they discuss is the ‘prescribed’ interpretation (even though I am not aware of research where such claims are made explicitly). The assumption of this book, for comparison, is that while it is impossible to predict all reader responses and
10
Language and literary narratives
possible interpretations of a text, the focus on cognitively based processes of interpretation should in principle open the floor to this discussion. Indeed, what is it that counts as an interpretation? Are there perhaps interpretations out there which depart excessively from optimal avenues of text processing? Certainly. And, also, are there perhaps language-based interpretations out there which are too restrictive? Of course, there are many. The point in these cases seems to be that the interpretive tools employed are either not specific enough or too rigid, and, I argue, these limitations are the result of broader assumptions about . . . yes, about language. Whether one believes that linguistic conventions are irrelevant to meaning or that they determine meaning entirely via grammar and discrete lexical packets, the distortion is almost equally dangerous. The model of language which at least attempts to avoid these kinds of traps is a model which assumes that all meaning is constructed, regardless of the level we are looking at. The semantics which would be satisfied with the characterization of the word such as a rock as [−animate] [+concrete] [+count] will not be able to open itself to the contexts where a rock is viewed as a geological specimen, material suitable for sculpture, material suitable for construction, a weapon, a tool, or a symbol of stability and certainty, or to the possible contextual ambiguities. By the same token, if a linguistic model assumes that the present tense is simply an aspect of the morphology of English and that it is different from the imperative form, it will not even address the possible interpretation of a common usage where these forms are difficult to distinguish, as in an airport BMW ad featuring an attractive picture of a car and the line Miss your flight; still make the meeting. The slogan is obviously not issuing an imperative, and it is not describing a situation which can be construed as the use of ‘present tense.’ What is required is an understanding that a similar use of a present form appears in non-predictive clauses of predictive constructions (If you miss the flight, you can still make the meeting), and that the verb form is to be understood as neither future nor present nor imperative in meaning, because the specific constructional meaning evoked is ‘non-predicted assumption to be used in further reasoning.’ Dancygier and Sweetser (2005) propose a broader discussion of how a verb form may take its meaning from a construction it participates in and how the use of the form may prompt the meaning of a whole construction. The point I am trying to reiterate here is that grammatical and lexical choices may carry more meaning than some grammars allow. On the other hand, from the language perspective, an interpretation may appear unnecessarily open-ended. Results vary, but there are clear cases where interpreting passes over central aspects of narrative meaning because of overreliance on cultural concepts. In Margaret Atwood’s novel The Blind Assassin the actual plot, or ‘what really happened,’ is not narrated directly, but requires construction by the reader. The crucial question of the plot is whether a man accused of arson actually did commit it. He is wanted for the crime, and the
1.3 Literature, language, and human nature
11
consequences of the fire ripple through the lives of all the characters, but no answer is provided. It is in fact available in the text, but only if the reader assumes that all its narrative strands, even if apparently incongruous, are present for a reason and that the form of the novel can be interpreted in a way which reveals the whole plot. However, the indirectness of the narrative actually distracted quite a few critics (Wilson 2006; Perrakis 2006; Roy 2006) into thinking that the surface incongruence is in fact yet another instance of the ‘instability’ of meaning, and into focusing on one of the themes common in contemporary criticism – violated women’s bodies. The unfortunate consequence is that they failed to consider crucial aspects of the plot and essentially misrepresented the plight of the main character. In cases like these, the problem is not that the ‘interpretation’ simply lacks depth and convincing power, but that the interpretive avenues chosen impoverished the result, while an analysis using the complexity of the structure to reveal the complexity of meaning might have worked better. Some critical work might be read to suggest that writing challenging literature intends to background the very idea of a coherent interpretation. From the linguistic perspective, such a suggestion would be difficult to accept. Within the framework I’m assuming, there is no reason to believe even the most challenging literary texts to be beyond our interpretive strategies by their very nature, although in fact readers do give up on texts which do not engage them enough. As long as a reader is prepared to do what she does in colloquial contexts, that is, read out of a text but also read into it, the most difficult narratives can be read and coherently interpreted. With this assumption, fragmented and ostensibly incoherent stories are indeed ultimate exercises in language use, because they stretch the cognitive abilities of ‘making sense’ to their limits. It has been my experience, as well as that of many of my students, that some practical expertise in cognitive analysis opened higher levels of our abilities as readers, by making explicit the strategies which remain implicit in less challenging instances of reading. 1.3
Literature, language, and human nature
While the nature of literary and linguistic meaning is still a hotly disputed subject, more recent arguments revolve around a much broader set of concepts involving human nature, evolution, the origin of culture, et cetera. There is a growing interest in various areas of the humanities in the study of the individual mind, independent thought, genetic endowment, cognitive processes, and the bodily basis of cognition, but we are far away from an agreement on how these questions are to be incorporated into the existing body of scholarship. One proposed approach relies heavily on the findings in evolutionary psychology and related disciplines. For example, Steven Pinker’s book The Blank Slate
12
Language and literary narratives
(2002) is a passionate defense of what we used to call ‘human nature,’ our shared as well as individual cognitive and evolutionary heritage which is not subject to cultural or social molding. While the evidence he presents to talk about, for example, education is clear and convincing, the argument becomes less clear when it comes to the final discussion of the arts. Pinker criticizes the current state of the arts rather harshly, and further points out that esoteric concepts such as ‘artistic value’ have direct correlates in ‘human nature.’ To conclude his argument, he discusses some literary examples. Unfortunately, it is not likely that the way Pinker chose will convince anyone seriously interested in literature or language. Several anecdotal examples not supported with any specific argument cannot count as evidence for the above rather strong claim which the reader is expected to accept. Intuitively, his claim that literature reflects human nature is equally unsubstantiated as insisting that it only reflects human culture, and the earlier extensive discussion of the tensions between nature and nurture do not suffice in supporting this very specific claim. An off-hand statement that reading initiates a cascade of neuronal and cognitive events does not explain how this is different from the avalanche of such events prompted by an appetizing smell or a beautiful view, or how humans have come to enjoy interacting with fictional artifacts of various kinds and keep on perfecting the means of creating them. While I agree with the general gist of Pinker’s story, it is somewhat too cursory to be satisfactory. I appreciate, though, that the arts were not the main focus of his project. However, another claim implied in much of Pinker’s work is convincing – the humanities should return to an attempt to understand humans as they are. In view of serious advances in the sciences the goal seems to be somewhat more realistic, and there is more and more work pointing in that direction. For one, Slingerland’s argument (2006) for what he calls ‘vertical integration’ (of arts and sciences) presents much of the already available evidence for the viability of such endeavors. Also, the growing interest in cognition within the study of music, visual art, performance, film, et cetera suggests that there is an awareness of the need for the humanities to at least interact with the explanations science offers and to prompt new questions in science.5 There is, however, the tricky issue of how exactly science fits in the humanities and what humanities need to do to expand their current understanding of what culture contributes to literary reading. In an issue of Criticism, for example, Evans (2007) makes specific suggestions for a more refined view of culture, such that would reflect the complexity of our times and fit better into the interests of criticism (and not cultural anthropology). In the same issue, van Oort (2007) discusses the potential for interaction between humanities and hard sciences and argues that a close methodological fit, such as the one proposed by Slingerland (2006), is not desirable, because humanities need to focus on the specifically human ability to use symbols in communication and culture, even
1.3 Literature, language, and human nature
13
at the cost of downplaying more traditional critical interests. While this is an interesting argument, I am not yet sure that the nature of symbolic representation is a rich enough base to enable the ‘after-theory’ humanities to flourish. It is also interesting that van Oort expects the humanities to include the discussion of language (I couldn’t agree more), but, surprisingly, puts cognitive linguistics among hard sciences, and thus outside the humanities. Further discussion of how cognitive linguistics might be useful to the humanities will be taken up throughout this book, but van Oort’s suggested dividing lines will probably surprise researchers on both sides. The point where I agree with van Oort is when he suggests that the humanities need their own methods, whether science-based or not. However, the call for the integration of disciplines is sometimes read as an encouragement to make humanities look like sciences. One such effort is represented by a group of literary critics now known as ‘Literary Darwinists.’6 Their claim is that literature is an evolutionary phenomenon and should be studied as such. Their main argument is that literature is a direct reflection of the human evolutionary set-up, so that literary texts reflect concepts such as mate selection, dominance, and other concerns of species survival and reproductive psychology. I am reading their work as a response to Pinker’s call for putting human nature first, but the specific choice made by this group is tantamount to substituting ‘mother nature’ with ‘mother evolution.’ As such, it is a misrepresentation of literary tradition and its cultural and humanistic role, and another bold attempt to find one key concept to explain a complex issue. The swing of the pendulum from culture to evolution does not even ask how culture is linked to evolution and glides gracefully over everything in between. Also, while the literary Darwinists focus on the few relevant aspects of human nature, and do not include historically determined cultural concepts in their discussion, they are also avoiding any recognition of what schools of psychology other than evolutionary psychology have to say. This also includes theories already used by other literary scholars, such as the theory of mind, cognitive psychology, or the psychology of emotions (Gerrig 1993; Zunshine 2006; Hogan 2003). The conclusion is that this specific attempt at marrying science and the humanities is not a particularly promising avenue. My own understanding of the problem is that building such bridges requires more than one step. One cannot go from evolution into literature directly, because too much is missing: human cognition, including the cognitive roots of culture; the cultural and cognitive role of art; and the role of language in representing both. The gap is not easily filled, but some attempts have been made. Throughout this book, I will rely on some of the concepts formulated by Donald (1991, 2001, 2006) and Deacon (2006). They both refer primarily to art, but many of their claims are naturally applicable to literature (some are not, but they do not undermine the use of those which do apply).
14
Language and literary narratives
Donald’s approach assumes that there is a dual explanation to the emergence of art. One is mimesis, as the primary and evolutionarily oldest mode of representation, while the other is self-reflection. He argues that art, which relies on both, is an activity intended to influence other minds, and it functions in the context of distributed cognition. It thus has a dimension that is both individual and social, as mental representations are also necessarily shared and requiring a mode of transmission (such as the early mimetic forms involving the body or purposefully created representation, and, in the later stages, a specialized form of communication, that is, language). The self-reflective nature of art applies not only to the individual, but also to the society as a whole, and, in later stages, to human nature. Donald also stresses that art does not emerge solely from an individual mind, but that the source of creativity is in fact in the distributed network of cultural memory and its technological development. Thus a cave drawing and a cathedral are both rooted in the cultural system the artists participate in, and even though they rely on currently available technology, they serve the reflective needs of the society as well as creating specific cognitive outcomes – monitoring collective memory, giving salience to cultural values, reinforcing the social network. Narratives can naturally be viewed similarly (cf. Herman 2006). The self-reflective nature of art is further discussed by Deacon (2006), who sees the role of art in terms of the way it affects individuals emotionally. Interestingly, Deacon proposes a category of ‘emergent emotional states’ which rely specifically on the availability of more than one cognitive agent or representation. Candidates for such emotions are irony, nostalgia, righteous indignation, or humor, because they imply an interaction of various cognitive constructs. The source of nostalgia, for example, is being able to access two mental representations at the same time – the one reflecting the current state, and the other rooted in the emotionally loaded memory. To put it in cognitive linguistic terms, the tension between different construals or mental spaces creates an emotional/cognitive reaction which is different from basic emotions (see Damasio 1999, 2003) such as fear. Deacon’s emergent emotions can only appear in the network of distributed cognition, so that both the foregrounded construal and its competitor are available (as when I feel indignant but the cultural norm approves). The concept also seems to tap into Damasio’s theory (1999) postulating the emergence of human consciousness out of just such complex emotions. While Donald’s and Deacon’s work gives us a top-down view of how the cognitive architecture of our species affects human production and reception of art (including literature), there are more issues to be considered in the bottom-up view. Donald makes it clear that art works in a two-way interaction, where the artistic object offers its structure and framing and the viewer/reader adds her or his own, and Deacon also stresses the symbolic roots of art, where the
1.3 Literature, language, and human nature
15
representation emerges in the recipient’s mind. There is, however, a whole set of questions pertaining to the cognitive work done in the close interaction with narratives. For example, the work in cognitive psychology reported in Gerrig and Egidi (2003) focuses on the postulated automatic and necessary types of strategies of narrative comprehension, emerging out of experimental work. There seem to be two major possible strategies for reading narratives: either text processing is primarily memory-based, or it relies on a search for coherence which prompts the construction of causative chains. While testing memory mechanisms may seem a safer and more minimalist route, researchers working closer to actual literary narratives (Emmott 1997) are inclined to propose reading strategies which are constructivist and coherence-based. Evidence from the studies of the therapeutic role of narratives (Klein 2003) also supports this approach, as subjects find that coherent accounts of traumatic events prevent recurring uncontrollable emotional responses to the memories of trauma. Causal coherence is one of several strategies proposed in Herman’s (2003a) discussion of the role of narratives in the organization of experience. Herman follows Bruner (1991) in viewing narratives as tools for thinking. The specific strategies Herman discusses are: chunking experience, imputing causal relations among events, framing events as instances of a certain type of occurrence, and construction of event sequences. Jointly, the strategies allow one to construct events as bounded concepts appearing in a sequence and linked causally (‘he proposed, she accepted, they got married’). One might add that these strategies are supported by a number of linguistically available frames (such as, what counts as a marriage proposal), scenarios (marriage requires prior consent of both parties), and attribution of intentionality (his proposal counts as the expression of his wishes), but the fact that such conceptualizations in fact rely on each other makes it difficult to determine which of the prompts is the crucial one. It seems that such instances of ‘organization of experience’ are in fact a kind of feedback loop: the emergence of narrative structure may crucially rely on such lower-level cognitive operations, while at the same time these operations support categorizations of behavior which make narrative structure possible. Furthermore, events can be categorized as participating in a narrative chain only when they can naturally be construed as having causal force – otherwise, they do not participate in narrative structure. It is thus clear that causality plays a crucial role in the concept of narrativity, but the way in which it participates in narrative structure is more difficult to determine. In any case, these low-level cognitive processes are an important aspect of the construal of narratives. To sum up, the resulting picture of art-based cognition requires that the properties of the brain and our core skills of symbolic representation be put in the context both of a social network which gives validity and purpose to the creative acts, and of the lower-level cognitive operations such as chunking and
16
Language and literary narratives
adding causal connections. Moreover, while art relies on our mimetic imagination, it becomes more complex as humans’ technological, cultural and communicative abilities are honed in the context of a human culture understood as a cognitive distributed network. It is perhaps possible now to see how literature owes its existence to our general cognitive abilities, to our ‘artful’ mind, and to the communicative needs of a group, all of which gave rise to more and more sophisticated use of language (see Dancygier [in press a] for a broader discussion). The combination of human cognitive set-up, the cognitive roots of culture, and the cognitive roots of language is what might be an account of literature which uses science without simply repeating its claims in a context which is not reducible to the concerns of scientific research. As more and more researchers notice, being able to see the human brain at work through fMRIs and scans will not tell a researcher in the humanities much beyond the obvious observation that something is happening in the brain as we listen to Mozart or read a novel. The humanities need their own tools to represent the ultimate forms of human language and cognition, though these tools cannot ignore what science has to offer. In this book, I will attempt to clarify one end of the spectrum – the contribution of language as a cognitive tool to the formal and interpretive complexities of literary texts. 1.4
Literary texts and communication
If questions of ‘evolution’ are relevant to this enquiry, it is to the degree to which we can attempt to answer the question of how literature has emerged in its present form. I am not at all interested in speculating whether literature is an adaptation or an evolutionary by-product,7 because the way such questions are posed assumes that once it has emerged it continues playing its evolutionary role and whatever changes in literary form is no longer relevant to the question. I believe that the questions which can be considered profitably are different ones: what form of communication has fiction evolved out of? what can the emergence of different genres and poetic forms teach us about both literature and language? did language have to change to accommodate literature, and if so, how? how do literary conventions depend on our folk understanding of the human mind and human emotions? These questions cannot be answered through evolutionary theory, but they are, in my view, directly pertinent to the question which seems to be the evolutionary Holy Grail: why do humans create and read fiction? In what follows I will assume that in order to understand the emergence of literature we have to place it in the context of communication. Broadly speaking, I find the picture emerging out of Tomasello’s (1999, 2008) argument about the origins of human communication to be a good starting point, especially in its interesting elaboration of the concepts proposed by Donald and its reliance on
1.4 Literary texts and communication
17
the concept of joint attention, recently broadly discussed by Oakley (2009). Tomasello sees the emergence of narrative as relying on the human basic need to share information and to arrive at social, and also linguistic, norms in the process. Further, Tomasello claims, telling stories gives rise to specialized linguistic constructions which allow tellers and listeners to connect events across different temporal planes, to keep track of individuals, and to take another person’s perspective. Indeed, the emergence of elaborate pronominal systems and of constructions reporting thoughts and feelings while clearly keeping track of whose thoughts and feelings they are suggests some form of feedback loop between narrative and linguistic forms. Tomasello’s model can also be useful as a starting point in an attempt to track the emergence of fiction as a step in the emergence of forms of sharing information. An attempt to establish social norms through storytelling requires at some point either that the narratives used to establish those norms become more salient than what is available in the immediate environment, or that the actual true stories are told in an exaggerated form. Perfect examples of this can be found in folk myths and stories. For example, Scalise Sugiyama’s (2005; Scalise Sugiyama and Sugiyama in press) work, including her account of stories of cannibalism told in various oral traditions of foraging peoples, poses some crucial questions regarding the history of storytelling as an information-sharing device. Do the vivid and horrifying descriptions of cannibalism she discusses actually represent the past of all of those peoples? In other words, did it actually happen that children could be cooked and eaten by their parents in time of famine? Or are these stories the products of generations of inventive storytellers who made the stories more and more vivid to make their messages more memorable? Have children actually been abducted and devoured when they left the safety of the settlement, or did they need to be discouraged in exaggerated ways from leaving the safety of the group? From the point of view of Tomasello’s account the question is as interesting as it is unanswerable, since it is impossible in these cases to separate the account of events which occurred in the past from the narrative selection and presentation of events which play the normative cultural role better. Since the question about the fictionality or truthfulness of early narratives cannot be resolved, it is probably more useful to consider the ways in which their messages may have been structured. One such way relies on linguistic markers of evidentiality and hearsay (similar to English adverbs like reportedly, but often taking inflectional form, as in Central Pomo, Makah, or Pawnee). While evidential markers appear in some languages to distinguish first-hand from second-hand accounts of events, other languages develop special storytelling forms such as the ‘mythic’ or ‘ancient’ past (in Upper Chinook, or Kiksht’). While evidential markers are primarily available in spoken communication, the ‘mythic past’ has developed specifically for narrative purposes. In other words, the more salient the cultural role of the narrative, the less salient
18
Language and literary narratives
the question of its truthfulness or the specific source of information. Once the text becomes a shared cultural artifact, it participates in the system of distributed cognition and is no longer treated as one speaker’s communicative contribution. Presumably, culturally salient stories are tools of distributed cognition and as such may develop a linguistic status of their own. It is also important to note that such stories are subject to the process which blending theory refers to as ‘compression’ – the primary process leading to new blends. An example of compression could be a very simple expression, such as the generations of tomorrow, where the deictic form referring to the day following the current one (from the point of view of the speaker) is used to refer to all such days in our future. The compression of all specific ‘tomorrows’ into one downplays their temporal or deictic location, and highlights the speaker’s generic intent. In the stories under discussion here, various potentially factual events are compressed into the one the story actually tells, moving again from the specific to the generic. In the process of compression, the story may undergo various embellishments, so that the generic aspects (the ones prompting ‘narrative transport,’ as Gerrig 1993 defines it) are more salient. Thus the gruesome image quoted by Scalise Sugiyama and Sugiyama, with the mother stirring the pot in which the children’s little feet and hands are still visible, is too strikingly detailed to be an actual description of an actual event from distant past. Its salience and descriptive detail seem to serve the purposes of the generic message and effective compression, not the accuracy of the storyteller’s account. One can thus conclude that such narratives are true and not quite true at the same time, because the singularity of the event recounted is not important from the point of view of the cognitively distributed goals of the group. Looking at oral storytelling from this perspective may help us see the connection between early narrative forms and highly elaborate contemporary ones. First, oral stories are limited by the capacity of human memory, but, at the same time, they expand that capacity by relying on the storyline as the glue which holds various events, times, people, and locations together. Written narratives capitalize on these narrative constructions, but they are gradually freed from the worry about anyone’s capacity to retain the story in its entirety, and narrative form becomes more and more complex. Second, the stories such as the ones discussed by Scalise Sugiyama (also in her earlier work) are clearly intended to achieve a certain effect in the listeners – compassion with victims, fear of predators, increased sense of dependence on social and family structures. But these effects are achieved as a kind of ‘narrative uptake,’ not because they are directly communicated. The mechanism whereby an elaborate linguistic construct can prompt ‘narrative transport’ (cf. Gerrig 1993) must have been present in the minds of the foraging peoples and it is present, in an enhanced form, in our minds. Still, regardless of the form and context of the story, the ‘point’ of the story is only available through constructing its meaning – not
1.4 Literary texts and communication
19
only by tracking the sequence of events, but, more important, by exploring all the ‘what-if’ scenarios which could have prevented overly confident children from being abducted or might help us understand ourselves and our world better. There is no limit to what a narrative can do, but it requires a mind which uses the story as it is told or written to construct its meaning. Apart from this broad context in which stories achieve their status of meaningful artifacts there is also the question of how a literary text is similar to or different from any other text, spoken or written. I will assume that language and its basic communicative context are the basis on which literariness is built. I treat the basic communicative context similarly to what cognitive linguists call the Ground (Langacker 1991, 2001; Coulson and Oakley 2005; Verhagen 2005) – the immediate surroundings, the presence of the speaker and the hearer, shared knowledge, shared visual and aural field, shared understanding of bodily experience, and, last but not least, the surrounding discourse context. The Ground makes various aspects of communication possible, since meaning can be derived from ostensive aspects of communication, from objects manifest in the environment, and from the interlocutors’ body language, gesture, facial expression, tone of voice, et cetera. In my understanding, the Ground conceived in this way also provides access to shared cultural and conceptual framing, so that the interlocutors can assume access to culturally salient concepts. The difference between the Ground of any specific exchange and the default communicative context is that the latter does not have to profile any individual interlocutors, but rather a generic concept of what engaging in communication is about. Similarly, what is required is a concept of a shared visual field, and not a concrete visual field accessible in a specific exchange. Another aspect of what is required for literature to build on is the assumption that linguistic expressions prompt conceptualizations. As Verhagen (2005) argues, any exchange may involve two aspects of conceptualization: the ‘subject’ of conceptualization (the speaker, the hearer, other participants, and salient aspects of the Ground) and the ‘object’ of conceptualization (the actual subject matter of the exchange). While the potential disagreement between the speaker and the hearer as to the actual conceptualization of some subject matter is always a possibility, what Verhagen also shows is how much of the actual linguistic structure is addressed at negotiating those conceptualizations. If a child visiting a zoo calls to the parent “Look, Dad! A dragon!” this counts as an attempt to conceptualize a phenomenon manifest in the Ground and in the shared visual field, but when the father answers “No, Jimmy, it’s a lizard/a crocodile/a snake . . .” this counts as negotiating a different conceptualization, with its attendant cultural consequences (‘we cannot start pretend play whereby you are a knight and I’m a wizard and we slay the dragon, but we can learn more about reptiles’). Naturally, there is no such potential exchange between the participants in the act of reading (reader versus author/narrator/text), but the
20
Language and literary narratives
negotiation of conceptualizations is built into the narrative structure. This is, I think, what Tomasello talks about when treating narrative as an engine to produce more complex linguistic constructions. These structures not only maintain temporal, referential or viewpoint links, but allow the construction of narrative meaning to happen. The greater part of this book is devoted to revealing the various ways in which narratives perform this task. It is important to be specific about how narrative context deviates from the default communicative context. First, ordinary communication features two interlocutors, the speaker (I) and the hearer (you), who, in the most standard set-up, have visual and aural access to each other and exchange their deictic roles in every turn of the conversation. In this default set-up what counts as communicated has to be said or done (or, from the point of view of the receiver, heard or seen). Naturally, this would never suffice, and that is how languages develop various forms of establishing categorizations, encouraging inferences, reaching beyond the available Ground, reporting past situations, et cetera. But the default, as I will try to show in chapters 6 and 7, remains surprisingly persistent in structuring different literary forms. Deviations from the default are common, and in most cases discourse is maintained via appropriate constructions, such as back-channeling in phone conversations (all the yeah and hmm expressions signaling continued presence), emoticons in e-mail, page layout and paragraphing in journalistic discourse, intonation patterns, rhetorical devices and repetitions in public speaking, et cetera. More interesting, however, there are instances of discourse which seem to assume the default in obvious absence of its crucial ingredients. People routinely speak to animals and plants, and also to cars, computers, and other gadgets (mostly phrases like Don’t do this to me!), while animals and objects commonly speak in stories, comics, and poems. Not only does this deviation not cause incomprehension, but most of the time it also goes unnoticed. It seems that the very fact of something being said evokes the default set-up, and thus ostensible communication is treated (linguistically) as bona fide communication. Such a ‘constructed’ communicative set-up is just an example of the fact that something being communicated suggests intentional communicative agents. Dennett (1997) talks about our ‘intentional stance,’ whereby we attribute agency to objects and animals, but it seems to be only a part of what Deacon calls a ‘representational stance,’ whereby we tend to treat things occurring around us as meaningful. In the context of a written narrative, both stances are involved. The presence of a text is treated by default as interpretable via our symbol-reading abilities, and as presented to us in its form by an intentional agent. Reading is thus a communicative act which has an addresser and an addressee, even though the communicative intent is available only through the form of the text in front of us. It is interesting to note how natural it is for readers
1.5 Why is fiction special?
21
to talk about ‘books’ or other texts as ‘saying’ something, and attributing intentions to authors (‘the author is saying X,’ ‘the author meant X,’ et cetera). These ‘fallacies’ seem to rely on the default understanding of communication. This seems also to be the case with extended narratives. There used to be a teller of an oral story, and the listeners may have assumed that the communication mostly goes one way, but these assumptions did not affect the communicative impact. Contemporary fiction is mostly written, and so the ‘communicator’ is hidden behind the text, but the concept of a narrator has substituted that of a teller. As I will argue throughout this book, the narrator (but not the author) may be no more than a linguistic construction (or constructions, since there are many kinds of narrators), and yet a written work of fiction is naturally seen as conforming to the communicative default. Also, earlier narrative forms are more preoccupied with mimicking the ‘somebody told me’ or ‘I saw it’ pretense, while less narrator-mediated forms emerge much later. Furthermore, visual prompts participate in storytelling alongside words, and often in ways which not only support words, but sometimes give them different meanings. 1.5
Why is fiction special?
Or is it? There are several recent attempts to address the question. Zunshine (2006) sees reading fiction as practice for our theory of mind, assuming, with Palmer (2004), that the access to characters’ minds is more important than the story itself. Oatley and his co-workers,8 for comparison, argue for some kind of cognitive gain, especially if the material read is sophisticated and has some psychological depth. Similar claims, though based in an evolutionary view, have been made in Boyd (2009), who argues for the development of fiction as driven by cognitive play. Researchers in the cognition of art look at fiction as a form of art, and try to view it alongside the path leading from cave drawings to Van Gogh. Others seek the value of fiction in entertainment, in crossing the boundaries of time and space, in exercising compassion, et cetera. However, there are many questions which are not addressed through these enquiries. First, why do people write fiction? The view from neuroscience, lucidly presented by Flaherty (2004), suggests that creative urges to write are to a degree a part of our biology. From the cultural perspective, there has never been a problem with low supply of aspiring writers, though we do not seem to question the fact that only some writers make a difference to our view of the world and have pushed the boundaries of narrative and poetic skills to new heights. There is plenty of fiction which seems no more than chewing gum for the eyes, and yet finds plenty of committed readers, and definitely counts as literature. There is also the new phenomenon of writing as a job. When Haruki Murakami, a very prolific and popular author, talks about his career, he simply
22
Language and literary narratives
says that he thought at some point in his life that he could make a living by being a writer (Murakami and Gabriel 2008). My point is that while we tend to focus on the writers who make a difference and build on that, a lot of storytelling seems to have much less sophisticated forms or less high-flown goals. It is a form of linguistic expression which builds on culturally supported cognitive needs. Furthermore, we ‘talk fiction’ much more than we are willing to admit, and our languages have elaborate structures allowing speakers to explore different scenarios. Conditionals and other hypothetical constructions in languages are varied and complex, no less complex than a fictional story, if not more, simply because of the compactness of actual constructions. There are many ways for speakers to mark evidentiality, epistemic stance, or emotional stance, and most of these linguistic forms are either lexicalized, or grammaticalized, or constructionally determined. Linguistic forms are typically understood to mean more than the facts described (consider implicature, linguistic politeness, indirect speech acts, and the like), and open themselves to reasoning more often than just stating the facts (hence so many attempts to treat language as based on some kind of logic, however misguided some of these attempts may be). My point is that the idea that colloquial language use is about facts while fiction is not is simply wrong. Language is very well equipped to deal with ‘what ifs’ and hidden meanings and is usually very clear in representing the speaker’s attitude, knowledge and the source of information. Contrary to common belief, statements that logicians love to work with, such as Grass is green or Mary is in London, are much less common than ones starting with if, I think, he said, or I wish. The question is, then, why treat fiction as . . ., well, pure fiction? In his delightful personal account of narrative techniques, Vargas Llosa (1997/2003) talks about the author as resembling “the Catoblepas” – a mythical creature which consumes itself, starting from the feet. As he argues, the themes an author will explore have to emerge from his own experience, though he can then give them any narrative form he chooses. There is thus a grain of truth in any fictional narrative, however weird it may be. Setting the case of realism in fiction aside, genres departing from ‘realism’ also are real to a degree. Even a cursory look at the ‘ultimately’ fictional genre of science fiction shows that while trying to construct storyworlds as remote from planet Earth as they can, authors cannot but rely on our reality in many of its aspects. At the other end of the spectrum, texts which are treated as novels but have been described (by their authors as well) as primarily autobiographical are versions of actual happenings. A perfect example of this phenomenon is J. G. Ballard’s acclaimed novel Empire of the Sun, which is a blend of fact and fiction. As a child, Ballard did spend three years in an internment camp in China, and the novel is based on his story. In his autobiography, The Miracles of
1.6 Narrative and grounding
23
Life, Ballard claims that most of the events described in the book happened in reality, though not to the people actually portrayed in the novel. He also explains the reasons for his choices, for example, for placing his character, Jim, in the camp without his parents, while in fact they were there with him. Ballard wanted to focus on a child’s experience without the protective layer which could only blur the message – a choice as much related to compression and narrative salience as the one in the cannibalism story. Fiction thus cannot be divorced from facts, some as basic as the existence of oceans, dogs, and bacteria, and some as specific as the conditions in internment camps in China during World War II. Furthermore, the need to categorize any given story as either true or not true seems to be important for cultural rather than linguistic reasons. It is important to know whether, let’s say, a politician accepted a bribe or not, because it is a legal issue. But whether a man named Darcy actually married a woman named Elizabeth Bennett, or whether they never met, or never even existed, is not relevant to the story. Pride and Prejudice is not a true account of what happened (as far as we know), but it can be read as such and may not ‘feel’ different in terms of factual plausibility from a story about distant cousins one can hear from one’s grandmother. Crucially, the latter story would have a different form: it would be shorter and use colloquial linguistic constructions. It thus matters little whether the story presents facts, but it is crucially important how it is told, and here is where the issue of literary discourse becomes interesting. To conclude, fiction has to be based on reality to have its effect. Reading biographies is often as moving and interesting as reading fiction, because the textual matter and the attendant constructed meaning or emotional response is triggered through the same medium and the same conceptual mechanisms. Uncovering the linguistic specificity of the medium and the conceptual mechanisms leading to narrative uptake is the goal of this book. 1.6
Narrative and grounding
One of the themes in the discussion of narratives which this book engages with is the cognitive status of narrativity. While definitions of narratives vary, they consistently stress temporality, causality, representation of minds, relationship to spatially and temporally defined storyworlds, problem solving, and the construction/resolution of some conflict (see Ryan 2007 for an interesting discussion). From the cognitive perspective, Herman (2003a) argues further that narratives are ‘tools for thinking,’ which help in the organization of experience, through the establishment of chunks of experience and causal links among them, recognition of types of events, et cetera. These lists include some more general concepts such as causality, and some which are narrativespecific, such as the ‘storyworld’ or ‘conflict.’ And yet there are also attempts
24
Language and literary narratives
to reach to the very conceptual core of the narrative and establish the role of narrative structures in thought processes in general. One such attempt is the book by Turner (1996), which sees basic narrative organization as a primitive concept and the source of the emergence of both linguistic and literary forms of expression. As Turner argues, perception of basic ‘spatial stories’ such as opening the door or throwing a rock is the starting point of higher levels of conceptualization. In this section, my goal is not to arrive at some definition of narrative which would reconcile all these descriptions, but to consider some problems in relating narrativity to conceptualization. Herman’s approach (2002, 2003a) identifies aspects of narrative which participate in conceptualization of events as linked into a story. Generally, Herman sees the story as a composite concept which provides a link between raw experience and its ‘storylike’ conceptualization. Turner, for comparison, views ‘storiness’ as a basic perceptual tool, leading to conceptualization of all kinds of experience as well as to projections of one ‘story’ onto another, leading to metaphor, parable, and more complex narrative forms. Essentially, Turner’s concept of a primitive narrative is a configuration of image-schematic concepts rooted in sensorimotor experience in which (causal) force or sequence plays a central role. Turner’s approach to narrative, and, by the same token, the entire theory of conceptual metaphor and blending, was extensively criticized in van Oort (2003), primarily on the grounds of what he perceives as the assumed direct link between perception and conceptualization. Since the issue, which I will refer to as the ‘grounding problem’ (after Harnad 1990), is clearly at the core of much of the current discussion of the narrative, I will attempt to review the questions briefly. In much of the work in cognitive linguistics it is assumed that meaning is grounded in perception and sensorimotor experience.9 Basic spatial configurations called image schemas (such as containment, force, balance, boundedness, et cetera) lead to the emergence of linguistic expressions describing concrete aspects of experience, but can then be projected into abstract domains of human activity. Thus most concrete categories as well as many abstract meanings are grounded in human perception and experience, via levels of projections which also distinguish multiple meanings of lexical items. For example, the preposition in can be used to describe concrete cases of containment, but also abstract states such as being in despair. Van Oort criticizes this explanation by arguing that it assumes a direct connection between perception and linguistic categorization, while what is needed is an explicit inclusion of a separate intervening level of symbolic representation. Van Oort refers extensively to Deacon (1997), who sees symbolic representation as the core of human cognition. He also evokes the claim that words, as symbols, are primarily in relationship to other words, and not to the speaker’s perceptions or actions. Finally, van Oort reads Turner’s discussion of the basic role of
1.6 Narrative and grounding
25
narrative as a hidden claim of the ‘story’ being an innate concept, very much like Chomsky’s universal grammar. It might be helpful to put this discussion in the context of recent experimental work in cognitive psychology. Surely, the idea of representation and categorization is at the centre of the discipline’s interest, and some of the research done in the last few years explicitly addresses the grounding problem – what does it mean for perceptions to drive conceptualization? how are categories formed in the brain? how do abstract concepts emerge? First of all, research starting as early as the 1990s (Barsalou 1999; Harnad 1990) suggests that treating meaning in terms of links within a network of abstract symbols in the mind runs into problems. One, the approach does not explain the observed emergence of arbitrary symbols from experience and perception, and, two, it offers no solution to the commonly asked question about the ways in which symbols connect to reality, since without some connection to the real world, symbols cannot perform any cognitive function. Experiments which ensue suggest that perception and action are central to higher cognition, without postulating intervening mental representations.10 Moreover, one of the claims is that experience can be used in the conceptualization of situations in other times and places, and the idea led to further investigation of how basic perceptions and actions can participate in conceptualizations without necessarily going through the configurations of arbitrary symbols. There seem to be many reasons not to treat discrete symbolic representation of categories as the basis on which meaning emerges. Recent work by Prinz (2002, 2004, 2005) makes a very strong case for the experientially grounded emergence of abstract concepts, including emotions. First, Prinz shows how even complex concepts such as ownership can be naturally explained through configurations of basic perceptual and sensorimotor concepts. Second, he sees concepts as necessary for action, and not for static categorizing, so that conceptual representations are variable and context-sensitive rather than participating in a stable system of symbols. Furthermore, experiments reported in Spivey, Richardson, and Gonzalez-Marquez (2005) show the smooth and pervasive interface between on-line language use and sensorimotor processes. Broadly speaking, research reported, among others, in Spivey (2007) and Pecher and Zwaan (2005a, 2005b) supports some early suggestions made in Gibbs and Matlock (1999) that tracing complete conceptualizations in the brain may not be possible because of the dynamic and continuous nature of such processes. At the same time, the research gives experimental support to two major claims of cognitive linguistic theories – one, that meanings emerge out of the basic levels of embodiment, and, two, that meaning is constructed on-line, for the needs of the specific act in which the conceptualizer is immersed, which does not preclude conceptualizations of situations not directly available in embodied experience. In other words, meaning is embodied and dynamic, even when what
26
Language and literary narratives
is being conceptualized is abstract, decoupled from the circumstances of the act of conceptualization, or leading primarily to an emotional response. Crucially, theories of embodied meaning do not require that arguing for a connection between experience and conceptualization puts one automatically on the side of extreme empiricism. While the authors who argue for embodied roots of abstract thought or emotions from a philosophical (rather than psychological) perspective (Johnson 1987, 2007; Lakoff and Johnson 1980, 1999; Damasio 1999, 2003; and Prinz 2002, 2004) agree that mind/body dualism is a convenient fiction, they are centrally interested in the emergence of conceptualizations, rather than in defending some form of empiricism. Even when experimental work confirms the role of image schemas in higher cognition, what is properly being described is the role of basic spatial configurations emerging from experience in concept formation. In other words, the upshot of the research is explaining the reason why an abstract word like status evokes the upper end of a vertical scale and a spatial expression like up top does too. Here is where van Oort appears to misinterpret the cognitive paradigm, by assuming that what is being said is that a use of a spatial preposition to describe a spatial configuration relies directly on perception without symbolic conceptualization, while its abstract meaning does not, whereas what is argued is that all kinds of meanings emerge on the basis of perception concepts. It is thus indeed possible to argue that basic spatially and temporally situated actions can serve as sources in conceptualization of abstract or displaced concepts, such as the ones underlying basic narratives – which has been Turner’s claim, formulated at a level closer to language than to cognition. The claim does not need to involve any hidden or explicit assumptions of innateness, because conceptualizations which emerge are not static or determinate enough to be candidates for such status. In general, claims of an innate mechanism related to language are not corroborated by experimental research. Second, while the symbolic nature of language is beyond dispute, the participation of symbols in various relationships with other symbols does not have to assume either discreteness or systematic correlations (such as binary oppositions, semantic features, et cetera). Also, our model of meaning has to capture the processes whereby new meanings emerge out of contextually prompted combinations of old meanings, and frameworks such as construction grammar and blending have made significant progress in that direction. Perhaps most important, though, treating symbols only in opposition to indexes and icons misses the point, for it creates an impression that only indexes and icons are linked directly to experience. While they have their own roles to play, symbolic concepts, as has been amply demonstrated, rely even more specifically on embodied experience. In other words, symbols are not detached from the real world; on the contrary, they are immersed in it via their reliance on image schemas and action. Categories and symbolic concepts are thus related to other
1.6 Narrative and grounding
27
categories and concepts not on the basis of the system of related meanings, but on the basis of underlying shared experiential concepts. To give just one example, if Spivey’s eye-tracking experiments reveal minute eye-movement upwards as a shared and consistent response to words such as up, tall, and respect (in clear contrast to down or submit), the connection thus revealed is among various applications of the basic image schema of verticality, from movement or location, through scalar measurement specific to cultural norms (a tall pygmy is a short Masai), to a social relationship. As an additional bonus, the proposed experiential treatment of meaning does not exclude cultural differences from the actual conceptualizations. Returning now to the concept of ‘narrative,’ we might say that in the context of the discussion above, it is not substantially different from any other form of conceptualization in that it relies on the dynamic emergence of representations prompted by the text (in terms of both the events portrayed and the emotional response to those events), and that it relies on embodied cognition. The specificity of how narrative chains of events are held together has much to do with the processes identified by Herman (2003a), but these processes are also essentially supported by experiential concepts. If I think of a portion of a story as ‘giving offence,’ I have to rely on the force-dynamic interaction between the participants, the emotional impact of their words and actions, the need to restore balance, et cetera. The fact that the events to be conceptualized are fictitious is not an obstacle to conceptualization. Naturally, narratives make heavy demands on cultural knowledge, memory, empathy, and other aspects of cognition, but they are not encumbered by the need to refer to discrete symbolic representations as independent steps in the process. Turner’s claim about the centrality of narrative thought and projection to conceptualization of any kind should not be read to suggest some evolutionary process which led to the emergence of the narrative module in the mind. In fact, claims about the evolutionary history of specific human abilities are becoming common. Since Fodor’s early work (1983), followed by the Chomskyan paradigm in linguistics and Pinker’s (1994) ‘language instinct,’ the search for modules in the brain to explain essentially human abilities seems to continue, even if jocularly. For example, a recent book by Dutton (2009), The Art Instinct, adds to the ‘instinct’ metaphor by considering art as a specifically human trait, doubtless originating in some evolutionary event or other. While any discussion of whether such innate modules in general are indeed a defensible explanation of the central aspects of human cognition is beyond the scope of this book, there is clearly no need to postulate a narrative module. Is there, then, any connection between basic narratives and evolution? While some recent work (Carroll 1995, 2005; Gottschall and Wilson 2005) connects evolution and literature, as well as evolution and art (Boyd 2005), these researchers make no claims about narrativity as such, and yet ‘narrative
28
Language and literary narratives
thought’ seems to pervade conceptualization. And as van Oort suggests, Turner’s idea of the ‘literary mind’ might be read as saying something similar. But while the pervasiveness of narrative is a fact, I hope to have shown that there is no need to postulate a narrative module or ‘instinct.’ Narrative-based conceptualizations may be indispensable (so I tentatively agree with Turner here), but they are readily explicable in terms of lower levels of embodied cognition. Also, while language is naturally treated as the basic environment in which stories exist, it is not indispensable to narrativity, since a story can be ‘told’ through visual means. The concepts required for a narrative to emerge (sequentiality, causation, chunks of experience, cultural framing of such chunks, imageschematic force-gestalts of conflict and restored balance, et cetera) are the same concepts which are required for other language conceptualizations to emerge, so we should perhaps advocate a stronger claim, such that narrative form relies on a specific cluster of such concepts. Moreover, recent work in both cognitive psychology and evolutionary theory allows one to talk in a sufficiently rigorous manner about human abilities which are not only cognitively central, but also culture-based. The work by Donald and Deacon mentioned above speaks of ‘coevolution,’ and the idea is given a more specific meaning in the theory of gene-culture co-evolution.11 The focus on cognitive sources of culture can at least suggest ways in which a centrally cultural non-primitive concept such as the ‘narrative’ may have become crucial to higher cognition, while remaining reliant on more primitive image schemas of force, sequence, or balance, on specialized language forms (extended structures which connect sentences in a variety of complex ways), and on the culture’s need to represent events and cultural norms in forms which guarantee their temporally extended availability. This discussion has not yet addressed a related but largely independent question concerning the cultural and linguistic emergence of narrative fiction. Zunshine’s work (2006) locates the explanation in evolutionary psychology and more specifically in the so-called ‘theory of mind’.12 The term is best known in the context of Baron-Cohen’s (1995) generally accepted explanation of communicative deficiencies summarized in the term ‘Autism Spectrum Disorder.’ At least some of the difficulties of autistic speakers are attributed to their lack of an understanding of how language and behavior can be read in order to grasp people’s mental and emotional states and adjust one’s language and behavior to meet their expectations. Zunshine builds on this ability to ‘read’ other minds to argue that fiction offers a form of ‘exercise’ in using it. There are some problems in such a use of the concept. First of all, ‘mindreading’ includes (and is tested with the use of) understanding of communicative components such as eye contact, facial expression, or body posture, and not exclusively language. It is also being used in explaining a variety of behaviors, including appreciation of humor, irony, and figurative language.
1.7 Approaching narratives
29
Reading cannot test all these skills, or serve as practice, because it relies on a narrative rendering, not lived experience, and is mediated entirely through language. If there is any disposition commonly associated with good mindreading which can also be applied to fiction, it is the ability to grasp the emotional impact of people’s words and actions. Interestingly, appreciation of mental and emotional acts has already been claimed to be the single most salient feature distinguishing narrative fiction from other narratives. Palmer (2003) argues that ‘fictional minds’ are the core of narrative discourse, even if narration presents mostly actions – this is what fiction is about, and the plot is in fact secondary to this primary focus; a similar point has also been made in the philosophical context by Currie (2007). I agree with Palmer because his observations elucidate the role of much of what needs to be narrated in a novel, but also because they explain how fiction stimulates an emotional response in the reader without relying on the whole constellation of abilities roughly summarized under the rubric of ‘theory of mind.’ Palmer’s claims can also be supported with recent research on the role of embodiment in representing mental states (cf. Zlatev et al. 2008), and I will return to the question in Chapter 4. However, accepting Palmer’s argument does not require a sharp distinction between, let’s say, narratives in oral traditions and narrative fiction as we know it. Any narrative form assumes the readers’ or listeners’ ability to feel for the characters and interpret the events from the characters’ point of view – to feel sympathy, envy, disgust, anger, et cetera. While some of these emotional responses may be complex, or even conflicted, they are better explained through our ability to feel empathy, and also to attribute emotional meaning to linguistic expressions – even simple experiential concepts such as up (alert, conscious, confident, in good mood) or down (tired, inactive, sick, unconscious, dejected). To conclude, I will treat extended fictional narratives – the primary focus of the project – as relying on our abilities to construct meaning and respond to it emotionally. But I will also assume that these abilities emerge, more or less in the ways described by Tomasello (2008), from the simplest linguistic forms based in embodied cognition, through more and more complex narrative forms, all the way to the complexities of contemporary stories. I will also argue, mainly in Chapter 6, that narrative ways of representing mental and emotional states of characters have evolved through literary periods and genres towards a more and more direct, and thus narratively more complex, access to fictional minds. 1.7
Approaching narratives
As the discussion above has indicated, narratives and narrative fiction have been subject to enquiry in a number of different disciplines, from psychology and cognitive science, through philosophy, literary criticism, stylistics and discourse analysis, to linguistics and narratology. Each of the disciplines contributes
30
Language and literary narratives
much to the discussion, while also having its own goals and methods. Where does this project locate itself then? The simplest answer would be – towards the linguistic and cognitive end of the spectrum. I want to propose an analysis which is informed by current research in cognitive science and which clearly addresses the linguistic phenomena specific to extended narrative texts. I will refer repeatedly to the concepts proposed by narratologists, but I also want to put this discussion in a context which is more open and more restricted at the same time, as it seeks to explain processes of conceptualization by looking at the fine-grained linguistic prompts for these conceptualizations. At the same time, the tools proposed should be useful in a discussion within stylistics, where language is discussed in terms of specific textual choices and their impact on the special quality of the individual text or genre. Like some previous work in cognitive discourse analysis (Emmott 1997), this project will also attempt to look at features of textual matter which contribute to text-wide aspects of interpretation. The hardest relationship to define is between this work and literary criticism. While I hope to open the linguistic tools to the role of cultural values or historical facts, I cannot make them central to the discussion, and thus much of the interests of literary critics will be downplayed. However, my interest in paths of interpretation is what critics should recognize as relevant to their field. There is also the question of pedagogy. In my work with students of literature, I found that introducing them to the concept of a frame or an image schema often gave rise to new and innovative interpretations, not because the concept could be found in one text or another (it would probably be found in any), but because of the special salience of some basic cognitive configuration in a text. In other words, they noticed things that required interpreting because they were highlighted once a cognitive concept became available. So cognitive analysis did not seem to restrict the playing field, it in fact opened it up. This book will thus focus on the link between language and cognition and the ways in which such correlations affect the construction of narrative meaning. First, I will outline the ways in which tools developed in the theory of conceptual integration can be applied to longer texts. I will start the discussion of narratives with situating the text-wide construal of narratorship in the conceptual integration model. The central part of the book will focus on discussion of viewpoint, looking at some ways in which sentence-level lexical choices and referential expressions affect narrative viewpoint. I will then move on to the role of representation of fictional minds in various storytelling genres, including the discussion of various types of discourse in the narrative, with special focus on the role of direct speech and thought. Finally, I will resume more general discussion in the concluding chapter.
2
Blending, narrative spaces, and the emergent story
When he woke up, the dinosaur was still there. (The Dinosaur, Augusto Monterroso)
Stories can, naturally, be treated as linguistic constructions. They are attested examples of natural written discourse, and, while it is possible to treat linguistic choices in the narrative as primarily rhetorical, there is also crucial reliance on very basic facts of language structure. Stories should also be talked about in terms of meaning. They are particularly challenging in this respect, if only because of the sheer length of the text and the variety of narrative options available. This project proposes a framework which makes the discussion of the meaning of stories possible. It relies heavily on analytic tools established within the discipline of cognitive linguistics (mental spaces theory, construction grammar, and frame theory), but it is primarily focused on the theory of blending, or conceptual integration. Fictional narratives rely on a variety of linguistic phenomena, but adapt them to their needs. To give just one example, they use personal pronouns and deixis in ways which exploit the default spoken setting, but set it against a different kind of deictic ground. Thus a pronoun like I is often used in novels to represent the narrator – a narrative construct, rather than an actual communicator. The discourse setting is that of a communicative act occurring across a spatial and temporal distance, with no clear addressee available (other than some generic reader). The consequences of such a set-up are farreaching, as they create the need for a different treatment of referential expressions designating characters and put the concepts such as the ‘present’ or the ‘past’ in a different context. These peculiarities will be discussed throughout this book. With respect to meaning complexities of fictional narratives, blending (cf. Fauconnier and Turner 1998a, 1998b, 2002) seems to be particularly well-suited to proposing an account of narrative structure. The concepts and mechanisms proposed in blending work translate naturally into the needs of an account of narratives, but also open the door to the discussion of blending mechanisms and properties which seem specific to the emergence of meaning 31
32
Blending, narrative spaces, and the emergent story
of longer texts. In what follows, I will elaborate the blending framework in ways which best respond to the needs of accounting for narrative discourse.
2.1
Applying blending to fictional narratives
The essential claim of blending is that conceptual packets (mental spaces or frames, here called inputs) can be activated in a person’s mind by verbal as well as nonverbal prompts and integrated into a new conceptual configuration (called a blend) to construct new meanings as discourse develops. The conceptual structure of the blend, while relying on projection from the inputs, achieves its own coherence by selecting only the relevant parts of the inputs and by compressing vital relations, such as time, causation, or identity. For example, while the inputs are two mental spaces one of which is present and the other past, the temporal distance between them may be compressed to yield a blend which is either past or present. The inputs remain activated and retain their separate conceptual status, but a specific blend may also project structure back to the inputs (backward projection), and thus reconstrue them for the purposes of the ongoing exchange. The blend is characterized by its own structure (emergent structure), but can then become an input to another blend, or series of blends. While blending can account for ongoing processes of meaning construction (as in following the course of a conversation or reading a comic strip and accumulating content until the joke gels in the final frame), they can also become entrenched as new expressions, such as compounds or idioms. Blending also accounts naturally for multimodal contexts, where visual, linguistic, and aural prompts work in combination, as in film. The way the blend emerges is largely dependent on the topology (internal structure) of the inputs. While in a number of cases inputs are mental spaces, they may also be structured by frames (as introduced in Fillmore 1985, 2006; Lehrer and Kittay 1992). For example, a blended construction such as When I learn to surf, I will move to California assumes the availability of two mental spaces: the future space in which I am a surfer and I move to California, and the present reality space where it is not the case. Each of these constructs has the structure of a mental space, with participants, temporal and spatial features, and internal structure (such as a connection between California coast and surfing opportunities), and they yield a blended space wherein being a surfer in California also retains other aspects of my identity: family ties, life story, et cetera. Crucially, the imagined future situation remains a mental space precisely because it has been constructed on-line, that is, within ongoing discourse, to indulge in an ‘alternative life’ for the duration of the conversation about surfing, and it will be deactivated when the discussion of perfecting surfing skills stops or when the topic is changed.
2.1 Applying blending to fictional narratives
33
Frames are significantly more stable, in that they structure our understanding of reality beyond the conversational context and are often associated with specific lexical items. While we all have the general concept of dying, we apply very rich cultural framing to concepts such as killing, murder, infant mortality, accidental death, or genocide. Each of these includes the fact of a person or persons dying, but they also include complex scenarios specifying the circumstances, causes, the degree of agency, cultural and social realities, et cetera. We can thus say that they participate in different frames, and that these frames are associated with assignment of guilt, legal consequences of guilt, moral values, medical problems, et cetera. Using any of the expressions listed above evokes all the associated framing, and, by the same token, appropriate social actions. The most important aspect of framing is the possibility of accessing the entire frame when only one aspect of it is mentioned. In the most obvious case, the discussion of buying a new car automatically implies the aspects of the frame which are not explicitly mentioned: selling, transfer, object transferred, price of transfer, et cetera. In the less obvious cases, when an aspect of the frame is chosen in a less predictable manner, we can also talk about frame metonymy – linguistic usage whereby an expression is used to stand for an associated aspect of the frame, while also evoking the entire frame. Cases of such pragmatically established metonymies were discussed, among others, in Fauconnier (1994 [1985], 1997). Fauconnier analyzes examples such as Room 4 has a visitor, which can be said by a member of hospital staff to refer to the patient in room 4 and announce that a visitor has arrived. In Dancygier and Sweetser (2005), the concept of frame metonymy has been expanded to include constructional frames, so that, for example, a specific use of the present tense to mark an unpredicted premise in a predictive reasoning signals predictive meaning even if no other constructional features are marked. Thus, a sentence such as You miss one more meeting, you’ll be fired counts as a prediction of dire consequences if an as-yet-unpredicted event of being late occurs, without using a conditional predictive construction in its entirety – the verb forms are enough for the frame to be evoked. In recent work (among others, Dancygier 2009), I argued that frame metonymy at the lexical level, when paired with frame metonymy at the constructional level (termed constructional compositionality), creates blended constructional patterns in which meaning emerges based on the specific blend between the evoked lexical frame and the constructional frame. For example, a sentence such as Iraq is a new Vietnam seems to follow the format of a copular construction in which the subject is a specified entity and the predicate is a non-specific name of a category (as in John is a teacher), but both proper names are used to evoke metonymically frames of US interventions and extended military conflicts (rather than ‘countries’), and thus the adjective new signals the
34
Blending, narrative spaces, and the emergent story
transfer of some of the framing of Vietnam (high cost, lack of success) into the framing of Iraq. As a result, Vietnam does not become a common noun (even though the frame is detached from its standard referential meaning and an indefinite article is used), while Iraq counts as metonymic for the Iraq war, rather than the country. The interaction among the frames evoked by both proper names, the adjective new, and the construction yields a blended meaning not available outside of the construction. I will discuss framing and reference in more detail in Chapter 5, but the connection between frames and blends has thus been briefly illustrated, and will be relied on in further discussion. A simple example of a blend is a compound expression such as computer literacy. The two inputs are frames of being (one) a computer user and (two) a person able to read and write. The frames are rich culturally. While literacy used to be considered the cornerstone of participation in contemporary society, computer use until recently was not considered an indispensable skill. The compound expression blends the two concepts to refer to a set of computer skills which are now considered essential. They are not restricted to reading and writing, but involve also an understanding of basic computer functions. While the ‘computer’ input contributes the understanding of which skills are basic, the ‘literacy’ input further qualifies those skills as required of anyone who wants to be considered minimally educated. The blend also informs both inputs, by defining computer skills as basic, and expanding the scope of literacy to match contemporary needs. Interestingly, the expression is then cleverly reframed in the name of a well-known Silicon Valley institution – Computer Literacy Bookshop, where the expression computer literacy refers to interaction with written language again, even though only computer professionals are the intended readers. It is thus a good example of how subsequent blends manipulate both the inputs and the earlier emergent meanings. A more complex blend is prompted by a clausal construction such as If I were your mother, I would teach you how to cook. It sets up an imagined situation which is a blend of the if-space (the speaker is the hearer’s mother) and the reality in which the speaker and the hearer participate (the hearer cannot cook, her mother never taught her, et cetera.). The emergent structure also creates a new imagined identity blend, wherein (under one of the possible interpretations) the hearer’s mother shares the commitment of the speaker to teach the hearer to cook. The blend is complex, but the mechanism that allows it to emerge is the same simple mechanism, and, crucially to the subject of this book, operates on inputs which are rich in contextual information, feature various participants, and present events and human dispositions as causing other events and dispositions. We can assume that to the speaker and hearer of the conditional all the contextual implications are known – the desires and personalities of the participants, the reasons why knowing how to cook is useful, and the reasons why the speaker cares about the hearer’s skills.
2.2 Narrative spaces as mental spaces
35
Similar meaning construction processes occur in longer narratives. Frames and mental spaces structure inputs, which then become integrated, possibly in ways specific to a reader, into the emergent blend. The process continues throughout reading, until the complete blend of the story emerges. In what follows, I will consider some details of the process. 2.2
Narrative spaces as mental spaces
Mental spaces were defined (Fauconnier 1994[1985], 1997) as conceptual packets set up primarily by linguistic expressions (though visual or gestural prompts are also common), and being used in the on-line mental processes of meaning construction. Thus a mental space may be a recall of a past event in the course of a conversation about the current situation. For example, if the speaker says I was really tired last night, but I’m better now so do tell me what you think about my book she calls up a past space in the first clause, describes her present state in the next, and evokes the spaces of the content of the book, the content of the hearer’s thoughts regarding the book, and the hearer’s expected conversational turn. All these are mental spaces, activated for the duration of this part of the conversation, and then becoming latent until evoked again, when, let’s say, the speaker later complains about the hearer’s bitter criticism to someone else. It is clear even from this example that mental spaces are constructs of varying complexity (from tiredness to the content of an entire book), attributable to various conceptualizers (here, the speaker or the hearer), and including mental representations of concrete objects, conversational moves, opinions, et cetera. It is often asked whether it is possible to confirm the presence of mental spaces in the brain, but the question seems to misconstrue the usefulness of the concept. Even if we could monitor someone’s spontaneous engagement in a conversation and discover which areas of the brain are involved (some have to be!), we could at best capture some of the brain structure relevant to the content of the conversation, but apparently not the thought processes prompted by the discourse. Also, as Spivey argues (2007), the flow of thought, when traceable in the brain, creates a trajectory which maintains a number of open options throughout – very much in agreement with what mental space theorists claim. Mental space analyses have shown how mental space networks can represent the on-line correlation between choice of expression and the resulting meaning.1 But mental spaces are not a strictly linguistic phenomenon, as recent work on gesture, art, comics, theatre, or film shows;2 rather, the theory is a useful tool in teasing out the details of various forms of meaningful interaction and discovering the correlations between such prompts and meaning. The above discussion of mental spaces and blending makes it clear that the framework has been mostly used to elucidate the emergence of meaning of expressions, discourse fragments, or pictorial/verbal artifacts. Fictional
36
Blending, narrative spaces, and the emergent story
narratives, for comparison, are much more complex and also simply longer. It does not seem possible to provide a clear explanation of how individual sentences may be subjected to the overall storytelling goals. To address this concern, I will argue that narrative construction proceeds through several levels of narrative organization, involves an emergence of partial narrative structures called narrative spaces, and is primarily organized by various forms of narrative viewpoint. A narrative space has most of the features of a mental space.3 It is set up through language expressions, often ones not unlike ordinary space builders – such as Once upon a time . . . or In a hole in the ground there lived a hobbit . . . and further structured by grammatical choices such as person or tense. It has its topology, such as space, time, cultural norms, or language spoken. It also has participants, some directly involved in the telling, called narrators, some being directly affected by the events, called characters, though there are also participants (intradiegetic narrators) who play both roles at the same time. Like mental spaces, narrative spaces can be elaborated in much detail, while others are barely mentioned. Both types of spaces are structured by viewpoint phenomena, though narrative spaces seem to feature more elaborate constructions of viewpoint, to be discussed throughout this book. To sum up, a narrative space is a mental construct participating in the emergence of the story, having distinctive topology and narrative status, and linked to other narrative spaces in ways which prompt story construction. Narrative spaces are different from other ‘worldlike’ concepts such as possible worlds (as applied to narrative) which are discussed in relation to the ‘real’ world as its alternatives.4 Narrative spaces are defined in parallel to and in correlation with lower-level linguistic phenomena participating in meaning construction, and, while they may apply to ‘worldlike’ constructs, they do not apply to texts in their entirety. An emergent story results from the blending of all of the text’s narrative spaces. The structure of narrative spaces is thus determined by the specific organization of the text, and no claims are made about their necessary completeness or coherence. The concept of a mental or narrative space also shares some features with the construct which has been termed a ‘text world’ (introduced by Werth 1999, and developed by Gavins 2003, 2007). Text world theory shares some of its assumptions with the mental spaces framework – for example, it also talks about linguistic forms being used to prompt conceptual constructs and accumulating as discourse progresses. It is primarily a theory of the interpretation of discourse, and has been applied to a number of discourse contexts – verbal and nonverbal – including the context of the narrative. The primary difference is possibly that the text world theory is much more text-driven than the mentalspaces-and-blending framework, while also being more focused on the analysis of discourse, rather than the elucidation of linguistic concepts. Mental spaces
2.2 Narrative spaces as mental spaces
37
theory has originally been conceived as a theory of meaning, and, while it is developing its textual applications, it still relies more heavily on semantics than on discourse. Also, mental spaces analyses are naturally extended into work on blending. I will be relying on mental spaces specifically, but I assume most claims can be ‘translated’ into the text world approach. A narrative space is thus a construct which is set up through linguistic means and continues being elaborated through some parts of the text (possibly all). It is also subsequently enriched through blending and gradually starts functioning in the network leading to an emergent story. In these respects it is thus similar to a mental space, which participates in extended discourse. What constitutes a crucial difference, however, is the nature of the discourse, since a narrative requires that its primary spaces be maintained and elaborated until the completion of the reading process, until their role in the text is fulfilled. To give a simple example, if a flashback space is set up, it can then play different narrative roles, not necessarily dependent on time, though in each case it would constitute an independent, temporally distinguished space. Setting up such a space can constitute background to the main story, or the beginning of the story proper, or a digression which explains some form of behavior. In any such instance the space will be incorporated into the story, though not only temporally. 2.2.1
Sentential level versus textual level
It may seem counterintuitive to claim that a narrative space may in some cases take the form of a sentence, but there is no reason to draw a specific line at any level of linguistic structure, because what matters is the narrative structure. For example, the text which won the contest for the shortest story is also a sentence. It was composed by Augusto Monterroso, and it is entitled The Dinosaur: (1) When he woke up, the dinosaur was still there.
The story has all the ingredients necessary, but asks the reader to do much of the story-construction work. There is a narrator (probably omniscient), and a character, plus a dinosaur. The when clause profiles a temporal location wherein the character awakens, which suggests that he was asleep before (in Spanish it might have also been a ‘she’). The word still in the main clause further suggests that before the character fell asleep, the dinosaur was also present. There is thus a sequence of events, there are characters in some relationship, there is a narrator, and there is suspense – how come there is a dinosaur in the presence of a human being? What has often been called ‘suspense,’ in my view, is the need to engage with the text and have the intention to complete the gaps in the story. In other words, the reader is required to use extra-textual and linguistic knowledge to complete the emergent structure of the blend being prompted by the narrative inputs.
38
Blending, narrative spaces, and the emergent story
In the case of Monterroso’s story, there are in fact several options. The dinosaur could be a real T. rex or triceratops, and then the story qualifies as science fiction, and the character is possibly in grave danger (hence the fact that he could sleep in the presence of a monster is intriguing indeed). Or it is a fantasy about a friendly dinosaur, who guards the character when he sleeps. If the spaces set up are not intended to represent the story’s reality but the character’s delusions (think ‘crazy scientist’), then the story might be a variety of ‘stream-of-consciousness,’ following the inner visions of the character. If the dinosaur appeared in a dream, when the character slept, then the reader is faced with a mystery – did the dinosaur cross over from dream to reality? or is the character still asleep, but dreams that he is waking? In fact, there is an impressive number of possible readings of the story, all dependent on how the reader (not the text) construes the spaces set up, and how the spaces are then linked by, for example, the presence of the same characters (‘he’ and dinosaur in the same ‘reality’ or not, dinosaur projected from dream to ‘reality’, et cetera). All the interpretations above are blends – a reality/fantasy blend, a reality/hallucination blend, or a reality/dream blend. What is more, as the ‘dream’ version suggests, it may not be possible or even interesting to resolve what kind of blend it is, so long as the story construction process engages the reader. Even in a longer narrative, the sentential level of the text may crucially contribute to the structure of narrative spaces. Consider example (2), where the main character in Eggers’s novel A Heartbreaking Work of Staggering Genius, Dave, is in his car, struggling with guilty thoughts about having left his small brother with a babysitter: (2) I will come home and the door will be open, wide. The baby-sitter will be gone and there will be silence . . . Blood on the walls, handprints soaked in blood. A note to me, from Stephen, taunting; . . . There will be a hearing, a trial, a show trial – How did you come to meet this man, this baby-sitter? We found a posting, on a bulletin board. And how long did your interview of him take? Ten, twenty minutes. (AHWOSG, p. 126)
The narrator is assigned to an independent space (story-viewpoint space; SVspace), which has the entire story in its scope. The story itself is contained in the main narrative space (MN-space), which consists of a number of narrative spaces (NS). The details of this set-up will be explained in Chapter 3, but example (2) illustrates what it means for lower-level expressions to participate in narrative space construction. The main narrative space in (2) is that of Dave’s life with his brother, Toph. As Dave is driving across Bay Bridge, his mind starts conjuring up an elaborate scenario wherein the babysitter kills Toph, and Dave is then put on trial for negligence. The sentence I will come home . . . thus sets up an imagined future
2.2 Narrative spaces as mental spaces
SV-space
39
Narrator: Dave present
MN-space
Narrator: Dave / ‘I’ present tense Dave’s life
NS1 Narrator: Dave / ‘I’ driving to San Francisco NS2 – crime Narrator: Dave / ‘I’ imagined future babysitter (killer) Toph (victim)
NS3 – trial Narrator: Dave / ‘I’ NS3 – Discourse space S1: Dave S2: prosecutor
Figure 2.1 Narrative spaces (example 2)
space, then developed with the gory details of the crime. Then, from the point of view of the imagined ‘crime’-space, another space is built, that of an imagined trial. Each one of these spaces is set up at the sentence level, and even though their activation does not extend beyond several sentences, they do play a role in the narrative construction (see Figure 2.1). First, the scenario developing in Dave’s mind is one of numerous instances in the novel where Dave’s unrealistic fears and his lack of confidence lead to erratic and exaggerated responses. The ‘crime-and-trial’ sequence of spaces is thus a part of a broader narrative construct representing Dave’s life. While seemingly local, it participates in the higher-level story construction process, which does not have to proceed linearly, in the way the text does, but elaborates spaces incrementally, throughout the text. Second, the example represents one of the key issues of story construction – the management of viewpoint. In terms of mental spaces theory, the spaces just described represent familiar patterns of viewpoint and focus. The space of Dave’s life is the viewpoint space from which the ‘crime’ space is being set up and elaborated, as Dave’s fantasy – it is thus the focus space. As discourse progresses, the ‘crime’ space takes the role of viewpoint space, and the ‘trial’ space is now in focus, and being elaborated. This viewpoint/focus structure explains why the trial is construed as the result of the imaginary crime, and not, say, of Dave’s driving habits.
40
Blending, narrative spaces, and the emergent story
The viewpoint/focus configuration explains how spaces form chains, such that a higher-level space is the viewpoint, and a lower-level space is the focus. However, in order to explain the processing of narratives as complex as novels, we also need an explanation of the mechanism which links various lower-level spaces into clusters contributing to the construal of higher-level spaces. I will discuss this mechanism as cross-input projection. At the same time, the content of the lower-level spaces contributes to the construal of the higher-level space, so that Dave’s fantasy is not only projected downwards from the highest space of his narrative, but also used in the construal of the narrative itself – for example, it gives the reader access to an understanding of what kind of narrator Dave is, and how his narrative as a whole relies on fact and fantasy alike. These higher-level construals result from what I will refer to as viewpoint compression, an integration mechanism which allows lower-level viewpoint to contribute to the higher level. The phenomenon also has its linguistic correlates, such that lower-level spaces are presented without linguistic indications of their being embedded in higher-level spaces. The entire fragment partially quoted in (2) is narrated alongside the description of Dave’s drive, and thus, stylistically at least, it is presented as belonging to the core of the story being narrated. The effect of engaging multiple levels of spaces and presenting them as one is the primary sign of viewpoint compression. I will discuss the phenomenon in more detail throughout the book. Beyond the sentential level, fictional narratives set up various configurations of spaces, distinguished by basic space topology (such as time and space, or participants). Some contemporary narratives rely on an even more fragmented structure, wherein major narrative spaces are presented as independent texts – there can be a novel within a novel, as in Atwood’s The Blind Assassin, different narrators, or stories ostensibly consisting of separated narratives, to be linked at the end, as in Alex Garland’s The Tesseract or in the movie Babel. Accounting for the astounding variety of such narrative choices would not be possible in any analysis, but I will discuss one text, The Blind Assassin, in more detail, to illustrate the processes involved. 2.3
Narrative spaces – an example
Margaret Atwood’s novel The Blind Assassin is a good example of how the multiplicity of narrative space operates in a narrative. The novel’s framing narrative is a memoir written by a now aged and lonely woman, Iris Griffen (née Chase). She lives in the small town where she grew up and writes the story of her life, to be read after her death by her estranged granddaughter. The main narrative space thus has a first-person narrator, Iris, remembering the past. Consequently, large parts of what we are reading are descriptions of Iris’s daily struggle to go on writing, but the main story of the novel is reconstructing
2.3 Narrative spaces – an example
41
Iris’s entire life. These two narrative spaces of ‘Iris’s present’ and ‘Iris’s past’ are the core of the narrative structure. Also embedded in the text is a novel, titled “The Blind Assassin,” ostensibly written by Iris’s sister, Laura. It is written in the present tense, with just two characters never identified by name – just a ‘he’ and a ‘she.’ These grammatical choices (no I narrator, present tense) help the reader maintain the embedded novel’s relative independence, even as it is gradually incorporated into the story. In the embedded novel, the two anonymous characters compose a pulp sciencefiction story, taking place on the planet Zycron, and featuring a blind assassin as one of the (again nameless) characters. This story is never completed, and thus its role in the entire text is difficult to see. Finally, the text is sometimes interrupted with short newspaper clippings, giving an official version of some of the events. The clippings often contradict the information provided by Iris’s memoir. They jointly represent a narrative space telling the official version of some of the events, reappearing throughout the text. The space is anchored by the format of the clippings and is reactivated as the need arises. Taken together, the clippings construct a narrative space which tells a story different from the one narrated in the rest of the text. All these major spaces are presented as independent texts, with different authors and characters, occurring in different locations and times. As Figure 2.2 shows, they are initially set up as independent spaces – and the set-up changes
SV-space
Narrator: Iris present
MN-space
Narrator: Iris Author: Iris past tense Memoir Narrator: Iris past tense Laura Alex Richard Father
Clippings Author: journalist Mrs. Richard E. Griffen Embedded novel Author: Laura Narrator: 0 present tense he she
Sci-fi story Author: he Narrator: 0 present tense mute girl (she) blind assassin (he) Lord of the Underworld King
Figure 2.2 Narrative spaces in The Blind Assassin
42
Blending, narrative spaces, and the emergent story
as the text progresses. Crucially, they all have to be treated as subordinate to the space of Iris narrating (they are within the scope of the viewpoint of the SVspace), and are gradually identified as complementary to the memoir she writes. Coherence links are suggested throughout and the fabric of the story emerges through various links across the narrative spaces. Even more important, all the viewpoints of all the spaces are finally compressed with the viewpoint of the initial set-up with Iris as the narrator. The case of The Blind Assassin illustrates all the central processes of story construction. I will attempt to elaborate them below, starting with the concept of narrative anchors. 2.3.1
Narrative anchors
I have introduced the concept of a narrative anchor in earlier work, focusing mainly on The Blind Assassin and F. Scott Fitzgerald’s The Great Gatsby (Dancygier 2007, 2008a). In both of these cases I argued that the fragmented nature of most narratives requires specific narrative mechanisms which provide coherence links across different narrative spaces. I defined narrative anchors as expressions which set up or suggest the availability of narrative spaces, but do not elaborate them right away. Such ‘place-holders’ may activate new narrative spaces and allow them to remain active, but the spaces are elaborated gradually as the text unfolds, and often contribute to the topology of other spaces constituting the story. The second function of anchors is ‘link-building’. The links they set up may prompt what I have called cross-input projection – spreading of topology from one input to another and building the coherence and completeness of the emergent story. To refer to the ‘clippings’ again, they are constantly crosslinked with other spaces, adding topology, foreshadowing events to be narrated later, but, first of all, presenting a viewpoint not available in other narrative spaces. The usefulness of anchors becomes clear when we realize that it is difficult to imagine a fictional narrative which would explicitly provide the entire content of its narrative spaces. Narrative coherence crucially depends on selection of the content needed at a given point in the story, and on suppression of all information which would reveal too much or drown the important narrative elements in a chaos of unbounded verbosity. Narrative anchors are difficult to define formally because they are not restricted to any specific aspect of narrativity which is typically discussed (plot, characterization, et cetera), but they may participate in all. They are best defined in terms of the processes they trigger which lead to the construction of the story. Various expressions can play the role of anchors, and the choice depends on the type of cross-space connection to be established. One can thus expect that well-selected referring expressions would lead to connections in terms of
2.3 Narrative spaces – an example
43
finding counterparts in different spaces and establishing identity (see 2.3.2 for more discussion), but descriptive expressions may play the same role. For example, the man named Alex in the main narrative space of The Blind Assassin has to be recognized as being the same as the one referred to only as ‘he’ in the embedded novel space, even though the two counterparts are characters in different stories, taking place in different space and time: (3) He stubs out his cigarette, reaches for another, thinks better of it. He’s still smoking ready-mades, a luxury for him. (The Blind Assassin, p. 133, embedded novel) (4) He produced a packet of cigarettes from his shirt pocket – Craven A’s, as I recall. He tapped one out for himself. I was a little surprised that he smoked ready-mades – it didn’t go with his shirt. (The Blind Assassin, p. 222, Iris’s memoir)
Making the connection is prompted by careful deployment of the same bits of information in both narratives, and by repetition of some of that information. Bits of characterization, such as (3) and (4), are spread throughout the narrative spaces via cross-input projections and jointly create the character’s identity which fits coherently in all the narrative spaces. Crucially for this example, as the page numbers show, the reader first sees the man in the embedded novel smoking ready-mades, but is also alerted to this trait by the use of still – which indicates a time prior to the scene when the same trait was observed. So when Alex Thomas is first introduced in Iris’s memoir (in a scene temporally prior to that of [3]) and the same information is presented as surprising, the reader can make the connection between the two characters. In fact, guessing Alex’s identity is made easy, but the identity of the woman remains elusive till almost the end of the novel. Anchors can also rely heavily on frames. As shown by Lermitte (2010), anchors may also be images which form an entire network of concepts and jointly give meaning to an abstract and difficult text. In Annie Dillard’s nonfiction text Holy the Firm the meaning is not so much a story as a reflection on religious and artistic passion and a life centered around such deeply felt emotions. The crucial image is that of a moth being attracted to the flame of a candle, dying, and acting as the candle’s wick as a result, allowing it to burn longer: She burned for two hours without changing, without bending or leaning – only glowing within, like a building fire glimpsed through silhouetted walls, like a hollow saint, like a flame-faced virgin gone to God. Passages such as this recur through the text, highlighting various dimensions of the emerging blends. The recurring images of burning as a sacrifice which in the end prolongs life allow Dillard to tell a poignant story of creative endeavors and religious devotions as acts in which an individual needs to ‘die’ in order to ‘keep the flame burning’ and go on living through what has been created. The anchors do
44
Blending, narrative spaces, and the emergent story
not simply ‘construct the story’ here, they construct its meaning through a network of blends and frame metonymies. Without them, the text achieves no narrative coherence. These various examples of anchors pose questions as to the actual cognitive nature of anchoring. It seems that it relies on a variety of possible construals and mappings. In the work done so far, I have distinguished several types. First, there are anchors which are quite naturally space builders. When the text of The Blind Assassin repeatedly refers to ‘notebooks’ a now dead character left in hiding, the content of the notebooks constitutes a narrative space which promises to fill huge gaps in the story. Naturally, when the notebooks are found, they do reveal crucial facts. Another type of anchor consists in evocation and repeated re-activation. A mention of Alex as wearing a blue, worker’s shirt but smoking ready-made cigarettes is a salient descriptive detail, and its reactivation in a different narrative space prompts the cross-input projection linking Alex to ‘him’ and making this cross-space identity available in both spaces. This type of anchoring relies as much on the reader’s attentiveness and memory as on the salience of the frames thus constructed and on the sheer number of anchors establishing and reestablishing the same cross-input links. Crucially, such anchors not only link the spaces, but also prompt projections of topology from input to input (narrative space to narrative space). The identity of the character named Alex in one input is cross-linked to the identity of an anonymous man in another, but the information from one input is thus also available in the other. Below I will present an extended analysis of a pair of anchors in The Blind Assassin which are central to the narrative meaning of the entire novel. Throughout the novel, these anchors play all of the roles described above – narrative space building, prompting cross-input mappings, establishing identity, metonymy, and metaphor. While this may be a special case, it does show the potential of anchoring in a complex story.
2.3.2
Anchoring: representation blends and frames
The main narrative space of the novel is the life of two sisters, Iris and Laura, and their involvement with a man, Alex Thomas. Iris is the narrator, while Laura, who has committed suicide, is not given a narrative voice. Even though one of the narrative spaces is a novel she presumably wrote, it turns out in the end that the novel was also written by Iris and published under Laura’s name. Throughout the reading, the text playfully alternates between suggesting Laura or Iris, leaving it to the reader to construct the true identity of the narrator of ‘Laura’s novel’ (I will refer to it as ‘embedded’). The authorship, in turn, also identifies the writer as Alex’s secret lover. While there are many salient anchors
2.3 Narrative spaces – an example
45
leading the reader to a conclusion, the main anchors are a photograph and a hand. They are initially mentioned in the first lines of the embedded novel: (5) She has a single photograph of him . . . the photo is of the two of them together, her and this man, on a picnic . . . he’s holding up his hand, as if to fend her off in play, or else to protect himself from the camera . . . Over to one side – you wouldn’t see it at first – there’s a hand, cut by the margin. (The Blind Assassin, p. 7)
The fragment is full of unanswered questions – who are the people? whose hand is visible at the edge? what happened to the man? and many more. They all emerge because of the fact that the concept of a photograph (similarly to picture, painting, drawing, et cetera) prompts a blend connecting two mental spaces: the representation space (distinguished by its form, colour, materiality, boundaries, et cetera) and the represented space (the actual reality it captures, with its temporal and spatial features, identity of participants, as well as, crucially, topology available beyond what the representation shows). In the fragment, only the representation space is provided, while none of the features of the represented space are clarified – hence all the unanswered questions which frame the entire ensuing narrative. It is thus a perfect example of a narrative anchor. It opens up a narrative (represented) space which is necessary to complete the representation blend prompted by the description of the representation space, and the anchor remains active until the blend is complete – until the end of the novel. The narrative of The Blind Assassin uses this anchor throughout to keep the reader guessing. It is reactivated regularly, and each time a new piece of the represented narrative space becomes available. The next mention, in Iris’s memoir, is as follows: (6) Then he took a picture for the paper with his camera . . . Alex Thomas raised his hand as if to fend him off. (The Blind Assassin, p. 223)
This appears in a fragment describing a factory picnic when the sisters first met Alex. The identity of the man in the photograph is suggested by the second mention of his gesture, but since both women were present at the picnic, the woman telling the story in the embedded novel is still mysterious. Finally, the newspaper report of the factory picnic is published: (7) One of the pictures was of Alex Thomas, with the two of us – me to the left of him, Laura to the right, like bookends. Both of us were looking at him and smiling too, but he’d thrust his hand up in front of him . . . The caption was, “Miss Chase and Miss Laura Chase Entertain an Out-of-Town Visitor.” (The Blind Assassin, p. 241)
The anchor of Alex’s raised hand is repeated again, but the identity of the woman in the photograph in (5) becomes even harder to discover. The fragment makes it clear that while the represented space is now known (Laura, Iris, and
46
Blending, narrative spaces, and the emergent story
Alex at the picnic) the representation in the photograph mentioned in (5) is incomplete. One of the women – Laura or Iris – is missing from the photograph, and the hand at the edge is thus a token of the presence of ‘the other woman.’ At this point, the identity of the narrator of the embedded novel has to match the woman in the photograph, and would also point to the one whose hand is visible. It is clear that the writer is playing this ‘now you see it, now you don’t’ game on purpose. In fact, the remaining (very rich) narrative is playfully providing ‘helpful’ information which alternatively points to Laura or Iris, and thus the text offers an abundance of anchors which do not materialize into complete spaces or unambiguous cross-input links until the end of the novel. In this context the ‘hand’ becomes another important anchor, and carries with it a rich array of meanings. At the most basic level, the hand in the photograph is simply an indication that there was a person sitting next to Alex – it is thus an anchor to the undisclosed identity of the other woman at the picnic. At the same time, it suggests that someone (the narrator of the embedded novel?) did not want the other sister in the picture – in the literal and figurative sense. The ‘hand’ is not only the hand of someone who was present in the represented space and time, but the hand that was left after someone tried to erase that person from the memory of the events. There are several levels of frame evocation here: ‘hand’ evokes the entire body and an entire person, the deliberate distortion of the original scene evokes someone’s intent, and the incomplete photograph is also an anchor to the missing part of the narrative, where the character forcefully removed from the story tells what had really happened. The double anchor of the photograph and the hand is a token of what the narrative needs in order to achieve coherence – the missing part of the story connected to the other woman. Later in the text we learn that Laura had access to the negative and produced two prints from it, each with one sister absent (save the ‘hand’); what is more, following a childhood habit of colouring black and white drawings, she stained each of the ‘hands’ a different color: Laura’s hand is yellow, Iris’s hand is blue. (8) But she’d cut herself out of it – only her hand remained. [yellow-Laura] ... “I have another one [photograph], for me.” “And I’m not in yours?” “No,” she said. “You’re not. None of you but your hand.” [blue-Iris] This was the closest she ever came, in my hearing, to a confession of love for Alex Thomas. (The Blind Assassin, p. 277)
This explains how there could be two photographs in the possession of two sisters – evoking two stories of love and loss related to Alex Thomas. However, both stories are in a sense written into both pictures, for even though the narrative might focus on one of the women, it will not be complete without the other.
2.3 Narrative spaces – an example
47
The hand/photograph anchor begins to interact with other anchors later in the text. In the embedded novel the photograph is kept in a gardening book, which also appears in the main narrative space (Iris’s autobiographical story), where she describes ways of occupying herself in the difficult period of her married life. “Perennials” is thus another anchor, identifying the narrator of the embedded novel as Iris: (9) She has a single photograph of him. She tucked it into a brown envelope . . . between the pages of Perennials for the Rock Garden, where no one else would ever look. (Prologue: Perennials for the Rock Garden) (The Blind Assassin, p. 7)
In the final chapters, Iris finally acquires her sister’s notebooks, mentioned in the very first paragraphs of the novel as a source of crucial clues to Laura’s suicide. (10) I riffed through the other notebooks. History was blank, except for the photograph Laura had glued into it – herself and Alex Thomas at the button factory picnic, both of them now colored light yellow, with my detached blue hand crawling towards them across the lawn. (The Blind Assassin, p. 627)
The hand’s suggested motion (crawl) evokes intended interference in the lovers’ encounter, as the body image of someone crawling evokes bent posture, such that the couple might not even notice anyone approaching; this in turn could be read as signaling evil intentions. The ‘crawling hand’ is narratively equal to a crawling person, someone who might then hurt the couple on the lawn. There are several important points to note here. First, this very low-level narrative choice has text-wide consequences, precisely because it appears in connection with the novel’s primary anchor. One of the points this analysis makes is that a narrative achieves its meaning through the compounded effect of meanings emerging at the lowest lexical and grammatical level. Second, it is important to note that the view of Iris as a destructive intruder which (10) suggests comes at the end of the novel, when Iris realizes that she unknowingly caused her sister to commit suicide. For Iris, the narrative voice of the story, to describe herself as my detached blue hand crawling towards them across the lawn, is to complete the untold story of the photograph. The representation blend has not only been fully integrated, through a match of the representation space and the represented space, but it achieves its narratively satisfying status through the interpretive viewpoint of one of the participants. The facts of the picnic as such would not tell the story that led to Laura’s death, but Iris’s realization of her own role in her sister’s life is the story being told throughout the book. In the end, ironically, the true story of the novel is told through the truncated photograph, not the photograph published in the paper, with all three of the main characters sitting together. In a sense, the way this anchor shapes the story in The Blind Assassin is by adding viewpoint and
48
Blending, narrative spaces, and the emergent story
causation to the events which otherwise would not have been framed to achieve any psychological effect. The role of the ‘hand’ anchor does not end there. While it structures much of what happens in the novel, it also participates in the construal of narratorship and narrative viewpoint. Narratorship here is not only the question of an epistemic viewpoint (who knew enough to tell the story in this way, a crucial consideration with respect to the embedded novel). Iris is ostensibly telling her own story, but it becomes clear as the text progresses that she is, perhaps primarily, telling the story of her sister’s life, as she finally understands her own role in it. In the last sections of the novel she talks about herself as the ‘hand’: Laura’s odd, extra hand, attached to no body – the hand that passed her on. The frame of the ‘hand’ now refers to a common understanding of the ‘hand’ as the body part capable of manipulating and doing things. The detached hand in the picture now stands for Iris’s decision to attribute the novel to Laura. The embedded novel tells the story of Iris’s secret relationship with Alex, but granting Laura the authorship has ironic consequences in view of the fact that the characters in the embedded novel are never referred to by proper names, only as ‘he’ and ‘she.’ By giving Laura the authorship, Iris’s ‘hand’ in a way also gives the participation in the story to her sister – something that could have been the case and which was clearly Laura’s deepest desire. In fact, people in Iris’s family understood the novel as Laura’s confession of a secret affair, so from their point of view, Laura was Alex’s lover. By referring to herself as Laura’s ‘extra hand,’ Iris removes herself from the story as its full-bodied participant – again, something the anchoring photograph supports. The anchoring image of the severed hand is reframed as the narrative voice. Iris explains this further: (11) I can’t say Laura didn’t write a word. Technically that’s accurate, but in another sense – what Laura would have called the spiritual sense – you could say she was my collaborator. The real author was neither one of us: a fist is more than the sum of its fingers. . . . Laura was my left hand, and I was hers. We wrote the book together. It’s a left-handed book. That’s why one of us is always out of sight, whichever way you look at it. (The Blind Assassin, p. 644)
These fragments provide crucial metanarrative comments. The tricky structure whereby a myriad of anchors keep on pointing to first one sister and then the other is now described in its full meaningfulness. Throughout the reading, one is tempted to keep on guessing which one of the sisters is the ‘she’ in the embedded novel, and the reader’s detective skills are rewarded when Iris admits that she wrote it and she was Alex’s lover. But the final comments quoted in (11) suggest that the shimmering narrative has a different purpose. The story is now seen as having two emergent structures, not just one. There is the narrative Iris constructs to set the record straight and explain her sister’s suicide, but there is also the narrative where she and her sister become one narrative agency
2.3 Narrative spaces – an example
49
and thus share in the experience of the story as well as the telling of it. The narrative thus reconstrues Laura’s life, by giving her a chance to experience what could not happen in the novel’s reality and to become an agent in the way the story is told. In the final sections the double photograph/hand anchor plays its last narrative trick. The embedded novel is spread chapter by chapter throughout the text, with other parts of Iris’s narrative taking much space in between. The Epilogue starts very similarly to how the embedded novel begins – see the fragments quoted in (5) and (9) for comparison: (12) The photo has been cut; a third of it has been cut off. In the lower left corner there’s a hand, scissored off at the wrist, resting on the grass. It’s the hand of the other one, the one who is always in the picture, whether seen or not. The hand that will set things down. (Epilogue: The other hand) (The Blind Assassin, pp. 649–650)
There is one striking difference, though. In the Prologue ([9]), the woman is clearly identified as Iris through the Perennials anchor (available later in the text), and the description of the photograph does not specify whether it is the left or the right hand that is seen next to the man. In the Epilogue, the hand is identified as the left hand, which clearly suggests that it belongs to Iris (when we imagine looking at the photo as described in [7], with Iris seen to the left of Alex). So the woman looking at the photo in the Prologue is Iris, but the woman looking at it in the Epilogue is Laura. This also explains why the ‘hand’ in (12) is the one that will ‘set things down,’ that is, Iris. What is more, the Epilogue also reconstructs the embedded novel to match the entire story of the photograph. The Prologue introduces Iris’s viewpoint, the text in the middle focuses on Alex and his lover, whoever she is, and the Epilogue adds Laura’s viewpoint. To make the analogy clearer, the women in the Prologue and the Epilogue are construed as ‘viewing’ their respective photographs, so that the concept of narrative viewpoint is clearly evoked. The anchor has now achieved its full potential and it explains what Iris means by saying that both sisters had their ‘hand’ in the embedded novel. The anchor described here requires a careful reader. Interestingly, critics (Roy 2006; Perrakis 2006; Wilson 2006) treat The Blind Assassin as just a postmodern exercise in instability of meaning. But the acknowledgement of the anchors, as I hope to have shown, does not trivialize the novel – on the contrary, it helps reveal its double bottom and its poignant view of what narrative fiction does. Just telling the story of what happened is not good enough, it seems to be saying, because the narrative may be the story that should have happened, but never did. The crucial role of the anchors discussed above depends largely on the nature of the concepts they metonymically evoke. The idea of a photograph is in itself a rich source of mental space topology. The other anchor, the hand, is also craftily
50
Blending, narrative spaces, and the emergent story
chosen. Being associated with the body but also with the human ability to manipulate, make new things, destroy things, et cetera, the ‘hand’ is easily seen as metonymic for a number of actions the person thus represented can undertake and perform. Moreover, all the actions associated with the hand also require a will and a subjectivity, thus in spite of not bearing any agency in itself, this body part can naturally stand for the subjectivity of a person. Atwood’s use of these anchors capitalizes on all of these aspects of the framing of photos and hands. The naturalness of the story construction processes triggered by the anchors supports Deacon’s (2006) discussion of how art works. He argues that in a broader context a single word such as hawk is much more than a symbol for an actual species of a bird, but rather a way to activate a network of concepts in the mind of the hearer/reader (bird of prey, powerful beak and claws, ruthless hunter, linguistic form, emotional response, et cetera). Furthermore, activating the image of the claws and predation may simultaneously evoke the concept of a victim, fear, and pain, or, for comparison, power, courage, attack, and pride. I’m reading Deacon’s claim as an acknowledgement of the aspects of the process of reading which rely not only on the constructing power of the words alone, but also on the richness of experiential frames already present as the background to reading. The anchors in Atwood’s novel seem to support this. While some are relatively simple (for example, Alex’s ready-mades), others interact with seemingly unrelated frames present in the reader’s mind. Also, all such frames evoke emotional responses needed in processing texts. Relying further on Deacon’s work, one can claim that anchors are an ultimate way of building on what he refers to as our ‘representational stance’ – the natural assumption of human cognition that something stands for something else. It is a more straightforward (though not simple) process in our use of language as a system of symbolic representations, but it is more complex in the narrative. Naturally, the narrative expressions are symbolic in the ordinary sense of the word, but they may be deployed in the text in a way which poses a challenge to our representational stance. We do know what a photograph is and we expect it to represent a recognizable reality. If that reality is not provided, it can still be constructed. We can thus say that narrative anchors are narratively salient expressions which rely metonymically on frames and exercise our representational abilities, but create suspense by providing necessary information only bit by bit. To sum up, the cognitive impact of anchors on the reading of narratives is multidimensional. At the simplest levels, anchors are frame metonymies (photograph for the representation blend it constructs, fire for pain, destruction, purification, light, et cetera). Moreover, the textual situation of anchors guarantees the emotional response Deacon sees as indispensable in the appreciation of literature. Perhaps more important, the patterns according to which anchors are reactivated and gradually enriched with new framing are also the patterns of
2.3 Narrative spaces – an example
51
narrative construction at every level. They participate in establishing temporal sequences of events, in revealing identities of characters, and in supplying missing elements of the plot or prompting their reconstruction. Overall, they provide the skeletal system of interacting frames which, as they are enhanced with new topology, point to crucial narrative spaces and connections between them, and thus lead to the emergent story. Narrative anchors capture more than the progressive accumulation of content and context, as described, for example, in Werth’s study of text worlds (1999) or Emmott’s discussion of ‘contextual frames’ (1997). They are also a more specific mechanism than what literary critics refer to as ‘motifs’, especially in the mental-space setting function. They exploit the mechanisms of frame metonymy, frame evocation, mental space set-up or evocation, and crossspace projection. At the same time, they capture the complex interaction between specific expressions used, emergent connections across various parts of the text, and the reader’s processing of it. 2.3.3
Reference and story construction
The discussion above shows how narrative anchors of various types play a role in prompting cross-input projections and the emerging cohesion of the story. Here, I will consider the ways in which referential expressions contribute to the emergence of the story. The central piece in the narrative-space structure of The Blind Assassin is the embedded novel, featuring only two characters, referred to through pronouns only. The pronouns he and she let the reader assume that these forms refer to subjectivities identified elsewhere in the novel, and thus a large part of the reading process consists in attempts to establish cross-space identity links which would explain who the characters are. He is soon identified as Alex Thomas, while the decision about the identity of the she oscillates between two sisters, Iris and Laura Chase. However, another narrative space of the novel, a science fiction pulp story Alex is composing, also features no names, using only descriptions of roles instead. The roles are specific to the text and not very inventive (‘blind assassin,’ ‘mute girl to be sacrificed,’ ‘king of the underworld’), but their presence in the overall novel and the metaphorical understanding of their characteristics (blind, mute, hidden) provide important clues to the understanding of the story (cf. Schnepf 2006 for more discussion). The story these characters participate in explains the facts from their lives which are not represented anywhere in the text, but the role of these facts is appreciated only through the matching of the roles with characters in other narrative spaces. Crucially, the entire construal of the sci-fi story represents the viewpoint of the in-novel writer, Alex, who is thus revealing the nature of his participation in the story. He was not given his own
52
Blending, narrative spaces, and the emergent story
voice anywhere else in the narrative, and thus the pulp sci-fi space is where his viewpoint is construed. Without the reliance on the role of referring expressions The Blind Assassin could not be imagined in its actual form. It forces the reader to construct dynamic frames to represent the identities signaled by proper names – otherwise, the entire reasoning of the choice between Iris and Laura as the ‘she’ of the embedded novel could not take place. It builds on the assumption that personal pronouns do have fully framed counterparts within the spaces of the text. It constructs roles which allow the reader to place characters in the appropriate frames. And it also relies heavily on the construal of the SV-space as profiling the narrator matched with a character (Iris), so that the spaces construed from another viewpoint are constantly matched with the SV-narrative. Crucially to the point of this project, the establishment of cross-input projections based on the links between names, role-descriptors, and pronouns is an exploitation of low-level linguistic mechanisms on the level of the extended discourse, such as a narrative. Knowing that a pronoun requires an antecedent in some mental space or other is what prompts readers into the attempts to seek referents of the expressions in the embedded novel in other parts of the narrative. Similarly, the frames set up in the sci-fi text, with their attendant roles, create the need for further referential links. All these referential connections are not simply available in the text of the novel: they need to be constructed (Figure 2.3 shows the final resulting set of referential links). In a sense,
SV-space
Narrator: Iris present
MN-space
Narrator: Iris Author: Iris past tense Memoir Narrator: Iris past tense Laura Richard Father Alex
Clippings Author: journalist Mrs. Richard E. Griffen Embedded novel Author: Iris Narrator: 0 present tense he she
Sci-fi story Author: Alex Narrator: 0 present tense Lord of the Underworld King blind assassin (he) mute girl (she)
Figure 2.3 Referential connections across spaces in The Blind Assassin
2.4 Emergent story
53
referential frames (profiling slots for a name, a pronoun, and a role) serve as prompts for the filling in of all the slots, even though there are no clear anaphoric or cataphoric patterns to rely on. Such a text-wide deployment of reference seems to be a characteristic of narrative discourse. This form of reliance on referential expressions is used in cross-input connections and the resulting enrichment of input narrative spaces with content. But we should also note that in the process of reading, the distance between the referential slots in individual spaces needs to be compressed, so that the final complete identity of characters emerges as a result. Blending and compression are thus participating in story construction, even at the level as basic as identity of characters and reference. Referential expressions in the narrative pose a complex and multidimensional problem. Starting with the distinction between first- and third-person narration, continuing to the less usual forms of second-person narratives and an occasional use of we, the range of usage of pronouns in narratives poses complex questions about the internal coherence within a text. In Chapter 3, I will consider pronouns used as text-wide narrative choices; Chapter 5 will look further at the question of proper names and roles, while Chapter 7 will consider the use of pronouns in Speech and Thought Representation (STR).
2.4
Emergent story
In much of the narratological literature the concept of the story is defined in relation to text.5 The text is the form (some also mention discourse as an additional element, cf. Toolan 2001) while the story is what the text represents. While the text may be fragmented, incoherent, temporally disorganized, the story is a temporal sequence of causally linked events leading to a resolution of some conflict or problem. The contrasting concepts seem very much like an elaboration of the form/meaning pair, but there is an additional claim, made with special clarity by Chatman (1978, 1990), arguing that the story as the abstracted chain of events remains the same when it is told through a different text or a different medium, such as a film. This may be related to the linguistic tradition of maintaining that propositional meaning remains constant under various formal changes, so that it is natural to see the ‘story’ as the meaning of the ‘text.’ Regardless of the source of the distinction, the ‘story’ is one of the most salient concepts in narratology, but it is typically not discussed in terms of its cognitive status emerging in reading. While Herman (2003a) talks about cognitive aspects of story construction such as chunking of experience or attributing causation, the concepts he evokes refer to the general concept of a story as a way to structure experience. The ‘story’ under discussion here, on the other hand, is the semantic correlate of a specific text.
54
Blending, narrative spaces, and the emergent story
The text, however, should not be seen as simply the words and syntactic structures, as it also relies on cultural availability of frames structuring the events described via the choice of words. The text is not just a collection of linguistic forms, it is also a rich resource of framing. To explain the emergence of the construct called a ‘story,’ I will treat textual choices as prompts guiding layers of blends leading to the sequential, causally structured configuration. Finally, the story is the final result of several modes of interaction with the text – reading the words, activating the frames, searching for correlates in one’s experience, making cross-space connections, blending narrative spaces, establishing identities, constructing tentative scenarios, storing them in memory, revising them as new events are narrated, responding emotionally, et cetera, et cetera, et cetera. The richness of the story as a construct can best be explained if we attempt to discover the processes that lead to the final result. Much of this work is already being done in various contexts – consider Gerrig’s work in psychology (1993), Emmott’s studies combining cognitive psychology and discourse analysis (1997, 2002, 2003; Emmott et al. 2007), or Herman’s renewed vision of narratology (2002, 2007). The goal of this study is to give the view afforded by cognitive tools used in the analyses of linguistic meaning. 2.4.1
Sequence of events and the story
One of the most contentious aspects of ‘storiness’ is its idealized status which, as narratologists argue, remains the same in different tellings. The claim calls for an explanation of where such a construct resides. It cannot be an overall interpretation arising in a single reader’s mind, since such a construct would be too specific. If, on the other hand, the concept is expected to remain constant across choices of a medium, across readers, or even across readings (as readers may change their perception in time), then it has to be ‘boiled down’ to the sequence of events and some conflict to be resolved. It would then be a very minimalist view of the story. The sequentiality of the story raises questions about the temporal discontinuities in texts, which are a norm rather than an exception. In fact, the need to disrupt temporal sequence appeared important in the most celebrated of epics – the Odyssey. It starts in the middle (with the episodes when Telemachus attempts to take control of his father’s household, while Odysseus is held captive on Calypso’s island), changes narrators, and does all the ‘modern’ narrating besides. But knowing why Odysseus is missed on Ithaca puts his entire journey in a different light and leads into the conclusion where the point is not that the journey ends, but that balance returns at last. Also, temporal infractions serve specific narrative purposes. Even in classical texts, not as experimental as contemporary ones, flashbacks are used for specific reasons. The story in Pride and Prejudice starts just before the ‘eligible
2.4 Emergent story
55
bachelors’ appear on the scene and proceeds, through various complications, until the two couples are happily married (it is unfair to Austen’s talent to summarize her masterpiece in this way, but time is all that is being considered here). In fact, there are important flashbacks in the novel: one when Wickham tells a false story of his youth which puts Darcy in a bad light, and the other when Darcy tells the true story of what happened between them. So it is by far not enough to describe these stories as flashbacks, since they present two opposite images of the main character, and entirely change the course of the story – the first makes marriage with Darcy impossible for a person with Elizabeth’s sensibilities, the other makes it very desirable. They do so not because of what ‘really’ happened in the story’s past, but because the versions of the past are crucial to characterization. The little deviation from sequentiality is thus not just acceptable, but central to the construction of characters’ dispositions. Broadly speaking, the list of temporal infractions in narratives is so long and varied that insisting on the centrality of this dimension of cognitive organization seems questionable. What is more, it is also questionable on experiential grounds. First of all, sequential interpretations of events profiled in linguistic constructions often emerge from construal, rather than from straightforward meaning of linguistic forms. They had a baby after they got married or They first got married and then had a baby clearly indicate the sequence of events with after, first, and then, but They got married and had a baby or They had a baby and got married only allow one to construe the sequence one way or the other based on assumptions of the clause order reflecting the order of events. But in real situations we only follow the sequence to the degree to which we experience the events ourselves or are exposed to their results. Temporal sequence is rarely relevant to our understanding of events, but knowing their consequences is crucial. The sequence is a questionable criterion even in the most sequential of stories, but epistemic stance and understanding of causation seem to matter much more (see Dancygier in press b for more discussion). The popularity of the ‘sequential’ core of narratives can be explained if we consider the cognitive construction of time. Stories develop through time and our conceptualization of time assumes something like a linear path progressing from prior events to the events that follow. The discussion of metaphors of time and their instantiations (Lakoff and Johnson 1999, Evans 2004, Nuñez and Sweetser 2006) postulates the emergence of our temporal concepts from our experience of motion through space (which not only changes our position but also takes time). Motion through time also assumes a specific orientation, such that the past is ‘behind’ us and the future ‘ahead’ of us. Stories also assume a direction, following a path from the story’s past to its future and its conclusion, among other reasons because the characters’ experience of the story’s events is modeled after our experience of time, but primarily because a sequential
56
Blending, narrative spaces, and the emergent story
organization is best suited to the representation of the causal organization, which seems more central to narrativity. In other words, even if the events do not appear in a sequence, the causal chains suggest such a sequence, and, as a result, the entirety of the story is best viewed as sequential, even though its telling may not follow the temporal chain. Also, even the spontaneous oral stories, such as the ones analyzed by Labov (1972) and Labov and Waletzky (1967), include components such as evaluation, where events from the story’s past or even future can be evoked to justify the telling of the main line of ‘complicating action.’ To conclude, sequentiality is useful as the final construal, and as such it is also the most salient one.
2.4.2
Story construction, vital relations, and optimality constraints
A ‘story’ can thus be discussed as a cognitive construct, a blend, emerging through the process of meaning construction triggered by reading. The role of the text is crucial in providing such prompts, but the emergence of the story relies to a comparable degree on the frames evoked in the reader’s mind, and on the construction of double-scope blends, integrated into the mega-blend. The story is the mega-blend arrived at in the interaction with the text. The processes of integration, as described above, rely on all the primary aspects of blending. Narrative spaces, such as the ones in The Blind Assassin, are set up as independent inputs, but then selective projection utilizes the elements needed in the blend. For example, the story of The Blind Assassin does retain all the hesitation about the authorship of the embedded novel, and retains Iris as the actual author (Figure 2.3). Also, subsequent blends throughout the process affect the understanding of the inputs, via backward projection. Finally, the story-blend itself undergoes the required integration and elaboration (so that, for example, Alex and he are fused into one participant, but the understanding of his identity relies on setting up a coherent view of his persona). Vital relations play a role in the construction of the story as well. Because of the two inputs, Alex and he are not really connected at all at first, then become connected by the relation of Analogy, until they fuse into the Identity relation in the emergent story, and in the final stage he is also linked to the ‘blind assassin’ via a Role–Value link. Similarly, the relation of the role of the author of the embedded novel is not clearly attributed to any of the Chase sisters, but in the final blend Iris is linked to that role, as well as the role of the lover in the embedded space and the role of the mute girl in the sci-fi story (she cannot tell anyone either about her relationship with Alex, or about her sister). The vital relation of Time also affects the blend, as temporally unconnected spaces (the memoir and the embedded novel) are compressed into the sequential organization of the story; the relation of Change becomes the relation of Causation as we
2.4 Emergent story
57
discover why Laura committed suicide, et cetera. Enumerating all the vital relation changes would take far too much space, but these examples support the claim that story construction progresses along the lines proposed in the blending framework. Construction processes described above also satisfy all the optimality constraints, as described by Fauconnnier and Turner (2002). The integration principle allows representations in the blend to be manipulated as a single unit, and that is what appears necessary in the story of The Blind Assassin, after all the identity relations are established. The blend maintains the inner structure of the inputs (topology and web principles), while tightening metonymies (as in the case of the hand and the photograph discussed above). Finally, the understanding of the entire story can be unpacked into an explanation of how the blend integrates the inputs, and the blend is necessary (good reason principle) to explain the events and the role of characters. Although The Blind Assassin seems to be an unusual example of the difficulties of story construction, it is not more than a useful one. One of the reasons to argue that the story is an emergent structure is to consider differences among readers. The emergent construct is not the same for all readers as different aspects of integration can come into focus on different occasions – even the critics’ responses to novels show that foregrounding certain aspects of the story yields a different blend in the reader’s mind. Thus the stability of the story in various media (such as film versus text) is an idea requiring further investigation. While it may not matter much if some bits of plot are omitted from the movie or if secondary characters are fused, the process of arriving at the story through visual means may highlight or downplay frames which play different roles in the reading. There is enough similarity in most cases, but because the process of arriving at the final blend is different, the results may also not be fully compatible.
3
Stories and their tellers
I am the man who comes and goes between the bar and the telephone booth. Or, rather: that man is called “I” and you know nothing else about him . . . (If on a Winter’s Night a Traveller, Italo Calvino)
The preceding chapter has outlined the concept of a narrative space as a textually prompted construct used in story construction. Narrative spaces can thus be distinguished through their temporal dimension, a consistent subplot, or the construction of a specific epistemic point of view. I have also explained how narratives achieve coherence through levels of blending leading to an emergent story. However, reading a work of fiction invariably assumes a text-mediated contact with the constructed fictional subjectivity often referred to as ‘the teller’ or ‘the narrator.’1 The concept is textually constructed, which has been stressed repeatedly, primarily in order to avoid simplifications whereby the narrator is identified with the author. At the same time, the illusion is compelling in many narratives, and the reason why the author is so naturally understood to be responsible for the way the story is told has not been answered in sufficient depth. While the ‘death of the author’ (as a viable narrative concept) has been widely publicized, the news may be exaggerated, as some flesh-and-blood authors begin to claim their right to be considered legitimate participants in the narrative exchange.2 Moreover, the ‘demotion’ or backgrounding of the author also avoids any interest in the question that is just as important as the question of the reader response and narrative transport, namely, the ability of a writer to construct a textually available subjectivity. What one might call ‘authorial transport’ is a cognitively unique ability whereby a writer can construct a mind other than his or her own and use it as a medium in making a story accessible to the reader. While the interest in the theory of mind has highlighted the psychological aspects of reader response (Iser 1974; Zunshine 2006), these approaches do not cover the mechanisms of story creation. My tentative suggestion is that creative storytelling requires the construction of very elaborate blends in which actual experience is embodied in a different context, after a process of selection and compression. Selection is necessary in order for the most salient aspects of 58
Stories and their tellers
59
experience (whether physical, historical, or psychological, first-hand or observed) to be used as the conceptual skeleton of the emerging blend. The cognitive ‘padding’ which adds specific character traits, circumstances, events, et cetera is the matter of the emerging narrative spaces. However, the crucial mechanism is that of compression (Fauconnier and Turner 2002; Turner 2006a; Dancygier 2004b). I argued in Chapter 1 that compression is the mechanism whereby specific stories of the oral tradition acquire their generic power to shape behaviors and maintain distributed cognitive networks. The crucial effect emerging out of the compression is also an enhancement of the central themes and increased salience of the intended ‘narrative uptake.’ In the written fictional narratives these tasks are equally important, but are also subject to the author’s choices. Naturally, any narrative selects the facts to be narrated and leaves much of the assumed story untold, but the kinds of selection would depend on the authorial choice of themes. The central compression mechanism, however, is the one whereby the entire story is packaged into a text where only some of the story’s events are told, and the specific compression choices lead to narratives of different kinds. At the very basic level, a fictional story affords a compressed view of the events, of the kind which is typically not available in actual experience. Only a novel gives one access to a picture as complete as is required in a form which is manageable, because it is compressed, and providing the required multitude of viewpoints. The kinds of compressions will affect all aspects of the narrative – time, viewpoint, characterization – but without the compressions the text could not achieve its salience. The processes of selection and compression play an important role in narrative form. In the actual narrative organization they are often subordinated to the choice of some subjectivity as the source of the narration, whichever term is chosen to describe it. While it is clear that such a subjectivity is a narrative construct, the reader has the impression that there is an intentionality and epistemic status behind the discourse, which is typically talked about as a narrator (Gibbs 1994, 1999). The intentionality is crucial in that the very act of storytelling assumes the intention to use the story in its proposed form to communicate some content, even if various forms of narrative experimentation disrupt the impression of consistency and purpose. The epistemic status is equally important as a constitutive feature of ‘narratorship,’ since, regardless of the scope of knowledge the narrator displays (ranging from omniscient, constituting a narrator who knows everything including characters’ thoughts, to unreliable, with a narrator who has a very limited view of the facts), a narrator gives the reader access to crucial narrative facts. In this chapter, I will focus on the aspects of ‘narratorship’ that involve specific configurations of narrative spaces, and also consider the macro-level narrative viewpoint in its relation to the textual profiling of storytellers. As
60
Stories and their tellers
Chapter 2 made clear, narrative spaces as cognitive constructs depend on various kinds of choices, and may not always take the form of narratives, not even skeletal ones. Here, however, I will focus on how various conceptualizations of the story’s teller(s) affect the narrative as a whole. 3.1
Narrators, narrative spaces, and viewpoint
While the focus of this book is not primarily aligned with the interests in narratology, nevertheless the distinctions of types of narration introduced by Genette (1980) and later developed in numerous studies (Rimmon-Kenan 1983; Bal 1985; Simpson 1993; Herman 2002) should briefly be mentioned here. The two questions which seem to be most pertinent to both the original distinctions proposed by Genette and their later expansions seem to seek a positioning of the narrating fictional subjectivity within the story or outside, and aligning it with that of a character or not. While these are indeed crucial distinctions, they fit the framework proposed here only to some degree. First of all, the dual contrast between ‘inside’ or ‘outside’ of the story cannot naturally be maintained in a number of narratives where there are in fact different storyworlds set up, possibly narrated through different types of narrator. Furthermore, as in the case of the ‘Clippings’ space in The Blind Assassin discussed in Chapter 2, there may be no single ‘narrator’ present in any form, while the coherence of the narrative space is achieved through an inexplicit but shared epistemic and moral viewpoint. While it is true that even the most fragmented narrative builds a conceptual structure which I have termed ‘emergent story,’ the boundedness of this construct is difficult to establish in a way which would clearly determine the position of the narrating voice. It seems intuitively convincing to claim that if a narrator is not at the same time a character, then he/she/it must have some knowledgeable position in a different domain, but it is interesting to ask also what that domain is and how the narrator gets access to the ‘story’ if he/she/it is positioned outside of it. I will thus attempt to talk about the ‘type of narration’ or ‘narrator’ as a feature of the narrative space configuration in any given case, and show how the spaces are linked via blending and viewpoint compression. As regards the cases where a narrating voice is aligned with that of a character within the bounds of the storyworld, the status of that voice also depends on the role of the specific character in the text. If the character/narrator tells his or her own story, in the first person, it is still typically the case that he/she relies on facts reported by others as well, and the balance between ‘narrating-the-self’ or ‘narrating-the-story’ is different in different texts. For example, the narrator/ character in Eggers’s autobiographical novel is almost solipsistic in his focus on his own consciousness (though even he relies on others in some fragments of the story), while the narrator/character of Chatwin’s Utz is in fact telling the story of another character, very much as an omniscient narrator would. While
3.1 Narrators, narrative spaces, and viewpoint
61
narratologists have made all the required distinctions, I will assume here that they can be elaborated and clarified in a broader context. The categories to be discussed below emerge out of the understanding of the concept of narrative space, as outlined above. Primarily, I will rely on the concept of narrative space as a source of narrative structure (among others, the features of the so-called storyworld), but, more important, as a source of viewpoint configurations. I will understand narrative space viewpoint as a particular use of the structure of the space. While the narrative may profile a number of characters, and temporal and spatial structures, and involve knowledge of certain facts, et cetera, the specific viewpoint, whether local or global, is a matter of selecting an aspect of narrative structure within the space as foregrounded or specifically profiled. In other words, viewpoint does not necessarily rely on a viewing subject, so that it is not necessarily ‘someone’s’ point of view, but it relies crucially on a selection of an aspect of space topology which provides a filter through which the events are narrated. Viewpoint may thus be spatial or temporal in the basic sense of aligning the narration with a given spatial or temporal dimension of the narrative space. But it can also be epistemic, when access to certain facts, but not others, is assumed (which includes any range of narrating voices, from omniscience to unreliable narrator), or ideological, when the very selection of facts presents the story in a specific light.3 In this approach, trying to come up with a classification of viewpoint types will not lead to a better understanding of the concept, though it may add to an appreciation of its complexity. In fact, narrative spaces, even more than mental spaces, can, and indeed have to, profile a number of possible viewpoint options in order to construct a topology suitable for narrative processing. While the ‘narrating fictional subjectivity’ may represent aspects of viewpoint which remain constant during most of the narrative, there are also local shifts of viewpoint which may affect expressions as small as individual sentences or phrases. As I will try to show in this chapter and the next one, viewpoint is best talked about as a configuration of viewpoint aspects, ranging across narrative spaces involved, and not as a uniform feature of a narrative fragment or the whole text. Even in a basic act of spoken communication one constantly operates in a network of viewpoint dimensions – the speaker’s and the hearer’s (with their epistemic and emotional stances), the time of conversation, the time of events talked about, the spatial locations experienced and mentioned, epistemic or other viewpoints available in the conversational record, in the broad context, or in the artifacts available, et cetera. Choosing to foreground one of these aspects in a construction does not make other viewpoint angles disappear, and shifts are common. If a speaker says in the course of a gossiplike exchange I’m not saying I expect he will propose, but he might, she is manipulating a number of viewpoints in just one sentence, even if we disregard the time and place of the conversation for the sake of clarity. There is her epistemic viewpoint
62
Stories and their tellers
(I expect), the hearer’s viewpoint (you are saying he will propose), the speaker’s rejection of attributed viewpoint (I’m not saying it), the man’s emotional viewpoint (he might be in love enough to propose), the distanced viewpoint marked by the past tense might (instead of may), etc. There is no reason to assume that narratives are any less complex in this respect. However, narratives can be talked about in terms of levels of viewpoint structure, which I will refer to as ‘macro-level’ and ‘micro-level.’ First, there seems to be a crucial difference between the consistent, narrative-wide choice of the way in which the narrating subjectivity is profiled, and the sentence-level shifts and manipulations of viewpoint. While the former emerges out of specific kinds of narrative space configurations, the latter elaborate the topology of individual spaces to give depth to the story. Second, macro-level phenomena reflect the text’s assumed epistemic viewpoint and contribute directly to the construction of the emergent story, while micro-level forms have more to do with the construction of specific events and conceptualizations. The two levels, however, are not independent, as the aggregate effects of micro-level shifts affect the macro level all the time, especially with respect to the epistemic viewpoint of a narrative space. Let me illustrate the interaction between levels with an example, before moving on to the general discussion of macro options (in this chapter) and micro phenomena (in Chapter 4). Example (1) comes from the autobiographical novel by Dave Eggers A Heartbreaking Work of Staggering Genius, where he tells the story of the years when he was bringing up his young brother, after their parents had died. The text follows the inner thoughts of the narrator/main character throughout. In the text preceding (1) he talks about a night spent with a woman, while Toph, the brother, slept at a friend’s house. (1) I wanted to be home in case he came back early . . . made it in time . . . The house was empty, and I dove into bed, fell back asleep, and when he came back home his brother was there, of course had been there the whole time, of course had never left. (AHWOSG, p. 112)
The first sentences, as the rest of the novel, refer to Dave (the sole narrator of the entire text) as I, while he refers to Toph. But when reference to Dave shifts from I to his brother, both levels of viewpoint are affected. At the micro level, the viewpoint no longer focuses on Dave’s intentions (make Toph believe that he spent the night at home), and shifts to the perception attributed to Toph (that indeed Dave never left the house). But it clearly involves the macro level as well, since the expression his brother assumes Toph’s view of Dave’s family role (available throughout the text), and Dave’s view of himself as ‘Toph’s/his’ brother (if the viewpoint totally shifted to Toph, my, not his, would be the required pronoun). The fragment thus manipulates viewpoint to let the reader appreciate both Dave’s intention to mislead, and the fact that Toph was convinced by the
3.1 Narrators, narrative spaces, and viewpoint
63
situation Dave had staged, both of which are relevant to the macro-level narrative, even though the effect is achieved through a minute shift at the micro level. Discussing viewpoint as the specific presentation of the content of a given narrative space makes it possible to acknowledge the reasons why a text may profile several spaces. In the case of The Blind Assassin, macro-level viewpoint varies significantly from space to space, while in other narratives the choice of macro viewpoint may be more consistent. Even though in all cases the narrative space is structured in terms of time, space, and the epistemic or emotional disposition of the teller, et cetera, macro viewpoint foregrounds some of the aspects of space topology and backgrounds others. For example, the consistent first-person narration of Eggers’s novel allows him to focus on the main character’s emotional states and obsessions, rather than on the consistent presentation of events; thus the actual story, even though basically sequential, has to be pieced together in the reader’s mind. In other texts the narration of events may take priority, but then a specific temporal focus will be chosen while other pieces of the timeline will be rendered through flashbacks or omitted altogether. In what follows, I will argue that the macro-level viewpoint is determined by various configurations of three types of spaces. First, there is story-viewpoint (SV) space, the space where the ‘narrator,’ or some features of narratorship, are located. The overarching narrative structure (the primary plot) can be referred to as main narrative (MN) space, which may profile its own narrator(s), but may rely entirely on the SV-space structure. Finally, a character in the MN-space can be selected as the narrator, and we can then talk about Ego-viewpoint, such that the character’s knowledge or intentionality provides the primary focus to the narrative mode. I will consider various examples of these categories below. I will also rely on two additional concepts introduced in cognitive grammar (Langacker 1990a, 1991) to clarify the kinds of possible correlations between construals prompted and the linguistic forms prompting them. While the verb may require a construal including an agent, a selected grammatical form may not profile the agent specifically. In an active voice construction, such as Tom broke the window, both the agent (Tom) and the patient (window) are on-stage participants, explicitly profiled by the form of the sentence; in the passive form The window was broken, however, the agent is off-stage, but nevertheless remains a part of the construal, since the act of breaking requires someone doing it. Various constructions can manipulate the ways in which construals are prompted by expressions, and narrative constructions can be claimed to do the same, with different elements of the space topology on-stage or off-stage for different lengths of time. I will describe some examples of various narrative constructions of tellers in terms of their relationship to the MN-space of the story being constructed. My goal will not be to cover exhaustively all available types, but rather to outline ways in which such constructions can be talked about. I will also attempt to rely on the linguistic correlates of these narratorial constructions, primarily tense and
64
Stories and their tellers
person, and refer to certain general ways in which these categories influence the narrative. For example, the use of past tense as the primary form of narration creates some distance between the narrator and the story, thus often opening more room for various degrees of omniscience. The use of present tense, for comparison, suggests a narration following the events as they develop and more immediate access to the minutiae of the story, but leaves less room for reflection and commentary. Furthermore, the choice of first person reinforces the effect of immediacy and direct access, while the choice of third person is often correlated with some distance and various degrees of ‘omniscience.’ These general observations will be further substantiated through the examples discussed below. In what follows, I will look at some of the common narratological constructions rendered by these choices and discuss examples which relate these concepts to configurations of narrative spaces. 3.2
Types of teller and epistemic viewpoint
The so-called omniscient narrator is a common choice of narrative form – from Henry Fielding’s Tom Jones to most Victorian novels, and all the way to contemporary literature. In many earlier texts, as in Tom Jones, the narrator is also ‘intrusive’ (addresses the reader directly), but as the novel form develops this form is used less and less. ‘Omniscience,’ however, remains a common feature of narrators. Considering that the concept in fact applies to both Henry Fielding’s and Virginia Woolf’s work, it does not seem to be a category salient enough to distinguish narrative techniques with any reliability. What it does refer to, though, seems to be the narrative space set-up whereby the content of all the narrative spaces of the text is accessible from an independent space in which all other spaces are embedded. This ‘outer’ narrative space, which I introduced above as ‘story-viewpoint space’ (SV-space), usually has little or no topology of its own in terms of time, setting, events, or subjectivities, but it houses a vantage point with the rest of the narrative in its scope. Because of the SV-space, the story is presented as if there were a deictic ground for the overall structure. I argue that the deictic elements such as time, speaker/hearer, or location are profiled even if the narrator is not. The vantage point is cognitively available in all cases, but the degree to which such a vantage point is associated with a narrating subject or not accounts for the various choices made.4 3.2.1
On-stage narrators
For example, an intrusive narrator, as in Tom Jones, speaks as follows: (2) Reader, I think proper, before we proceed any farther together, to acquaint thee that I intend to digress, through this whole history, as often as I see occasion, of which I am myself a better judge than any pitiful critic whatever. (Tom Jones, p. 17)
3.2 Types of teller and epistemic viewpoint
SV-space
65
Narrator: ‘I’ Reader: ‘you’
Time: present
Character: he/she
MN-space
Tense: past
Character: he/she
Character: he/she
NS1
NS2
NS3
NS4
Metanarrative space
Figure 3.1 Intrusive narrator (example 2)
Here is an on-stage (fully profiled) narrator who addresses the reader directly and comments on expected reactions to the narrative (see Figure 3.1). The SVspace thus has a certain topology – the narrator is on-stage, the time is later than story-time. All characters are referred to in the third person. One of the results of this set-up is that there is clear deictic distance between the SV-space and the MN-space, including each of the narrative spaces (NS) profiled. While perhaps artificial, this narrative set-up constructs its SV-space as close as possible to the default discourse space, with the subjectivities of the speaker and the hearer, and as such is also the closest to what might be considered the most salient narrative set-up. In fact, the narrator’s mention of critics also sets up a metanarrative space, where the entire novel and its discourse are available for independent evaluation. This kind of set-up is also interesting with respect to its epistemic viewpoint. The source of all the information the story provides is localized in one space. There is no need for the story’s narrative spaces to be evaluated from the epistemic perspective or matched with appropriate subjectivities – all is in the hands and mind of the narrator. This narrative form is not often used now, although it is worth noting that it was also used by Bulgakov (to be discussed below), with a highly humorous effect.
66
Stories and their tellers
The narrator may further attune his/her presence to the needs of the narrative construction. In the first pages of Henry James’s novel The Portrait of a Lady the narrative starts with an “I” narrator, commenting on what he intends to render to the readers. The “I” voice introduces the scene through a fine and extended description of the sunlit lawn where tea is served, but then the narrator’s presence gradually blends away as the characters begin to speak and take their roles in the story. The narrator’s voice in the first paragraphs is only used to give the setting and the atmosphere of the first scene, but the narrator’s presence is thus defined for the duration of the novel. Though this narrative choice has little influence over the text as a whole, it is an interesting confirmation of the need for the SV-space as a separate cognitive construct available.
3.2.2
Off-stage narrators
It is interesting to compare these examples with the more contemporary form of omniscience, as in Anne Tyler’s novel Ladder of Years. The main character, Delia, comes to a job interview, and is waiting in the office: (3) She crossed to the desk, which was bare except for a telephone and a typewriter. She lifted a corner of the typewriter’s grey rubber hood. Manual; not even electric. (She had worried she would find a computer.) She gave a small, testing spin to the swivel chair behind it. Good afternoon, she would say. I’m here to ask if . . . No, not ask. Ask was too tentative. (Ladder of Years, p. 94)
This is a more common type of omniscient narrator, who knows Delia’s spatial and temporal location and follows her actions (she crossed to the desk, lifted a corner of the hood, et cetera). More important, the narrator knows what Delia registers consciously in the environment (bare desk, manual typewriter), and what her thoughts are. The narrator is thus off-stage (does not speak as an independent subjectivity), but all that is said about Delia is relayed through the SV-space, which is located temporally later with respect to the main narrative space. The only glimpse of a narrator’s subjectivity is the comment in parentheses: She had worried she would find a computer. It is less important that the sentence flashes back to Delia’s thoughts before her arrival, but it is crucial that it is an explanation of the reasons why Delia checked the kind of typing equipment she would be using if she were to get the job, and why the phrase not even electric is to be interpreted as an expression of relief rather than dismay. The almost invisible narrator is thus present here and guides the reader in interpreting Delia’s words and actions. In other words, while off-stage, the narrator’s subjectivity is still present as a source of coherence and organization of the story.
3.2 Types of teller and epistemic viewpoint
67
The next example comes from Virginia Woolf’s modernist novel Mrs. Dalloway: (4) But – but – why did she suddenly feel, for no reason that she could discover, desperately unhappy? . . . no, it was not Sally Seton saying that Richard would never be in the Cabinet because he had a second-class brain (it came back to her); no, she did not mind that; nor was it to do with Elizabeth either and Doris Kilman. (Mrs. Dalloway, pp. 120–121)
Here is an ostensibly similar third-person, past-tense narration (which Humphrey [1954] describes as ‘indirect internal monologue’), but the narrator does not say much about the character’s actions and focuses on the flow of thought in the character’s mind, with just a minimal level of commentary – it came back to her. Throughout the novel, narration of events is minimal, and transitions between episodes are primarily transitions from one consciousness to another: (5) “The time, Septimus,” Rezia repeated. “What is the time?” He was talking, he was starting, this man must notice him. He was looking at them. “I will tell you the time,” said Septimus, very slowly, very drowsily, smiling mysteriously at the dead man in the grey suit. As he sat smiling, the quarter struck – the quarter to twelve. And that is being young, Peter Walsh thought as he passed them. (Mrs. Dalloway, p. 70)
At first, the reader is given access to Septimus’s visions, occasionally interrupted by scraps of conversation between Septimus and his wife. But in the last sentence, the viewpoint abruptly shifts to Peter Walsh, and it remains in his thoughts for some paragraphs. These are but a few examples of the way in which a narrative of this type includes much more than the specific viewpoint of a specific character at any given point in the story. There are many Ego-viewpoints represented (Clarissa, Rezia, Septimus, Lady Bruton, Miss Kilman, et cetera) through reports of the characters’ acts and thoughts, but the narrating voice is minimal and the distance between the SV-space and the MN-space is also much reduced. The set-up with Ego-viewpoints is represented in Figure 3.2. While the case of example (3) is different in that one Ego has been selected (Delia), and (4) and (5) represent a multiplicity of parallel Ego-viewpoints, the general narrative space structure is similar. Also, the dotted lines which link characters to the zero narrator represent the fact that the viewpoint of a character becomes the narrative viewpoint for so long as that character’s viewpoint structures the narrative – the entire novel, as in (3), or short fragments, as in (4) and (5). This ‘uploading’ of viewpoint to a higher level constitutes the narrative effect which will be discussed throughout the text – that of viewpoint compression. It is this compression to SV-space which yields the effects wherein the narrator is off-stage and yields the storyviewpoint to a character.
68
Stories and their tellers
SV-space
Narrator: 0 Reader: 0 Time: present
MN-space Character2: he/she Character1: he/she Character3: he/she
NS1
NS2
NS3
Tense: past
NS4
Figure 3.2 Ego-viewpoints
3.2.3
Omniscience and narratorship
There are thus many ways in which the omniscient narrator participates in the story construction. The narrator may be on-stage, and thus be given an independent voice, or off-stage, orchestrating the narrative. But in both cases the subjectivity talked about as a narrator is part of the topology of the SV-space. The third person in all of the instances discussed above is a device setting up a deictic centre of the narrating subjectivity outside of the main narrative space. Also the choice of the past tense serves similar purposes – it distances the main narrative space from the SV-space. The concept of ‘omniscience’ relies on the setting up of a deictic center independent of the story’s main space, which then provides a bird’s-eye view of the structure of the story. The choice of the subjectivity being on- or off-stage has stylistic consequences, but exerts much less influence over the way the story is read. At the same time, the narrating subjectivity, whether intrusive or completely backgrounded, is a source of a variety of blends, involving both the characters and the reader. The intrusive narrator requires that the reader or ‘narratee’ blend him/herself with the position of the addressee in the fictional deictic set-up of the SV-space. The interaction still proceeds only in one direction (from the
3.2 Types of teller and epistemic viewpoint
69
narrator to the reader, but not the other way round), but the reader is given the illusion of a personalized narrative account which addresses the reader’s storyconstruction needs directly and guides the reading process. In a sense, the intrusive narrator style is an interesting (even if now rare) narrative form which openly acknowledges that the story told is not necessarily sufficient for the reader to grasp all that is required. It is also interesting to point out that even the most intrusive of narrators is not restricted to the SV-space as if it were an ivory tower, and may actually signal some degree of involvement in the main narrative space. (6) He now lived, for the most part, retired in the country, with one sister, for whom he had a very tender affection . . . This lady was now somewhat past the age of thirty. (Tom Jones, p. 17)
In (6), the narrator tells a story which is presumably past with respect to the SVspace, and yet occasionally uses the adverb now in the way which is mostly found in free indirect discourse (see Nikiforidou 2010; also Chapter 7). In the case of (6) such constructions are found in narration, and not in represented speech and thought fragments. This is a result of the narrative set-up in which the narrator is explicitly communicating things to the reader, so that narration in this case turns into a form of represented speech, but of the narrator, not of a character.5 The narrator of Tom Jones does not act as a character as well, and yet needs to mark a presence both in the MN-space and in the SV-space – as a witness-like participant in the former and as a narrator in the other. Crucially, the use of now here is referring to the time of storytelling, not the time of the story’s events, so it helps establish the state of affairs in the MN-space with respect to the moment when the narration begins. It negotiates the relationship between the SV-space and the MN-space.6 The omniscient narrator in Tyler’s Ladder of Years is different. The reader is not an overt addressee, the narrator is off-stage, and the narration does not explicitly signal the SV-space. The narrator can naturally blend with the character and report the thoughts (Manual; not even electric.), but also comments on these same thoughts by evoking another moment in the story, whether previously narrated or not (She had worried she would find a computer.). Such narrative moments suggest that the narrator can access any point in the MNspace at any point in the narration. In a way, the narrator’s viewpoint alternately blends with the character-viewpoint in MN-space, or returns to the SV-space viewpoint, and this alternation between spaces constructs a specific form of omniscience. But the blend favors one character, so that the viewpoint alternation affects two stable viewpoints in two spaces. The difference between Tyler’s narrator and that of Mrs Dalloway is mainly in the proportion between the SV-space viewpoint and Ego-viewpoint. While Tyler’s narrator reports various actions and movements, Woolf’s alternates from character
70
Stories and their tellers
to character and intervenes only occasionally. There are no transitions from one Ego-viewpoint to another, or between the inner and outer states of characters, so that narrative spaces are naturally compressed into one flow of the narrative. The primary blend of this kind of narration is compressing the SV-space viewpoint with subsequent Ego-viewpoints in MN-space, and backgrounding the storyviewpoint. To sum up, these ‘omniscient’ blends are varieties of a similar initial set-up, and there are probably other varieties possible. What remains constant is some degree of profiling of the SV-space, while differences result from various degrees of compression of the distance between the narrator viewpoint in SVspace and any specific character viewpoint in the MN-space. The higher the degree of compression, the less narrating the narrator does and the closer the narrative gets to some variety of the ‘stream-of-consciousness’ technique. 3.2.4
Tense, person, and distancing
It seems that the distance between the SV-space and any of the narrative spaces is most saliently marked by the choice of the third person, not of past tense. Some texts, such as Neal Stephenson’s Snow Crash, maintain the third person and shift to the present tense (see Figure 3.3), and the effect may be similar to some of the narration modes described above: (7) The Oldsmobile keeps popping. Jason can’t think of what it is until he realizes that people are shooting at him. Good thing he let his uncle talk him into springing for full armor! When he figures that one out, he actually gets psyched. This is the real thing, man! He’s driving around in his Olds and the bastards are shooting at him, and it just don’t matter. (Snow Crash, p. 147)
The Ego-viewpoint here is maintained for some length of the narrative, and the characters whose viewpoints are thus represented change very often. The use of the present tense signals temporal viewpoint compression up to the SV-space, and the effect is that the narrative now favors keeping record of the on-line character’s thoughts and perceptions (The Oldsmobile keeps popping; This is the real thing, man!). In spite of the present tense, on-line form, the narrator still uses varieties of STR (he realizes that people are shooting at him), and narration, but the consistency with Ego-perspective is achieved differently. Even in the narrated parts (When he figures that one out, he actually gets psyched.) the narrator chooses expressions which would seem more appropriate to the character whose Ego-viewpoint is used here. It is thus an interesting blend of the narrator’s epistemic viewpoint and the character’s discourse style, which makes the subsequent Ego-narrator choices less conspicuous and gives coherence to longer narrative fragments. At the same time, viewpoint compression here affects not only the character viewpoint (as in Figure 3.2), but also temporal viewpoint.
3.2 Types of teller and epistemic viewpoint SV-space
71
Narrator: 0 Reader: 0 Time: present
MN-space Character2: he/she Character1: he/she
Tense: present Character3: he/she
NS1
NS2
NS3
NS4
Figure 3.3 Present time and present tense (example 7)
For comparison, first-person/present-tense narratives (typically not considered ‘omniscient’) go still further in compressing the SV-space and the MNspace, so that in effect the SV-space viewpoint and the MN viewpoint become the same (see Figure 3.4): (8) I open the window, turn up the volume. I pass two cars at once and get on the highway . . . Through the toll, the light, onto a ramp, onto the bridge. Now I can’t turn back. The Oakland shipyards to the left, a billboard encouraging the saving of water. (AHWOSG, Eggers, p. 126)
In (8), the narrator and the character subjectivities and viewpoints are fully blended, but the SV-space does not disappear. The record of events such as opening the window or passing other cars narrates the driver’s actions and measures the passage of time since Dave (the Ego-narrator) left home, independently of his on-line record of perceptions (shipyards, billboard), and thoughts (I can’t turn back). What seems to be the most salient characteristic of this narrative mode is the completeness of the narrator/character blend, so that all parts of the story, including the events which took place prior to the storyline in the MN-space, are told with the same voice.
72
Stories and their tellers SV-space
Narrator: ‘I’ Reader: 0 Time: present
MN-space Main Character: ‘I’ Character2: he/she
Tense: present Character3: he/she
NS1
NS2
NS3
NS4
Figure 3.4 Ego-viewpoint in the present tense (example 8)
There are small parts of the story told in the past, but the difference is that the ‘past’ parts are not intended to be an on-line account of one person’s participation in the events, but are a more ordinary distanced narrative. An example from the same novel is given in (9): (9) When my father was in intensive care, about a day and a half from throwing in the towel, a priest was sent, presumably to administer last rites. After meeting him and ascertaining the purpose of the visit, my father quickly dismissed him, sent him out. (AHWOSG, p. 40)
Characteristically, the style of passage (9) is markedly different from (8), as the events are rendered in distanced form, including past tense, markers of epistemic distance (presumably), passive voice (was sent) and formal vocabulary (ascertaining the purpose of the visit). These brief moments when Dave leaves his frantic on-line persona aside and assumes a narrating role with all seriousness of purpose mark viewpoint shifts, but also reconfigure the type of narrator Dave is and can be. While most of the text is in the frantic present-tense, those composed past-tense parts may be read not just as shifts of temporal perspective, but also as glimpses of a different ‘narratorship.’ While both (8) and (9) belong to the book’s MN-space, their style is adjusted to what is being narrated – past
3.2 Types of teller and epistemic viewpoint
73
family events can be presented in a reflective and serious way, but the ongoing roller-coaster of bringing up a young brother calls for a different narrative style. The multiplicity of smaller spaces set-up is also represented in Figure 3.4 by the smaller circles around more major narrative spaces. The types of narration sketched above are distinguished, via constructional means, as reflecting a specific relationship between the SV-space and the MNspace. The compressions along the lines of temporal distance (past or present) or the distance between a deictically determined communicator in the SVspace and a character in the MN-space (first or third person) yield several types, even though they are not always clearly distinguishable. The important point is that in most cases the specific viewpoint configuration remains as the default set-up throughout a novel, so that various specific parts of the narrative structure are subjected to the specific SV/MN relationship. For example, all the numerous Ego-viewpoint blends in Mrs. Dalloway are naturally assumed to be subjected to the MN-space of the day of Clarissa’s party. Similarly, the narrative of Snow Crash blurs the Ego/narrator boundaries through consistent stylistic form. Also Eggers’s narrative, which is less stable in its viewpoint setup, remains within the bounds of an Ego-narration. Features of style, however, contribute significantly to how the narrative space set-up will be framed. Eggers’s narration is scattered and shifts narrative-space-viewpoint all the time, and thus, by comparison, the characters in Mrs. Dalloway seem to ‘think’ in voices too similar and stylistically colorless to represent abrupt viewpoint shifts, even though the Ego-viewpoint changes all the time. The stylistic dimension (choice of register, degree of distance) thus seems to create the quality of narration which obliterates or highlights the viewpoints profiled. The smoother the transition from narrator to Ego, the more cohesive the narrative seems. The scattered quality of Eggers’s story persists even when the Egoviewpoint is maintained. Example (10),7 which follows (8) immediately, without even starting a new paragraph, has to be read from the point of view established in (8) (Dave, driving to San Francisco, after he left his brother with a babysitter): (10) I will come home and the door will be open, wide. The baby-sitter will be gone and there will be silence . . . Blood on the walls, handprints soaked in blood . . . There will be a hearing, a trial, a show trial – How did you come to meet this man, this baby-sitter? We found a posting, on a bulletin board. And how long did your interview of him take? Ten, twenty minutes. (AHWOSG, p. 126)
Most readers would agree that Dave’s worries about his brother are exaggerated, so that he could be viewed as an ‘unreliable’ narrator. The concept was
74
Stories and their tellers
introduced by Booth (1961), and was recently discussed by Semino (2007) as related to what stylisticians refer to as ‘mind style.’ Whichever term is used, the point seems to be that some narratives are relayed through an Ego who is either unaware of or unable to grasp the nature of the events (for example, Benjy in Faulkner’s The Sound and the Fury). As Semino rightly points out, there are a variety of ways in which the specificity of the constructed mind influences the narrated content. The so-called unreliable narrator is reliable enough for the reader to construct the story, but the narrative viewpoint is indeed expressed somewhat differently. Primarily, unreliable narrators are presented as only capable of a very specific epistemic viewpoint, so that the narrowness of the understanding of events strikes the reader as inaccurate. Whether such a narrator is accessible through the first-person or third-person style, other viewpoints are not directly available. It seems that such cases constitute another blend, such that the SV-space is entirely blended with the MN-space relayed through one specific Ego. Such a narrative space naturally calls for the reader to construct alternative viewpoints, prompted only by the bizarre consequences of following the viewpoint available. In the case of (10), the readers understand that a parent would most likely not be put on trial for leaving a child with a babysitter, but does it make Dave ‘unreliable’? He is, after all, narrating his obsessions and worries, which are a crucial part of the story told. It seems that the viewpoint is entirely subjected to the goals of the narrative, and such goals may not include giving a reliable account of events. As the examples above suggest, the goals are more often focused on representing the character’s epistemic and emotional perspective. The examples above all point to the specificity of distance in narrative discourse. Dancygier and Vandelanotte (2009) and Vandelanotte (2010) discuss the varieties of distancing devices in literary texts – specific forms of represented speech and thought, temporal distance, metalinguistic distance, epistemic and emotional distance – mostly available through textual access to more than one viewpoint-marked narrative space. In the fragments considered above, the concept of narratorship itself seems to emerge from the varieties of temporal and epistemic distance between the major spaces involved (SVspace, MN-space, Ego-space). While the temporal distance (e.g., present SV/past MN) and epistemic distance (SV access to story/Ego access to story) seem to be the most common devices, the stylistic features marking emotional, evidential or social distance contribute substantially to the impact of the text, including perception of viewpoint. Ironically, Mrs. Dalloway may seem more coherent, in spite of its galaxy of narrating Egos, while Eggers’s text feels scattered, even though there is only one Ego narrator, because its emotional temperature keeps jumping up and down and his mind wanders in most unlikely directions. Overall, the viewpoint texture of a narrative as a whole depends as much on the general space set-up as on the stylistic profiling
3.2 Types of teller and epistemic viewpoint
75
of various aspects of viewpoint in those spaces and the foregrounding or backgrounding of various types of distance. 3.2.5
Constructional compositionality
Viewing the narrative options described above as constructions highlights important facts about interpretive consequences. There seem to be four constructional parameters involved: narrator on- or off-stage, person, tense, and the profiling of Ego-viewpoint (or several). All these parameters represent interactions between the SV-space and the topology of the MN-space, suggesting various degrees of distance and compression. The degree of profiling of the SV-space correlates with the degree of distance between the SV-space and the MN-space. If the narrator is an onstage part of the SV-space, the degree of compression with the MN-space is rather low, as the storyworld and the character are not seen as participating in the narration. For comparison, an off-stage narrator invites less distance and more compression, especially with many Ego-viewpoints profiled. The yielding of narrative voice to Egos (as in the fragment from Woolf) downplays the role of the SV-space, reducing the distance to the temporal dimension, and yielding a narrative where the character perspective dominates and the organizational role of the narrator is not foregrounded. Intrusive narration and multi-Ego narration are thus two extremes along the spectrum of narratorial involvement. In this context, the choice of first-person or third-person narration is also dictated by the degree of distance involved – with first-person narration being fully compressed into the SV-level (so that the independent narrating voice is not construed by the set-up), and the third person giving the narrator opportunities to intervene and comment. The choice of tense is less consequential. I have assumed that the SV-space is to be treated as ‘present,’ since the narrating is always done in the narration’s present. This yields a rather straightforward set-up in the case of narration in past tense, as it only underscores the distance between the SV-space and the MN-space. Narration in the present tense, however, does not suggest that the time of the story’s events is the same as the time of the story-telling. Rather, it diminishes the distance between the spaces, and allows the story’s time to be compressed into the SV-space – thus, even if the SV-time is present time, it does not mean it is the same as the story’s present tense, which will still be past time, wih respect to SV. There is still conceptual distance between the SV- and MN-spaces. This brief overview suggests that the actual form of any given narrative emerges out of a combination of these constructional parameters, and is primarily defined by the degree of remaining distance between SV- and MNspaces. Crucially, not many narratives are entirely consistent in the choice of the narrative space configuration, and such shifts can be explained only against
76
Stories and their tellers
the template of the SV-space. The SV-space thus needs to remain separate from other narrative spaces, so that the overall coherence of the story can be established at some stable narrative level. 3.3
Second-person narratives
Some of the issues raised above can further be clarified with a look at the less common form of narratives – with second-person pronoun as the primary SVspace-related choice. These instances are related to the SV set-up described above, but exploit it in different ways. In fact, there are more significant differences across examples than, let’s say, among third-person narratives. One of the texts where the second person dominates is Oriana Fallaci’s novel The Man, describing the life of Alekos Panagoulis, her close friend. She dedicates the book to him (For you), then, in the introduction, promises to tell your tale and describes Panagoulis as my only possible interlocutor. The story follows, with Panagoulis consistently referred to as you. (11) You thought about the amnesty all night, sometimes believing it, sometimes not, and when you didn’t believe it, you felt serene, when you did believe it, your conscience was split in two. (The Man, p. 132)
Interestingly, the second-person form is used here to represent the inner thought of the main character, so that the text contains all the narrative elements the reader expects. There is also an I in the novel, referring to Fallaci herself, also framed as the narrator: (12) As if stung by a wasp I stood up. I said I had to leave you and go find a hotel. You answered categorically: “You’re not going anywhere. You stay here.” (The Man, p. 150)
The novel thus explicitly sets up a deictic frame, with the writer as an equivalent of the speaker, and the main character as the addressee (see Figure 3.5). The addressee position here is not that of the reader or narratee, who remains a standard recipient of what the text communicates. To accommodate the you, we need an extra space of the reality, in which Fallaci and Panagoulis were friends, and we need to establish a cross-mapping between the two of them as people and as narrator/character and character/addressee respectively. The novel’s set-up thus explicitly comments on reality and fiction, while also blending the two grounds – the imagined deictic ground wherein Pangoulis is an addressee, and the novel’s narrative ground, where the addressee is the generic reader. The set-up is different in another novel which relies heavily on you – Calvino’s If on a Winter’s Night a Traveller. The structure of the text is very interesting, with the novel itself partially embedded in it, and with various excursions into other (presumably nonexistent) novels. You in the text refers
3.3 Second-person narratives SV-space
77
Discourse space Speaker: Oriana Hearer: Alekos
Narrator: ‘I’ Reader: 0 Time: present
Reality-space Oriana Alekos MN-space
Tense: past Character: Oriana
Character: Alekos NS1
NS2
NS3
NS4
Figure 3.5 Second-person narrative (examples 11 and 12)
to the reader, but the reader is present not only in the SV-space, as is usually the case, but also as a character whose experience of reading and interacting with the text is the primary theme of the novel: (13) The novel you are reading wants to present to you a corporeal world, thick, detailed. Immersed in your reading, you move the paper knife mechanically in the depth of the volume. (If on a Winter’s Night a Traveller, p. 42)
The present tense and the progressive aspect in (13) also suggest that the intended construal is that of the ongoing process of reading (with the exception of fragments from embedded imaginary novels, which are differently framed). The MN-space is initially structured as taking place in a railway station, with the reader as one of the characters and the narrator as the other. The text also implies that the scene is developing within the reader’s field of vision: (14) I am the man who comes and goes between the bar and the telephone booth. Or, rather: that man is called “I” and you know nothing else about him . . . (If on a Winter’s Night a Traveller, p. 11)
However, the reader’s observations are relayed through the narrator’s voice, as in (15). (15) It is obvious that I am a subordinate, I do not seem the sort of man who is travelling for personal reasons or who is in business for himself. (If on a Winter’s Night a Traveller, p. 14)
78
Stories and their tellers
SV-space
Discourse space Speaker: Narrator Hearer: Reader
Tense: past
MN-space
Character-Reader: ‘you’
NS1
Narrator: ‘I’ Reader: 0 Time: present
NS2
Character-Narrator: ‘I’
NS3
NS4
Figure 3.6 Second-person narrative (examples 13–15)
Similarly to The Man, the novel profiles the narrator and the addressee as characters, but neither of them is correlated with reality, so that the crossmappings reach across MN-space and SV-space (see Figure 3.6). The two characters are thus compressed up to the SV-space, but the narrator retains his ability to reach into the mind of the character, even when the character’s thoughts are about the narrator as character. The identity of the I character is not developed to include an independent subjectivity, and he is reduced to his narratorial function and the perception of the reader-character. Second-person narratives (cf. Fludernik 1994) thus rely on mechanisms described above for first- and third-person texts. The presence of the narrator and reader slots in the SV-space is exploited for cross-mappings. However, while the reader is only profiled in the text in intrusive narration, in the examples above other cross-mappings are set up – with a character in the MN-space, or with an addressee in the reality space. Furthermore, the viewpoint configuration such as in (14) and (15) includes a blend wherein the I-narrator does not report his or her own story, but another character’s, cross-mapped with the reader. Finally, such a set-up does not invalidate the added layer of narrative space structure wherein there is an actual reader reading Calvino’s novel.
3.4 The teller, the author, and the character
79
This discussion has not mentioned the narrative use of you which is very similar to the colloquial discourse usage – the generic you, as found in suggestions, how-to descriptions and other contexts where a general pattern of behavior is being described. Because this generic form is not specific to narratives, I will not discuss it any further. Thus far, I have tried to trace some of the constructional correlates of general viewpoint mechanisms in a fictional story. The tense and pronoun choices discussed here are specifically those which can be naturally explained in the overall narrative space set-up. There are also more local, micro-level pronominal choices, sometimes affecting the concept of narratorship, but these will be discussed in Chapter 5. At the same time, there are also text-wide consequences of how the teller of the story is construed, especially with respect to its author, as tellers can be discussed from the point of view of the overall coherence of the text. The question of how the conceptualization of the teller affects the story will be considered in the remainder of this chapter. 3.4
The teller, the author, and the character
Recent work by J. M. Coetzee explicitly addresses the complexities of the role of the teller in the understanding of a text (see Dancygier [2010] for broader discussion). Coetzee himself is an author of fiction as well as nonfiction, and his recent novel, Elizabeth Costello, combines the two. In the text, he includes various articles and lectures he wrote or delivered as J. M. Coetzee and attributes them to a fictional author, his alter ego Elizabeth Costello. The resulting blend represents authorship from two perspectives: on the one hand there is the subjectivity named on the cover, who only speaks through the authored text, while on the other hand there is the actual person, having reasons to write what she does. Costello is different from any other author we know only from a vignette on the dustjacket in that she also represents the desire to be understood and heard, strong personal beliefs, and tiredness with fame. It is natural to read her as a blend of the fictional subjectivity and that of the actual author. In the next novel, Slow Man, Costello appears again. She walks into the apartment (and life) of a character, Paul, who loses his will to act after a crippling accident. This narrative construction, whereby there is an author inside the narrative space, gives a view of authorship which questions a number of default assumptions. While other characters in the novel as well as its critics suggest that Costello’s interventions are those of an author writing the novel who wants to put Paul in her book, this seems to be an oversimplification. True, Costello tries to prompt Paul to action, but not in order to use the events as narrative material, but rather in order for the narrative to ‘take place.’ As she (or Coetzee) sees it, it is the character’s responsibility to ‘create’ the story by living it. The story of Slow Man does not exist in Costello’s (or anybody else’s) mind,
80
Stories and their tellers
and the reader thus reads a novel as existing in a narrative space independent of the author’s or narrator’s or character’s subjectivities, but driven directly by their agency as characters and writers. The presence of Costello as an embodied narrative choice (to live a life worth writing about and let the novel emerge from it) constantly reminds Paul that a different story could be lived and told, but if we recall that she is in fact Coetzee’s alter ego, the choices she offers are choices Coetzee might be associated with. In effect, the character is not ‘put in a book,’ but the author is. In the next novel, A Diary of a Bad Year, Coetzee’s next alter ego, Señor C., speaks more openly of his role and claims his right not to be seen as ‘merely’ a set of rhetorical tricks. Coetzee himself is a critic, but in his own case, he speaks through his novels on his own authorial viewpoint and his understanding of what fiction is. He seems to find the pronouncement of the author’s death (even if it is meant just to be provocative) to be an insult to his personal and intellectual involvement in the process of writing and, while he is willing to yield most of the power to the character and the logic of the story, he does not like the idea of giving it all to the reader. To sum up, Coetzee’s recent work suggests a re-evaluation of the narrative which goes in two directions. On the one hand, fiction needs to be freed from the overpowering subjectivity of a teller, and given its own place as a representation of a story. On the other hand, stories and characters need their authors to take shape and be convincing. While the author cannot write a story that a character refuses to live, life is full of stories which are lived in a way worthy of an author’s attention. The viewpoints which thus emerge are all subjected to the viewpoint of the story as such. 3.5
Multiple tellers
The question of narratorship and the cognitive status of a teller becomes particularly interesting in the cases where the text as a whole profiles more than one teller, while ostensibly telling the same story. The case of narratives which rely on various focalizers is not the one under scrutiny here; instead, I will focus on cases where the text designates different characters or different identifiable narrators to tell different partial stories. I have discussed one such case in an earlier work (Dancygier 2008a), Jan Potocki’s The Manuscript Found in Saragossa. The text has an interesting history, and a still more unusual composition. It was written in French by Jan Potocki, a Polish aristocrat living and writing in the late eighteenth and early nineteenth century, whose life story could make a fascinating novel of its own. The text was first published in two different parts, so that even reconstituting the entire book in its intended form was a historical and critical challenge. The ‘novel’ consists of 165 stories (or so one story-counter claims), connected in different
3.5 Multiple tellers
81
ways through embedding and digression, told by a number of different tellers, and yet, in some way, constituting a coherent narrative of the story of the main character, Alphonse van Worden, and of the Moorish family of the Gomelez, to whom he is related through his mother. The intricacies of the narrative in this case challenge the reader in many ways, as even keeping track of which level of embedding a given story belongs to is not a straightforward task. As I argued elsewhere (2008a), the book does form a coherent whole not because the stories naturally form one overarching storyline (they don’t), but because various stories contribute to the establishment and maintenance of viewpoints. In other words, tellers are not necessarily responsible for a piece of narrative structure, but their stories represent ways of interpreting the overall story being told, and thus the whole is not only a fascinating narrative, but also a reflection on honor, religion, tolerance, the value of wealth, as well as mathematics, philosophy, and history. One of the most salient questions is the choice between religion and faith on the one hand (this is the time of the Inquisition!) and honor on the other hand, as principles guiding a moral man. Indeed, these are the choices the main character, Alphonse, faces throughout the text, but the options are presented through different stories, told by different people in different places and times. Thus there are a number of stories whose conclusions suggest that faith is the best foundation, while others could be read as mocking religious devotion, and still others reinforce the idea that honor is the best guide. Within his own story, Alphonse chooses honor, and gains respect as well as riches. The power of the Manuscript, then, is in the setting up, elaboration, and reinforcement of viewpoints, while their continued presence and salience in the text are constructed through a variety of smaller narratives, none of which can sustain the subtlety and multidimensionality of the moral and philosophical debate. Thus the number of ‘tellers’ and narrative contexts is the source of variety and interest, but the main viewpoints guiding the story and its main character are established over and above the various narrators, through the construction and selection of the stories themselves. There are also cases where the story profiles more than one teller but in the end remains unspecific as to the subjectivity which the story is to be attributed to. There are numerous examples of this in contemporary literature, but I will discuss one in detail here – Mikhail Bulgakov’s The Master and Margarita (M&M) – to propose also an analysis of how the misleading structure fits the viewpoints represented. The novel has two major narrative spaces: the story taking place in Moscow, in the late 1920s, told by an omniscient intrusive narrator (M&M1), and a novel about Pontius Pilate and a wandering philosopher Yeshua who is crucified at the end (PP&Y). The narrator of the PP&Y novel is also omniscient. M&M1 is a story of frightening, but also hilarious events which occur in Moscow as a result of the visit by Satan (traveling as Mr. Woland) and his team. Among other
82
Stories and their tellers
things, Woland organizes a ball where all the sinners and souls committed to eternal damnation have a day’s break. He needs a lady to accompany him, and an appropriate woman is soon found. Her name is Margarita. The PP&Y novel was written by a man living in Moscow, who only appears in the text as the Master (his name is unknown). In the novel, the story involving Pontius Pilate, Jesus (here Yeshua), and Judas, and leading to the Crucifixion is told in an entirely realist mode, with no apparent allusions to the religious context of the story. The writing of the novel and the reasons why it was never published (in fact, the manuscript was burned) form one of the narrative themes in M&M1, where the Master is one of the characters. From the narrative perspective, the most interesting aspect of the PP&Y novel is how it is given to the reader. The first installment is told by Woland (Satan) to two incredulous men – the critic Berlioz and the poet Ivan Bezdomny. However, it is presented as an eyewitness account – Woland describes what he saw and heard in Pilate’s palace. Ivan the poet is shocked by this encounter and eventually ends up in a lunatic asylum, where the next fragment of the novel appears to him in a dream. Finally, Margarita recovers the entire manuscript (Satan gives it back to her, adding that ‘manuscripts don’t burn’) and reads it (throughout the remainder of the M&M1 story), which is how we get to the end of the text. The last words, as announced earlier by Margarita, are “procurator of Judea, the equestrian Pontius Pilate.” The text of the PP&Y novel is thus presented at once as a novel on a subject entirely divorced from the reality of Moscow in the post-revolution years, and as an account of Pilate’s encounter with Jesus and of the crucifixion, which is fictional (it is a novel), but also factual (eyewitness report). On yet another level, Ivan and the Master are both insane, and Margarita, by usual standards, loses her sanity as well – so perhaps these are just ravings of lunatics – but then, how come they give a very sane account of events which are typically seen in more otherworldly terms? The choice of narrators is the only source of such questions. The search for authorship within the storyworld is equally elusive. PP&Y is the novel written by the Master (even though it exists beyond the manuscript or the author’s consciousness). But in the final scenes the Master is allowed to free Pontius Pilate and is thus still responsible for his characters and has the power to change their fate – as we expect a true author to do. But now he not only ‘writes them,’ he can change their lives in the reality the novel represents, which is now a part, though an invisible part, of the reality as we know it. Also the authorship of M&M in its entirety becomes questionable. The final words of the Master’s novel, “procurator of Judea, the equestrian Pontius Pilate,” also end the story of Ivan the poet, and the entire book. The words become a powerful narrative anchor which marks the limits of authorship as based on magical insight into the truth, not on the power to create fiction.
3.5 Multiple tellers
83
In this sense, the Master, the poet, and the actual writer Bulgakov are all candidates for authorship-as-insight, while the Master and Bulgakov are blended as possibly actual writers of the words. Indeed, there is much in Bulgakov’s biography that matches the misfortunes of politicized criticism which destroy the Master, but, first of all, the final words seem to be the ultimate wink of this self-ironic and at times amusing, at times deeply moving novel. It does not really matter who wrote what, just in case some eager critic will want to use it against the writer. The way in which Bulgakov plays with the concept of authorship and with the boundary between fact and fiction is of course deliberate. The truth or fictionality of the story of the Crucifixion is crucial to the worldviews which are competing in the historical context throughout Bulgakov’s life. The political impact of Bulgakov’s text is beyond the interest of this analysis, but there are important narrative lessons to be learned. There are things authors can talk about which are just for the enjoyment of their readers, but there are also things in socalled-fiction which are deeply true, even though there is no proof of their truth in any historical document. It is the truth not owned by the author, but perhaps shown to him in a moment of reflection, or in a dream. The blurring of all crucial narrative categories in this important novel is deliberate, and it is achieved through manipulation of narrative structure worthy of a true Master. From the point of view of the emergent story, the case of M&M is also interesting as a result of its choice of tellers. The two primary narrative spaces (M&M1 and PP&Y), which start out as separate (the ‘Woland-in-Moscow’ story and the Pilate/Yeshua story) go through interesting changes of status as the text develops. Initially, they are linked through Woland, who is a character in M&M1 and the first teller of the PP&Y part. Later, the link is further developed through Ivan’s dream. Then Ivan meets the Master and the Master’s novel first enters the narrative scene. The fact that the reader has thus far had access to PP&Y through major characters in M&M1, and can only then reconstrue the story as the Master’s novel has crucial importance for the narrative structure. The progression is from eyewitness report to a dream, to a novel, and thus the truth-or-fiction issue is set up from the start as impossible to resolve. Having different narrative spaces as homes for construals of the Crucifixion story makes it possible for these spaces to maintain three different viewpoints throughout the text, even though the sequence of events told is not disrupted at all – Woland starts the story, Ivan’s dream continues it, and then the Master’s novel takes care of the rest. This example supports the suggestion made earlier that the sequentiality of the emergent story is of secondary importance to the meaning of the narrative, while available viewpoints shape the emergent structure in substantial ways. To put it simply, if the Crucifixion story is true, then Satan exists and can interfere in contemporary affairs, while Ivan and the Master are sane and have been given access to the truth. If it is a delusion or fiction, then we have read a
84
Stories and their tellers
funny fantasy story. This choice is established textually at the beginning, since the first scene of the novel has Ivan and the critic, Berlioz, discussing the question of whether Jesus really existed or not. The whole conversation can thus be viewed as a narrative way of establishing the two viewpoints which drive the ensuing narrative – similarly to Potocki’s text. The viewpoints, not the facts told, decide on how this book is read and understood, and both viewpoints remain active till the very end. 3.6
Narrative space embedding
Thus far, I have looked at instances where the SV-space either has little structure or is only a source of temporal and epistemic viewpoint, but plays little role in the story. However, there are also cases of straightforward space embedding, where the SV-space contributes elements of the story and often changes the emergent story significantly.8 The role of embedding in emergent narrative structure can be illustrated with the narrative space configuration of Ian McEwan’s novel Atonement. The text, a typical third-person/past-tense omniscient narrative, consists of three parts. It starts when the main character, thirteen-year-old Briony Tallis, writes a play, The Trials of Arabella, a rhymed story full of kitsch emotions and childish ideas of love. Although the play does not participate in the story proper, it provides an opportunity to characterize Briony as an aspiring writer, a person prone to using big words, and treating fiction as seriously as she treats life. The play foreshadows the coming events, where Briony witnesses sexual encounters between her sister, Cecilia, and a man named Robbie. She interprets these events as Arabella probably would have, as signaling brutality rather than romance. After another character helpfully provides the term ‘maniac,’ Briony has her vocabulary to describe Robbie, and when another girl is raped, she blames it on him. This part of the narrative clearly defines Briony’s over-excited confusion of the adult world around her with the fiction she writes. In a sense, Briony tells two fictional stories – of Arabella and of the rape, both structured with the same authorial disposition. The story develops dramatically. Robbie is sentenced, then freed to join the army, while both sisters take part in the war effort as nurses. Later in the war, Robbie returns to England, and Briony meets with him and Cecilia to acknowledge that her accusations were wrong. This ends the main narrative space of the novel, and the text is signed “BT, London 1999.” This brief addition, however inconspicuous, changes the entire viewpoint structure of the novel. The narrator is no longer a creation of the (implied?) author, McEwan, but the main character, Briony. The novel we read is not the entire story constructed by McEwan, but only its part. And we also learn why the text had to be written, as the last words are: She knew what was required of her. Not simply a letter, but a new draft, an
3.6 Narrative space embedding
85
atonement, and she was ready to begin. A “new draft” of what? This narrative anchor, as well as that of the date, are crucial to the final shape of the emergent story. The fourth part of the novel, titled London, 1999, mentions various drafts of the novel (1940, 1947, 1999) and acknowledges that Briony has been changing the facts from draft to draft until she finally decided to tell the whole truth of her crime, but not say what really happened to Cecilia and Robbie. She suggests a possibility that they died in the war and never saw each other after the police took Robbie away, but we can only guess that that was what ‘really’ happened. Throughout the majority of the book the text constructs a viewpoint of a narrator who knows what happened and does not judge anyone, but in its last twenty pages it changes everything. An additional narrative space is set up where the true author writes a story which will count as ‘the truth,’ including the name of the real rapist, now married to the girl he raped. It is in fact an explicit construction of an SV-space. The ‘embedded novel’ was developing throughout the first 350 pages as an ordinary set-up with a narrator off-stage, and a temporal and epistemic viewpoint yielding a third-person/past-tense narrative. But the final revelations add a lot of topology to the SV-space, which now has a firstperson/present-tense narrator who also identifies herself as the author of the story told in the MN-space. The MN-space thus becomes an embedded space, in the sense typically used in mental spaces theory, which means, among other things, that the stance and viewpoint of the higher SV-space are projected into the lower MN-space. This means that the topology of the ‘embedded novel’ space now inherits the relevant bits of the topology of the SV-space, so that, among others, there is a cross-space connection linking Briony-the-character to Briony-the-author. But the epistemic viewpoint of the MN-space had to change, in that we now know that the ‘true’ story resides in the SV-space, along with all the previous drafts where different names were used and different circumstances were described. The new set-up also crucially changes the intended sense of atonement, which now has three different meanings. In the embedded MN-space, atonement for Briony’s guilt is achieved by clearing Robbie’s name and accepting the blame. The novel itself, as a text to be read, counts as atoning for the crime by changing its result and giving Robbie and Cecilia a life together. In effect, for a reader in the space in which the embedded novel exists as an artifact, there is a happy ending. For the author, however, atonement is only to be sought. The various drafts she alludes to, with different versions of what happened, are attempts, and she finally settles for the version which is the closest to the truth, but changes the ending – and that is how atonement might be achieved. To conclude, the establishment of the SV-space is consistently used to structure the story’s general viewpoint and to put the MN-space story in perspective, or even change its structure. Regardless of whether the set-up is
86
Stories and their tellers
textually obvious, or whether the embedding space is smoothly blended with the embedded space, as in the Tyler or Woolf examples, the narratives discussed here rely on the embedded set-up and its effects – primarily on the projection of SV-space viewpoint into the MN-space.
3.7
Narrative viewpoint and narrative spaces
What the examples above suggest is that one might consider narrative viewpoint in terms independent of the subjectivity of a specific teller, but instead as a feature of the narrative space. Major narrative spaces, being mental constructs and basic units of narrative structure, are equipped with rich topology, including spatial and temporal grounding, the record of events and/or conversations, and, first of all, the characters. In the various cases of profiled narrators, as described in Section 3.2, the viewpoint is at least partially relayed through the fictional subjectivity of the narrator. The need for the ‘narrator’ construct, whether independent of the characters in the space or blended with one of them, emerges out of the default communicative set-up, whereby content communicated should be attributed to a communicator. But the presence of the construct does not do much more than provide a focal point which gives the reader access to the topology of the space. What we have seen in the texts briefly described above is that narrative viewpoint can be complicated by creative manipulation of spaces, rather than reconstrual of narrators. There may be additional spaces set up, or viewpoint may be spread over a number of different narrative spaces and tellers, as in Potocki’s novel, or the space topology and epistemic viewpoint may alternatively match one character or another or remain shared across two subjectivities. Finally, as in Bulgakov’s text, distributing narratorship across different subjectivities can contribute to the establishment of multidimensional or competing viewpoints. In none of these cases is the viewpoint uniformly attached to a subjectivity of a narrator; rather, it is a characteristic of the narrative space. While it is perhaps more typical for narratives to profile tellers, they remain a structural solution to a conceptual problem – how does the reader get access to the facts s/he needs in order to comprehend the story and see its broader implications. At the level of the reader’s construal of the emergent story, the selection of an aspect of narrative topology to serve as the focal point is a way to direct the story-construction processes.
4
Viewpoint: representation and compression
There were two men in my father’s chair. (Coasting, Jonathan Raban)
The preceding chapters have outlined ways in which narrative viewpoint is construed and linguistically expressed at the level of major narrative spaces of a story. However, viewpoint phenomena of more partial nature pervade narrative discourse at every level. There are several terminological issues which require clarification at this point. ‘Viewpoint’ and ‘perspective’ are the broadest terms in use, with respect to both macro-level phenomena (such as narratorship, temporal and spatial anchoring of the narrative, et cetera) and micro-level phenomena, which are often talked about in terms of Genette’s (1980) concept of ‘focalization,’ or ‘who sees.’1 Much of this terminology evokes visual access to situations, though the nature of focalization may include other ways in which the narrative is aligned with a fictional subjectivity, whether of a narrator or of a character. To avoid the visual implications, Toolan (2001) proposes to talk about ‘orientation’ instead, which in turn evokes spatial configuration of some kind. All three terms, ‘focalization,’ ‘orientation,’ and ‘viewpoint,’ in fact cover a broad range of phenomena having to do with the specific use of narrative space topology – temporal, spatial, perceptual, and the like. Much of the relevant discussion is also complicated by the assumption that there is a difference in narrative discourse between stylistic forms which ‘tell’ (such as narration) and those which ‘show’ (those which are assumed to represent, rather than describe) (cf. Booth 1961; Rimmon-Kenan 1983). The distinction, while useful in many ways, calls for a clarification of other concepts. Narration, in this sense, is the domain of the narrator, who (which?) is a narrative construct only – so if ‘telling’ the story is what is indeed talked about, this involves rather strong assumptions at play about a narrating subjectivity or a ‘teller.’ In fact, ‘telling’ seems to be a more complex concept. In Chapter 3, I suggested that there are three sources of the conceptual set-up underlying textually determined construction of the narrator’s ability to tell the story: the viewpoint configuration involving the SV-space and the MN-space, the reliance on the default communicative set-up, and the choice of grammatical forms such as tense or person. In 87
88
Viewpoint: representation and compression
such a set-up, the ‘showing’ would also be the domain of the narrator, except that the ‘voice’ is temporarily given over to a character. But how can a character have a ‘voice’ independent of the narrative choice of the specific situation and wording? The so-called direct speech in narrative fiction should not be taken as a token of plausible conversational discourse, as any discourse analyst knows very well. Natural discourse is full of interruptions, hesitations, overlaps between speakers, unfinished sentences, discourse markers, et cetera, in sharp distinction to the narrative representation of discourse. The idealization does not consist only in smoothing out the performance, but also in very careful choice of which exchanges between characters will be represented and how. Direct discourse in fiction is as ‘narrative’ as narration or free indirect discourse, but it is designed to ‘tell’ (not really ‘show’) the story through a different conceptual medium (for further discussion of direct discourse see Chapter 7). ‘Viewpoint’ as a feature of a narrative space, rather than of a fictional mind, is also a generic concept which underlies much of the discussion of represented speech and thought, especially forms such as ‘free indirect discourse,’ which have been claimed to represent a ‘dual voice’ (Fludernik 1993). One of the ‘voices’ then would be that of a narrator, and the other of the character, but both types of fictional minds are primarily a source of viewpoint, rather than discourse as such. The difference between a fictional mind which is ‘seeing’ and one which is ‘speaking’ may be partly tenable with respect to represented speech, which would then be an instance of ‘speaking,’ but does not seem justified in application to thought, given that the unspoken content of characters’ minds definitely involves those characters’ viewpoints (or ‘seeing’) and is to a much lesser degree aligned with specific forms of expression. The forms of speech and thought representation (STR) are usually considered to be an independent narrative phenomenon, and are definitely distinguishable as linguistic constructions (see Vandelanotte 2005, 2009 for a very thorough analysis), but they do assume a viewpoint of a speaker or thinker. What is more, the indirect types of discourse described within STR are often (misleadingly) talked about as representing actual words or thoughts, while they can at best be claimed to be narrative-specific representations of their content.2 Yet another question is the distinction between ‘speech’ and ‘thought’ in STR, although in most cases both are assumed to be represented by almost identical constructions, whether described as direct or as indirect.3 In natural language use, the distinction is crucial, in terms of addressivity, accessibility, and the specificity of propositional content. In narrative fiction, for comparison, the distinction is almost nonexistent, and I will discuss more specific exmples in Chapter 7. Generally, I will assume that both have the role of representing conceptualizations, with little need to distinguish sharply between communicated and noncommunicated ones, since the overarching goal is to communicate them both to the reader.
Viewpoint: representation and compression
89
It seems that the treatment of ‘thought’ in narrative discourse emerged out of the possibility to use similar ‘reporting’ constructions to handle both speech and thought. In spoken discourse, it is equally possible to say Patsy said she was in love or Patsy thought she was in love, and Patsy’s disposition thus described is the same. What is different is the management of epistemic commitment of all the subjectivities involved. In the ‘say’ example Patsy went on record about her feelings, or the ‘narrator’ is going on record about correctly interpreting Patsy’s words (which may have been different). In the ‘think’ example there are more possibilities. Patsy may have said something suggesting she suspected her ‘symptoms’ indicated being in love (so the speaker is interpreting); or she may have communicated what her suspicions were (such as I think I am in love, though I’m not sure); or the speaker indicates that Patsy did think she was in love but no longer does; or the speaker means to suggest that Patsy was wrong, even if Patsy didn’t realize it at the time. All these interpretations arise out of various epistemic stance configurations prompted by the constructions, but the possibility which is in fact least likely is that the speaker is ‘reporting’ Patsy’s thoughts as they were, simply because in natural discourse we do not assume access to other minds, we can only construct our understanding of them and communicate that. Why, then, would narrative discourse be so different? It is only special in one respect – that it excludes the possibilities wherein the ‘narrator’ was in communication with Patsy, but otherwise it leaves the door open to options favoring ‘construal’ rather than ‘reporting.’ Thus narrative discourse inherits some of the constructional meanings available in spoken discourse, but it also adds the possibility that the narrator had access to Patsy’s mind without talking to her. This, in turn, is a viewpoint phenomenon, which can only be defined in terms of viewpoint configurations discussed in Chapter 3. Moreover, the idea of narrative access to minds calls for an explanation going beyond the usual treatment of STR (see Chapter 7 for more discussion). Before I move on to the sections detailing specific instances of microviewpoint besides STR, I will look briefly at the viewpoint potential afforded by seemingly straightforward representation of speech of characters. Example (1) illustrates some of the issues raised above. In the fragment, the Ego-narrator of Margaret Atwood’s The Blind Assassin, Iris, accounts for the information she received about the death of her sister, whose car went off the bridge: (1) I was informed of the accident by a policeman: the car was mine, and they’d traced the license. His tone was respectful: no doubt he recognized Richard’s name. He said the tires may have caught on streetcar track or the brakes may have failed, but he also felt bound to inform me that two witnesses – a retired lawyer and a bank teller, dependable people – had claimed to have seen the whole thing. They’d said Laura had turned the car sharply and deliberately, and had plunged off the bridge with no more fuss than stepping off a curb. They’d noticed her hands on the wheel because of the white gloves she’d been wearing. (The Blind Assassin, p. 3)
90
Viewpoint: representation and compression
Ostensibly, the fragment is an elaborate report of various stretches of speech and thought. There are thoughts of the narrator, as she guesses how the police got to her and why the policeman sounded respectful, there are the policeman’s words describing his interpretation of events and the information he had gathered from the witnesses, and there are the eyewitness reports which include descriptions of what happened, and in what manner. At this level, the fragment is saturated with STR constructions, with words and thoughts attributed to four different individuals. The fragment appears on the first page of the novel, and introduces some of the major narrative spaces and viewpoints to be elaborated. There are several levels determining how the fragment shapes the reading of the remaining parts of the text. First, there are several participants mentioned, but only three are given salient narrative roles: Iris (the Ego-narrator), Laura, her sister, whose death begins the story, and Richard (as we learn later, he was Iris’s husband). The actual speakers – the policeman and two witnesses – are not referred to by name and are never mentioned again. They are, in a sense, sources of information about the circumstances of Laura’s death, rather than characters proper. Primarily, their reports are used in establishing two possible interpretations of what happened – as an accident or as suicide (see Figure 4.1). Jointly, they construct the crucial point of the entire narrative, rather than having a specific role as characters whose speech and thought need to be represented. Second, the reported utterances, independently of their speakers, give different accounts of what happened. The policeman, who knows that Iris’s husband is a prominent businessman and politician, who wouldn’t appreciate an ugly case of suicide in his family, helpfully provides a possible explanation of Laura’s death as an accident (he reports it to Iris in this way). But he sounds speculative (may have caught, may have failed), and admits that the witnesses told a different story. The witnesses are very clear, report what they ‘saw’ and ‘noticed’, and use words like deliberately – clearly describing a suicide. Between the policeman’s hesitant epistemic stance and the witnesses’ strongly positive epistemic stance (supported by visually, and thus directly, acquired evidence), it is clear that suicide is the more likely explanation, but there may be reasons why it won’t be described that way. In effect, the reported words play no role other than the construction of two takes on the facts – suicide versus accident – and frame the rest of the novel as an attempt to explain both the suicide and the efforts to hide it. As such examples suggest, reporting utterances in narrative discourse may have goals pertaining to the overall structure of the story, and are not necessarily local attempts to render the ‘scene’ (RimmonKenan 1983). Contrary to the narrative form chosen, the reported discourse fragments here count as a kind of proleptic ‘summary,’ or, as I would prefer to call it, as narrative space builders. In order to consider them in this way, though, we have to assume that viewpoint spreads (in Fauconnier’s 1994 [1985], 1997
4.1 Viewpoint and representation SV-space
MN-space
Narrator: Iris present Narrator: Iris past tense
NS1 Laura dies DS 4 Clipping vpoint: accident
91
vpoint: accident
vpoint: suicide
vpoint: accident
vpoint: suicide
vpoint: accident
vpoint: suicide
DS 1 S: policeman H: Iris DS 2 S: policeman H: Iris vpoint: accident
DS 3 S: witnesses H: policeman vpoint: suicide
Figure 4.1 Reported voices and viewpoint (example 1); S = Speaker, H = Hearer, vpoint = viewpoint
sense) through compression across narrative levels, from the lower to higher levels of the story, and thus participates in story construction. This understanding of viewpoint will be explored throughout the chapter. The issue of STR and the various constructions involved will be taken up in Chapter 7. In what follows, I will attempt to clarify the conceptual underpinnings of micro-level viewpoint, starting with examples involving visual perspective and modes of visual representation, and leading towards understanding micro-level viewpoint as a representation of experiential phenomena and of conceptualization. I will then return briefly to the concept of focalization. 4.1
Viewpoint and representation
While the general tendency in narratology is to treat Genette’s ‘seeing’ metaphorically and move away from the visual understanding of viewpoint,4 I will start out here by showing that even literally visual viewpoint has specific discourse functions pertaining to narrative viewpoint in general, and helps in explaining viewpoint phenomena on the micro level of the narrative. The potential of visual viewpoint and visual representation is well illustrated by the extended fragment, from the first pages of Jonathan Raban’s travel narrative Old Glory, to be discussed here. The fragment, titled The River, shows a series
92
Viewpoint: representation and compression
of micro-level viewpoint shifts and compressions which use the narrative as a way of constructing a broad representation of a concept. Thus, what seems to be a non-narrative series of visual representations is a story of the emergence of a concept which is crucial in framing the entire text. The book is a travel story in which the narrator/traveler sails down the Mississippi in a small boat. It starts in the present tense, as a description: (2) It is as big and depthless as the sky itself. You can see the curve of the earth on its surface as it stretches away for miles to the far shore. Sunset has turned the water to the color of unripe peaches. There’s no wind. (Old Glory, p. 11)
The description suggests an observer taking in the view, even though it is phrased with the viewer off-stage, only as a conceptualizer. In the next sentences, the description becomes less literal: (3) A fish jumps. The river shatters for a moment, then glazes over. The forest which rims it is a long, loping smudge of charcoal. You could make it by running your thumb along the top edge of the water. (Old Glory, p. 11)
In the text, the forest is not compared to a “smudge of charcoal,” it is one. And in the very next sentence it is suggested that the description is not that of an actual view, but of a pictorial representation being created by someone. In fact, it is impossible to decide what the intended interpretation is, as the representation in the observer’s mind and the representation in a picture are blended, and thus the viewpoints are also blended. The compressed ‘representation/reality’ viewpoint introduces the story as one of both real travel and of narrativized aesthetic experience. However, the next paragraph changes the viewpoint (and the status of the “river”) once again: (4) It is called the Mississippi, but it is more an imaginary river than a real one. (Old Glory, p. 11)
“It” refers to the blended construct introduced in the first two paragraphs, but it also acknowledges its being mostly “imaginary” – so that the ‘objective’ existence of either the river or the painting is questioned. The next paragraphs shift into the past tense and profile an on-stage ‘I’ narrator, but the story is still not about events the narrator participated in, but about the gradual emergence of the concept of ‘the Mississippi’ in the conceptualizer’s mind. In the following paragraphs the reader is taken through various pictorial representations of the river which gradually shaped the traveler’s conceptualization and hi