How to Show Things with Words: A Study on Logic, Language and Literature 9783110899627, 9783110179958

Table of contents :
Preface
List of figures
List of tables
Abbreviations
Introduction
Part 1: Prolegomena
1. The linguistic structure of narrative transmission
1.1. Introduction
1.2. The showing-telling distinction
1.3. Narrative transmission as cognitive distance: From evidential modalities to indication signs
1.4. The role of tense, aspect and ‘Aktionsart’
1.5. Concluding remarks
2. Linguistics in narratology: A critical historical survey
2.1. Introduction
2.2. Ingarden, Stanzel, Hamburger: The neutralization of the ‘episches Präteritum’ as a past-tense form
2.3. Müller: Quantitative indicators and beyond
2.4. Weinrich’s textlinguistic theory: A tense-centered approach to backgrounded narrative discourse
2.5. Uspensky: Synchronic and retrospective viewpoints as a function of tense and aspect oppositions
2.6. Barthes: The semiotics of ‘L’effet de réel’
2.7. Chatman and Prince: ‘Aktionsart’ revisited
2.8. Caenepeel: Perspectivally situated vs. perspectivally non-situated sentences
2.9. Chafe: Displacement and immediacy in conversational language vs. displaced immediacy in narrative fiction
2.10. Concluding remarks
3. The narrating stance as locutionary subjectivity
3.1. Introduction
3.2. Speech-act theory and narrative discourse
3.3. The philosophical research on linguistic functions
3.4. The phenomenological make-up of the narrating stance as locutionary subjectivity
3.5. Locutionary subjectivity as a function of tense, aspect and ‘Aktionsart’
3.6. Summary and conclusion
Part 2: The temporal-perspectival organization of discourse
4. Tense
4.1. Introduction
4.2. Reichenbach’s theory of tense
4.3. Tense in narrative discourse
4.4. Tense, perception and memory
4.5. Concluding remarks
5. Aspect
5.1. Introduction
5.2. Classificatory systems of aspectual oppositions
5.3. Viewpoint aspect and point of view: A first view on the role played by imperfective meaning
5.4. Imperfectivity as a two-edged aspectual form or another view on viewpoint aspect and point of view
5.5. A note on iterativity
5.6. Concluding remarks
6. Aktionsart
6.1. Introduction
6.2. Vendler’s aspectual classes
6.3. Formal semantic approaches to ‘Aktionsart’
6.4. Concluding remarks
7. The effects of Aktionsart on narrative transmission
7.1. Introduction
7.2. -STAT eventuality descriptions
7.3. +STAT eventuality descriptions
7.4. World-knowledge based event semantics
7.5. Concluding remarks
Conclusion
Appendix 1
Appendix 2
Notes
References
Index of names
Index of subjects

Citation preview

How to Show Things with Words

W DE G

Trends in Linguistics Studies and Monographs 155

Editors Walter Bisang

(main editor for this volume)

Hans Henrich Hock Werner Winter

Mouton de Gruyter Berlin · New York

How to Show Things with Words A Study on Logic, Language and Literature

by Rui Linhares-Dias

Mouton de Gruyter Berlin · New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publicalion Data

Linhares-Dias, Rui, 1955How to show things with words : a study on logic, language and literature / by Rui Linhares-Dias. p. cm. - (Trends in linguistics. Studies and monographs ; 155) Includes bibliographical references and index. ISBN 3-11-017995-4 (alk. paper) 1. Language and languages - Philosophy. I. Title. II. Series. P107.L56 2006 401-dc22 2004021948

ISBN-13: 978-3-11-017995-8 ISBN-10: 3-11-017995-4 ISSN 1861-4302 Bibliographic information published by Die Deutsche Bibliothek

Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at . © Copyright 2006 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Cover design: Christopher Schneider, Berlin. Printed in Germany.

Para a Vania Para os meus pais, Manuel e Romualda Para ο Ivo, in memoriam

Preface Even if the successful application of formal logic to natural-language semantics does not require one to be a logician, as Landman claims, still it is reassuring to have their support for problem-solving tasks. To my regret, my logician-friends were too far away to prevent me from possibly incurring in errors that I would like to have eschewed. And yet formal semantics is only part of a lifetime interdisciplinary research program whose findings are finally materialized in this book, at the intersection of logic, linguistics, philosophy and literary theory. As a full mastery of the relevant literature was simply impracticable, some topics are barely discussed and others passed over into silence. Fortunately, books are not expected to provide full coverage of their subject matters or to be completely flawless and accurate products; their goal is to get some stimulating proposals into the public arena for critical consideration among those who share the same interests. It is my turn to enter the lists. Vania proved an indefatigable partner in elaborating the indices, and my parents provided the only financial support that I ever had. To get the book published was the least I could do in memory of Ivo. January 2006

Table of contents Preface List of

vii figures

xv

List of tables

xix

Abbreviations

xxi

Introduction

1

Part 1: Prolegomena 1. The linguistic structure of narrative transmission 15 1.1. Introduction 15 1.2. The showing-telling distinction 16 1.2.1. Quantitative indicators of scene and summary 16 1.2.2. Markedness in narrative pace: The dramatic mode 19 1.3. Narrative transmission as cognitive distance: From evidential modalities to indication signs 24 1.3.1. The problem of mimesis 24 1.3.2. The linguistic coding of epistemology 25 1.3.3. Evidential retrieval and perception verb complements (PVCs) 34 1.3.4. Evidentials as indexicals: The origins of showing ... 40 1.4. The role of tense, aspect and 'Aktionsart' 45 1.4.1. Tense and evidential skewing 45 1.4.2. Episodic and generic sentences 47 1.4.3. ±STAT eventuality descriptions 49 1.5. Concluding remarks 51 2. Linguistics in narratology: A critical historical survey 55 2.1. Introduction 55 2.2. Ingarden, Stanzel, Hamburger: The neutralization of the 'episches Präteritum' as a past-tense form 58

Table of contents

2.3. Müller: Quantitative indicators and beyond 2.4. Weinrich's textlinguistic theory: A tense-centered approach to backgrounded narrative discourse 2.5. Uspensky: Synchronic and retrospective viewpoints as a function of tense and aspect oppositions 2.6. Barthes: The semiotics of 'L'effet de reel' 2.7. Chatman and Prince: 'Aktionsart' revisited 2.8. Caenepeel: Perspectivally situated vs. perspectivally non-situated sentences 2.9. Chafe: Displacement and immediacy in conversational language vs. displaced immediacy in narrative fiction ... 2.10. Concluding remarks

3. The narrating stance as locutionary subjectivity

66 69 78 84 89 93 101 105

107

3.1. Introduction 107 3.2. Speech-act theory and narrative discourse 109 3.2.1. Sentence types as carriers of illocutionary force .... 109 3.2.2. The form-meaning relationship problem 112 3.2.2.1. The abstract performative analysis 112 3.2.2.2. Illocutionary-force coding operators 114 3.2.3. Statement-making sentences as a necessary but not sufficient condition for mimetic narrative discourse 115 3.3. The philosophical research on linguistic functions 117 3.3.1. Setting the stage 117 3.3.2. Anton Marty: Intimation and arousal; meaning as function or illocutionary point vs. meaning as prepositional content 118 3.3.3. Indicative signs in Husserl's ideational theory of meaning 122 3.3.3.1. Indication vs. expression 122 3.3.3.2. Intimation of non-extensional expressions and intimation of non-linguistic acts proper 124 3.3.4. Buhler's organon model of language 130 3.3.5. Jakobson's extended functional model 132

Table of contents

xi

3.4. The phenomenological make-up of the narrating stance as locutionary subjectivity 133 3.4.1. The intentional stance 133 3.4.2. The concept of objectivating act 136 3.4.2.1. Signitive vs. intuitive objectivating acts .. 137 3.4.2.2. Doing away with non-objectivating acts .. 144 3.5. Locutionary subjectivity as a function of tense, aspect and 'Aktionsart' 146 3.6. Summary and conclusion 150 Part 2: The temporal—perspectival organization of discourse 4. Tense 157 4.1. Introduction 157 4.2. Reichenbach's theory of tense 158 4.2.1. Tenses as ordered triples of time-points 158 4.2.2. The ambivalence of Reichenbach's reference point 162 4.2.2.1. The 'erratic fate' of a seminal concept .... 162 4.2.2.2. Reference time and temporal anaphora 163 4.2.2.3. Reference point and temporal perspective 169 4.3. Tense in narrative discourse 173 4.3.1. The choice of a formal system 173 4.3.2. Hybrid improved tense-logic (HITL) 175 4.3.2.1. Simple past tensed discourse 178 4.3.2.2. Sequences of simple past and past perfect sentences 181 4.3.3. Discourse Representation Theory (DRT) 184 4.3.3.1. The simple present and the present perfect 188 4.3.3.2. Present and past tense systems: A comparative analysis 192 4.3.4. Si nemo a me quaeret, scio, si quaerenti explicare velim, nescio 195

xii

Table of contents

4.4. Tense, perception and memory 4.5. Concluding remarks

202 209

5. Aspect 5.1. Introduction 5.2. Classificatory systems of aspectual oppositions 5.2.1. Perfective vs. imperfective aspect 5.2.2. Phase and habitual aspects as separate categories 5.2.3. Internal vs. external aspect forms 5.2.4. A tentative aspectual reclassification 5.3. Viewpoint aspect and point of view: A first view on the role played by imperfective meaning 5.3.1. IMP: An aspectually sensitive tense 5.3.1.1. Tpt-based approaches to internal perspective 5.3.1.2. TPpt-based approaches to internal perspective 5.3.2. PROG: A state-forming aspectual operator 5.3.2.1. Synchronic and diachronic linguistic evidence 5.3.2.2. PROG = +STAT modulo the subinterval condition 5.3.3. The phenomenology of state perception 5.4. Imperfectivity as a two-edged aspectual form or another view on viewpoint aspect and point of view 5.4.1. The plural form of imperfective sentences 5.4.2. The habitual reading of the French IMP 5.5. A note on iterativity 5.6. Concluding remarks

211 211 213 213

6. Aktionsart 6.1. Introduction 6.2. Vendler's aspectual classes 6.2.1. Syntactic tests and entailment relations 6.2.2. A brief informal characterization 6.2.3. Aspectual multivalence: Telic and atelic coercion

257 257 258 258 263 . 266

216 220 224 225 225 226 230 233 234 236 239 245 245 249 253 255

Table of contents

6.3. Formal semantic approaches to 'Aktionsart' 6.3.1. Dowty's aspect calculus 6.3.2. Beyond aspectual feature algebras 6.3.3. Multi-indexed modal analysis of gradual becoming 6.3.4. Lattice-theoretic event semantics 6.3.5. Two-dimensional dynamic temporal logic 6.4. Concluding remarks 7. The effects of Aktionsart on narrative transmission 7.1. Introduction 7.2. -STAT eventuality descriptions 7.2.1. Protracted and non-protracted situations 7.2.1.1. Achievements 7.2.1.2. Accomplishments 7.2.1.3. Activities 7.2.2. The phenomenology of event perception 7.2.2.1. Motion proper vs. simple displacement 7.2.2.2. From survey-knowledge to routeknowledge 7.2.2.2.1. The bird's-eye perspective 7.2.2.2.2. The field-perspective 7.2.2.3. Nonmotional changes of state 7.3. + STAT eventuality descriptions 7.3.1. Stage-level predicates 7.3.2. Individual-level predicates 7.3.3. Thetic and categorical judgements 7.4. World-knowledge based event semantics 7.4.1. The 'specious' present hypothesis 7.4.2. Revisiting the phenomenology of event perception 7.4.2.1. Achievements 7.2.1.2. Accomplishments 7.2.1.3. Activities 7.5. Concluding remarks

xiii

270 270 274 277 285 289 294

295 295 297 297 299 300 315 325 325 332 332 337 344 350 350 360 369 375 375 376 376 379 383 383

xiv

Table of contents

Conclusion

385

Appendix 1

401

Appendix 2

411

Notes

415

References

483

Index of names

519

Index of subjects

527

List of figures

1. Nested structure of evidential modalities 2. Model of epistemological stance adoption according to Mushin 3. Evidential and epistemic modalities 4. The classification of clausal perception verb complement types 5. Correlation of aspect and 'Aktionsart' with evidential stance adoption 6. Binary oppositions underlying point of view found in the literature 7. Overberg's algorithm for measuring the variations in timeratio of discourse-time to story-time against the average narrative rhythm ('mittleren Erzählrhythmus') 8. The Ogden and Richards's triangle of signification 9. AC as a causal dyadic relation in Barthes's reduced model 10. The immediate mode of speaking according to Chafe 11. Speaking/writing in the displaced mode (adapted from Chafe) 12. The displaced immediacy in storytelling (adapted from Chafe) 13. Sadock's-style P-marker for performative structures 14. Biihler's organon model of the semantic functions of language 15. Communicative factors and linguistic functions 16. The act-structure underlying locutionary subjectivity 17. The noetico-noematic correlations in the form-matter schema 18. Husserl's classification of intentional acts schematised by Mohanty 19. The ordering relations between the points S, R and E on the time line

28 33 34 36 53 57

67 85 86 101 102 103 112 130 132 134 137 139 160

xvi

List of figures

20. Set-theoretical inclusion as a linear order(ing) on events 168 21. The R-point as intersentential anaphora and as temporal perspective 172 22. 'Fictive present' as a backshifted series of temporal perspective points 180 23. (A) synchronous perspectives of Mendilow's fictive present and past 183 24. Perspectival structure of texts (156) and (163) in E,R,S-terms 194 25. The sources of input to consciousness according to Chafe .... 202 26. The flux of time as diagrammed by Merleau-Ponty 203 27. The R + Ε-reading of perfective/imperfective aspect 213 28. Comrie's classificatory system of aspectual oppositions 214 29. Thruth-conditions for iterated achievements with in/definite nouns 219 30. Structure of nuclear, core and extended predications 220 31. The developmental stages of So As: onset, nucleus and coda 222 32. The overall aspect(ual) system of Functional Grammar 223 33. Temporal relations between sentences in the PS and the IMP 229 34. Perspectival shifts between sentences in the PS and the IMP 232 35a First stage: if PROG φ is true at 7, φ holds of its superinterval J 237 35b Second stage: if φ is true at J, so is PROG φ at subinterval Ij C I 237 35c Third stage: if φ is true at «7, PROG φ holds of every InClcJ 238 36. Restriction function f\S of Ε J^O to S C E 243 37. The order-preserving failure of iteratively recounted eventualities 254 38. Change as a transition between the truth values of propositions 272 39a A" is indefinitely extensible to the left in two-valued logics .... 273 39b Jis indefinitely extensible to the right in two-valued logics ... 273

List of figures xvii 40. VerkuyPs basic structure of aspectual construal 41. Functional relation between s, / and π for the arguments #1, ...,xn 42. Increasing function from time-points to truth values 43. Krifka's representation of the Theme-to-event homomorphism 44. The Moens-style nucleus-structure of events 45. Execution sequence of nuclei-structures for Accomplishments 46. Execution sequence of nuclei-structures for Activity verbs 47. The nucleus-structure of Achievements according to Naumann 48. The terminative function t : ( s o p ) for the arguments χι, ...,xn 49. Interval generated by model-theoretically interpreting [+ADD TO]-marked verbs in terms of the successor function s : Int (Γ) -»· Int (Γ) 50. Participancy function relative to the (partitioned) subset Ε of A 51. Graph of terminative function sop : E/R* —> Ia 52. The terminative function composed of (318) and (308) 53. Closed interval generated by the terminative function applied to A 54. The (facial) structure of achievements 55. Cross-section of the incremental structure of accomplishment-terms 56. The incremental path-structure of accomplishment predications 57. Graph of the characteristic functions of a/telic situations 58a Perceptual path of activities on a set-theoretical reading 58b Perceptual path of accomplishments on a set-theoretical reading 59. Graphic representation of telic and atelic eventuality-description types

275 276 283 287 289 291 292 293 302

304 307 308 309 309 310 312 314 316 317 317 318

xviii 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

List of figures

The referential structure of atelic situation types 319 The referential structure of telic situation types 320 Graphic illustration of the spatial zones in (333) 326 A&S's illustration of the CoL verbs listed in Table 20 328 To step into/leave the road in RCC-terms 330 To cross: sequences of RCC-relations grouped into a five-tuple of zones 330 To travel: a path-gapping verb and a placeholder for spatial layout 331 Tree-form SDKS for text (339) given in (343) 334 SDKS for (339) augmented by (345) 335 Nick's path to Hirsch's rooming house and back to Henry's lunchroom 343 The incremental process according to Rothstein 345 Rothstein's event structure for accomplishments 346 The one-many trace relation from events to their run-times 352 One-many relation between events and stage-level predicates 354 One-one correspondence between discourse- and story-time 354 S-induced development of activities and accomplishments ... 359 Mapping objectivating acts onto narrative modes 387 Mapping inherent meaning onto objectivating acts 395 Product or composite function of / : A —> Ο and :O^T 395

List of tables 1. Pseudo-markedness shift in narrative discourse rhythmic oppositions 2. Interaction between evidential and epistemic modalities 3. Two-dimensional display of Stanzel's typology of narrative situations 4. Genette's Voice χ Mode (Distance 4- Perspective) Cartesian product 5. The overall architecture of Weinrich's syntactic articulation of the tense system 6. The Cartesian product of reportive types and epistemic modes 7. Caenepeel's Referential Centres and associated Meaning Structures 8. Aristotle's view of meaning-conferring acts after Daubert 9. Husserl's view of meaning-givenness according to Daubert 10. Terminological variations (in German) on the announcement function 11. Traditional and Reichenbach's nomenclatures for his nine tense schemes 12. Feature-valued representation of DRT and Reichenbach's tense forms 13. Brinton's model for the classification of English aspectual oppositions 14. Tentative aspectual classification based on Comrie's and Dik's systems 15. Graph of the characteristic function of feature-bearing people 16. Syntactic tests and entailment relations for Vlendler's verb typology 17. Feature matrix of Vendler's four-way classification of time schemata 18. The effect of NP arguments and PPs on basic inherent meaning types

20 38 60 61 71 81 94 124 125 127 161 186 217 224 246 262 263 269

xx

List of tables

19. The logical structure of aspectual classes in Dowty's aspect calculus 20. CoL verbs as a set of ordered triples of z-valued assignments 21. Descriptive operators for lexical entries of motion verbs and prepositions 22. Network of CRIL-expressions for representing route description (355) 23. The dispositional predicate types as represented in Krifka 24. Cross-classification: private/nonprivate χ stage/individual predicates

270 327 339 342 366 368

Abbreviations AC BI Col CoP CR CRIL CUM DC DL DMP DR DRS DRT DT EC ES FG FOL GA GCS GPSG HITL IL IMP INCR IPVC ITL KB LF MP MS MtE MtO NEP

aspect calculus bare infinitive change of location change of position construction rule conceptual route instruction language cumulative reference disconnected dynamic logic defeasible modus ponens discourse representation discourse representation structure Discourse Representation Theory discourse-time external contact event semantics Functional Grammar first-order logic geometric agent geometric concept specification Generalized Phrase Structure Grammar hybrid improved tense-logic intensional logic imparfait incremental relation infinitive perception verb complement improved tense logic knowledge base logical formula modus ponens meaning structure mapping to events mapping to objects natural event predicate

xxii

Abbreviations

NTL NTPP PDL PHITL ΡΟ PPVC P-Q-P

PREP PROG PS PVC QUA RC RCC Rpt SCUM SDRT SIP SNG SO A ST STAT STP SVC TDIP TP TPP TPpt Tql TS UG UoO Z-IH Z-IT Z-OH Z-OM Z-OT

nominal tense logic non-tangential proper part prepositional dynamic logic backwards-looking fragment of HITL partial overlap participial perception verb complement plus-que-parfait

projective terms progressive passe simple perception verb complement quantized reference referential centre Region(al) Connection Calculus reference point strictly cumulative reference Segmented Discourse Representation Theory strictly internal path singular reference state of affairs story-time stative set terminal point serial verb construction temporal discourse interpretation principle temporal logic tangential proper part temporal perspective point text qualifier text structure Universal Grammar uniqueness of objects inner-halo zone inner-transit zone outer-halo zone outer-most zone outer-transit zone

Introduction 1. There is a deep-seated belief among opera-goers and opera-lovers that any really successful overture is that which manages to provide as plain succinct a foreshadowing of the subsequent drama or comedy. If in a wonderful tone picture such as the overture to Tannhäuser we do indeed apprehend two broad leitmotiven in admirable contrast, where the solemn chant of the Pilgrims ceaselessly struggles with the wild insistence of Tannhäuser on his trenchant song in praise of Venus, the truth pure and simple is that the baffling plot of Le Nozze di Figaro ensures that its equally remarkable brief overture can hardly achieve more than creating a delightful atmosphere of gaiety and amusement. In a sense, the introduction to a book is subject to the same fate. Thus, in those monographs focussing on a particular subject matter, it is by no means difficult to circumscribe the range of data addressed nor to state how they depart from concurrent literature on the field. However, when it comes to a complex interdisciplinary research study straddling the border between formal logic, linguistics and philosophy to solve a longstanding issue originating in narrative-theoretical soil, the introduction is almost as hopeless as the overture to classical opera and so cannot but simply sketch what the book is and is not about. The whole thread of the argument will then be deferred to Chapter 1. 2. Considering that the term 'deixis' is etimologically derived from a Greek root whose meaning is 'to show', 'to point' or 'to indicate', it seems reasonable to assume that How to Show Things with Words is essentially about the class of linguistic expressions used to relate utterances to the spatio-temporal co-ordinates of the act of utterance: I/you, this/that, here/there, now/then, today/tomorrow and the like. Demonstrative reference processes are really alluded to apropos of Husserl's essentially occasional, as opposed to objective, expressions and also in the context of Bühler's so-called demonstratio ad oculos, not to mention that the indexical function of perceptual intimation is certainly part and parcel of any serious attempt to shed new light on the narrative-theoretical issue of proximal vs. distal stance adoption.

2

Introduction

Narrative discourse, however, is of a kind that is chiefly characterized by a number of negative features such as present-tense exclusion and the absence of deictics of time and place (Benveniste [1959] 1966) (except with those forms of the past-tense system referred to below.1) Instead of lexical and grammatical features of speaker-based centres, narrative discourse is anchorable to an allocentric reference system contextually salient in the so-called diegetic(al) world it gives rise to, and avails itself of anaphoric expressions like the temporal adverbials for ordering the chain of events along the time axis, on the one hand, and the spatial locutions for describing the environment, on the other. On the face of it, it is certainly not in the deictic sense of the word that natural-language narrative systems can show anything at all. Consider now the following excerpts from L 'Education Sentimentale: (1) a. Jamais Frederic n'avait ete plus loin du marriage. D'ailleurs, Mile Rogue lui semblait une petite personne assez ridicule. Quelle difference avec une femme comme Mme Dambreuse! Un bien autre avenir lui etait reserve! II en avait la certitude aujourd'hui ... (Flaubert [1869] 1983: 409) 'Frederic's thoughts had never been further removed from marriage. Besides, Mademoiselle Roque struck him as a somewhat ridiculous little thing. What a difference there was between her and a woman like Madame Dambreuse! A very different future awaited him! He was certain of that now ...' (Flaubert 1964: 347) b. II se demanda, serieusement, s'il serait un grand peintre ou un grand poete; et U se decida pour la peinture, car les exigences de ce metier le rapprocheraient de Mme Arnoux. II avait done trouve sa vocation! Le but de son existence etait clair maintenant, et l'avenir infaillible. (Flaubert 1983: 59) 'He asked himself in all seriousness whether he was to be a great painter or a great poet; and he decided in favour of painting, for the demands of this profession would bring him closer to Madame Arnoux. So he had found his vocation! The object of his existence was now clear, and there could be no doubt about the future.' (Flaubert 1964: 61)

Introduction

3

The co-occurrence in (la)-(lb) of imperfective past tenses with deictic temporal adverbs like aujourd'hui ('today') and maintenant ('now') shifts the perspective from the narrative-discourse generating source to the internal sphere of the reflector-character's own consciousness, thus expressing his thoughts and feelings in free indirect speech form. Again, focalization shifts are a matter of psychological point of view and should not be confused with issues of proximal/distal perspective dependent on spatio-temporal properties of eventuality descriptions and their bearing upon phenomenological aspects of event cognition, as those subject to investigation in the sequel. 3. If it will have become quite clear that this investigation is not about space-time deixis and pronominal/demonstrative reference systems, nor about the division of/oca/labour between narrator and characters, to speak of a phenomenology of event semantics leaves one at a loss. The best way to reveal what the present book is really about is thus viewing a sample of the material to be studied later on in the text. With every single entry of quoted speech expunged from the first one, there follows a couple of examples to start with: (2) Nick followed the woman up a flight of stairs and back to the end of a corridor. She knocked on the door. ... Nick opened the door and went into the room. Ole Andreson was lying on the bed with all his clothes on. He had been a heavyweight prize-fighter and he was too long for the bed. He lay with his head on two pillows. He did not look at Nick. ... Nick went out. As he shut the door he saw Ole Andreson with all his clothes on, lying on the bed looking at the wall. (Hemingway [1928] 1984: 374) (3) She sat at the window watching the evening invade the avenue. Her head was leaned against the window curtains and in her nostrils was the odour of dusty cretonne. She was tired. Few people passed. The man out of the last house passed on his way home; she heard his footsteps clacking along the concrete pavement and afterwards crunching on the cinder path before the new red houses. (Joyce [1914] 1993: 42)

4

Introduction

Based on what might sensibly be called his innate stylistic intuition, the reader who has skimmed through these texts will have a feeling, no matter how loosely structured, of what it is they have in common. The motto might well run along these lines: All facts, both "material" and psychic, are shown to the reader as if he directly "saw" them. The narrator - of course only illusorilly - has the function of a mere object-glass of a photographic camera or of an exactly registrating scientific apparatus. (Mukafovsky, quoted in Dolezel 1967: 548)

To see how Mukafovsky accutely seizes the quiddity of (2) and (3), one must first consider what kind of texts they are both opposed to and which constitute the default or untutored mode of storytelling. The following examples bring out the contrast: (4) Ilfallait chaque mois payer des billets, en renouveler d'autres, obtenir du temps. Le mari travaillait, le soir, a mettre au net les comptes d 'un commergant, et la nuit, souvent, U faisait de la copie a cinq sous la page. Et cette vie dura dix ans. (Maupassant [1885] 1973: 164) 'Each month they had to meet some notes, renew others, obtain more time. Her husband worked in the evening making a fair copy of some tradesman's accounts, and late at night he often copied manuscript for five sous a page. And this life lasted for ten years.' (English translation in Brooks and Warren [1943] 1959: 111) (5) L'annee demiere Jean escalada le Cervin. Le premier jour il monta jusqu'a la cabane H. II y passa la nuit. Ensuite il attaqua la face nord. Douze heures plus tard il arriva au sommet. (Kamp and Rohrer 1983: 260) 'Last year Jean climbed the Cervin. The first day he climbed to cabin H. He spent the night there. Next he attacked the north face. Twelve hours later he arrived at the top.' (6) Emma Woodhouse, handsome, clever, and rich, with a comfortable home and happy disposition, seemed to unite some of the best blessings of existence; and had lived nearly twenty-one years in the world with very little to distress or vex her. (Austen 1994: 5)

Introduction

5

A little knowledge of literature and its history in the recent past suffices to realize that both groups of texts are meant to (re)introduce a discussion of long standing in literary criticism and narrative theory about two narrative rendering techniques dubbed showing and telling by Anglo-American critics at the beginning of the twentieth century. Based on summary as the default or unmarked mode of storytelling, telling is nothing but a brief and generalized report of a series of events which covers extended periods of time as well as a variety of locales. To the extent that the discourse-time is shorter than the story-time, the narrative generating (in)stance cannot afford to dwell upon details and is thereby compelled to adopt a vantage point of view from the height of which the deeds described are retrospectively apprehended. Showing, by contrast, represents rather the incorporation of dramatic principles into narrative discourse in scenic-like (re)presentation form, consisting in a quite detailed account of a specific 'space-time region'. The quantity of information available in terms of text length is so close to the sequence of events in the space-time region that the fictional world is apprehended from an immanent temporal perspective and the listener accordingly becomes a(n eye)witness to the course of action. In present-day work subsidiary to Genette's narrative theory, however, showing and telling surge forth as mimesis and diegesis, respectively, and under this hellenophilical guise the aforementioned pro-illusionist approach to the structure of narrative transmission is simply dismissed on the grounds that natural languages do signify without imitating. If so, then not even some recent findings in current linguistic research seem to escape so zealous a commitment to pseudo-semiotic truisms. In fact, just as the deictic particle 'i- in (Uto-Aztecan) Ute narratives represents the so-called mode of immediate deixis by means of which "... the hearer is brought into the middle of the scene so that the action takes place right in front of his/her eyes." (Givon 1982: 40), the very act of affirming inherent in the thetic forms of judgement couched in ^α-sentences of Japanese folk tales has perceptual grounds, creating the illusion that one is witnessing the events (Kuroda 1992). (The reader is referred to section 7.3.3 for a more detailed treatment.) In a similar fashion, it has also been argued that 'scene-describing' stative predicates figuring in the descriptive parts of narrative fiction should be taken as the perceptual observations made by an onlooker,

6

Introduction

either the narrator or some reflector-character in the diegetical world, so that readers vicariously re-live these perceptual acts (Dowty 1986). (The reader is again referred to section 5.3.2 for further information.) Now, since the progressive form is a state-forming aspectual operator, the following passage makes just the same point: Emmon Bach (pers. comm.) pointed out that in stories the Progressive Form can pertain to States in such a way that the objects which are in a given state are temporarily experienced by the narrator (Galton 1984; Dowty 1979). Thus, a sentence like The village was lying in the valley seems to report a state of a village as seen by the narrator who is telling the story as if he just had entered the valley. The use of the Progressive Form tends to actualize its ephemeral nature: the state is reported from the point of view of the narrator. (Verkuyl 1993: 37; my emphasis on 'as seen')

and there are other 'villages' in real literary corpora like the French, where aspectually sensitive tenses such as the IMP play the same role as state-forming aspectual operators as the English progressive form. In Madame Bovary, for instance, Flaubert uses the imperfective form to indicate that Rouen is seen by an observer placed above the town:2 (7) Puis, d'un seul coup d'oeil, la ville apparaissait. Descendant tout en amphitheatre et noyee dans le brouillard, eile s'elargissait au delä des ponts, confusement. La pleine campagne remontait ensuite d'un mouvement monotone, jusqua toucher au loin la base indecise du del pale. Ainsi vu d'en haut, le pay sage tout entier avait I'air immobile comme une peinture; les navires a I'ancre se tassaient dans un coin; le fleuve arrondissait sa courbe au pied des collines vertes, et les lies, de forme oblongue, semblaient sur l'eau de grands poissons noirs arretes. Les cheminees des usines poussaient d'immenses panaches bruns qui s'envolaient par le bout. On entendait le ronflement des fondemes avec le carillon clair des eglises qui se dressaient dans la brume. Les arbres des boulevards, sans feuilles, faisaient des broussailles violettes au milieu des maisons, et les toits, tout reluisants de pluie, miroitaient inegalement, selon la hauteur des quartiers. Parfois un coup de vent emportait les nuages vers la cote Sainte-Catherine, comme des flots aeriens qui se bmsaient en silence contre une falaise. (Flaubert 1966: 287)

Introduction

7

These data clearly show that from cross-linguistic typological studies to cutting-edge inquiries carried out within formal semantic research, it will have become a commonplace that whenever speakers/narrators are assumed to have seen whatever situations they set out to describe, readers/narratees automatically adhere to that perceptual parameter. It is thus likely that some knowledge of lexical decomposition analysis might have prevented Genette from prejudging a valuable inheritance. In fact, pairs of morphologically unrelated verbs like to see and to show bear the same semantic relation to one another as do pairs of verbs related by causative morphological constructions in other languages. Thus, to show is nothing but to cause others to see that one has seen, not in the non-epistemic sense of arousing an intuitive act character by means of which an object or situation is apprehended 'in the flesh', but rather as a sort of text-induced process of epistemological transfer such that the recipient will decode the perceptual information source as something he might have experienced if he had been on the spot. Showing and telling are hence just an epistemological epiphenomenon of the proximal or distal self-positioning of the narrative instance3 as regards the space-time coordinate system of the diegetical world where the event(ualitie)s described are deemed to have taken place; these eventuality descriptions are then said to be shown whenever they come close enough to the narrative mediating stance3 so as to be seen, or rather taken to be simply told whenever perceptually inaccessible from the spatio-temporal location it occupies in the universe of fiction. Thus, irrespective of describing space-occupying second-order entities or focussing instead on the psychic phenomena of reflector-characters, texts (2) and (3) are structured according to the order of perception and the listener vicariously partakes in the same direct experience made out without the priviledge of knowing what is about to come. The same does not hold of the other excerpts in the sample, however. In fact, instead of describing what happened on a single occasion, text (4) is nothing but an iterative report of a ten-year long period, whereas (5) is based on an analepsis-generating discourse relation4 with the side-effect of mismatching descriptive and temporal order(s). The adoption of retrospective vantage points also extends to text (6), betraying cognitive processing systems of prepredicative experience; all facts, both 'material' and psychic, are now told rather than shown.

8

Introduction

Though the showing-telling distinction has loomed larger than any other in twentieth-century narrative theory and literary criticism (indeed, to such an extent that the list of references thereon is almost as endless as Leporello's famous catalogue on his master's conquests), the discussion has nevertheless revolved around normative judgements passed on the basis of aesthetic considerations on the craft of fiction. As early as 1921, however, one of the leading figures of New Criticism had made it clear that: And yet the novelist must state, must tell, must narrate - what else can he do? His book is a series of assertions, nothing more. It is so, obviously, and the difference between the art of Defoe and the art of Flaubert is only in their different method of placing their statements. Defoe takes a directer way, Flaubert a more roundabout ... (Lubbock 1921: 63; my emphasis)

proving his awareness that narration was a fact of natural language and that issues of method should be traced back to linguistic structure; the only reason why the matter was not pursued any further was the lack of tools which modern linguistics provides for us to fill in the gap. Thus, if determining the source of knowledge for some course of events in languages endowed with morphological evidential-marking systems simply requires that closed classes of morphemes be properly decoded, the situation is quite different in those other families of languages on which the common run of current narratological research focus, since no paradigmatic evidential contrasts are part of clause structure. Although natural languages are by no means unifunctional systems, still the strictly subjective manifestations of language-using personae are not linguistically coded in formally traceable sets of morphemes. On the face of it, our first task is just to draw on and to elaborate on the rich philosophical research tradition on the functions of language in order to lay down the foundations of self-expression phenomena, while casting the net wider to encompass epistemological issues, too. Now, just as the 'i- and ^α-particles as well as the English progressive have been claimed to bring narratees into the midlle of the scene, there are arguably other linguistic devices to achieve that same effect. If so, then our second task is to infer the 'epistemological imprint' of the narrating stance from a specific set of surface discursive markers.

Introduction

9

4. In her 1977 pioneering book on the language-literature interface, Pratt had to find her own way through people belonging in each other's vicinity but unwilling to exchange programs or street guides. Thirty years later, though literary scholars are no longer unfamiliar with linguistic theories nor are linguists indifferent to literary studies, they do not share the same language nor do they use the same tools. In a study straddling the border between linguistics and philosophy to solve narrative-theoretical issues on a sound interdisciplinary basis, one way to redress the balance between the interwoven research areas is certainly to pitch the discussion at a level appropriate to all of them. Thus, each new chapter provides, by way of introduction or otherwise, the background information relevant to the subject under discussion. The exposition, on the other hand, was made as simple as possible, and for preventing what is called 'math anxiety' (Landman 1991), 'heavy Quinean artillery' (Taylor 1985) has been kept to a minimum, which is not to say that it should be of no interest to the cognoscenti. And now a word about the overall organization of the book. Chapter 1 resembles the symphonic suite from an opera or ballet, in that it offers a synoptic account of the common thread of the book. The contours of the basic distinctions adumbrated in this introduction are then toughened up with new texts and typological descriptions. Next, a batch of examples taken from typologically different languages illustrates how evidentials are used to build up narrative structures, as a foil for considering the indicative processes of evidential retrieval available in those languages with no obligatory coding of epistemology. The chapter ends with a series of brief introductory considerations on how the related linguistic categories of tense, aspect and Aktionsart are brought to bear on the two opposite modes of narrative rendering known as showing and telling. Chapter 2 is a critical historical survey, necessarily incomplete, of some of the main contributions made in narrative theory to address the 'showing effect' of natural language reports of non-verbal events. To the extent that every new theory inevitably grows out of a feeling that those which precede it are manifestly insufficient or inadequate, this survey of, and/or debate with, the aforementioned contributions provides stimulating theoretical insights for subsequent developments in matters of tense, aspect and Aktionsart.

10

Introduction

Chapter 3 constitutes an attempt to set up I'instance racontante as a source of epistemology on a par with illocutionary force marking, which is a new form of self-expression called 'locutionary subjectivity'. For that matter, the shortcomings of speech act theory are laid bare in the light of German philosophical research on linguistic functions, following a path from Marty's semasiology to Bühler's organon model and where Husserl's Kundgebende Funktion (announcement function) accommodates the intentional acts transpiring in narrative discourse as part and parcel of the speaker's overall communicative intention. Because these psychic phenomena are either intuitive or non-intuitive, the showing-telling distinction can thus be accounted for according to the locutionary subjectivity manifests itself in one or the other way. This being the case, the main task of the remaining four chapters, which lend a formal semantic vein to the second part of the book, is to cast about for the specific linguistic categories responsible for the temporal-perspectival organization of discourse underlying that particular function of announcing the kinds of psychic phenomena on which the showing-telling distinction has its epistemological roots. Chapter 4 provides an in-depth analysis of verbal tense based on Reichenbach's widespread three-point structured conception thereof. The ambiguity of the concept of 'reference point' is first disentangled, and after a brief excursus on the formalization of temporal anaphora, the main focus is on temporal perspective in the (con)text of narrative. The outcome of approaching tense via hybrid tense-logical systems or within the framework of DRT essentially amounts to the following: tenses with overlapping event and reference points, on the one hand, are unmarked forms to set the stage where the course of action occurs; tenses with non-overlapping pairs of such points, on the other hand, introduce sidelines for events viewed as past from that vantage point. Ultimately, the 'displaced immediacy' associated with the simple past is of no consequence provided that the proto-retentional horizon of (direct) experience be preserved under memory. Chapter 5 reviews well-known systems of aspectual oppositions in order to arrive at a tentative reclassification against which to gain new insights on how viewpoint aspect bears on narrative transmission. Just as the English progressive is a state-forming sentential operator, the French imparfait is also taken to be an aspectually sensitive tense;

Introduction

11

in either event, it is because the consequent stage-level predicates are perceptual state-descriptions that these imperfective forms are generally considered a device for heightening the sense of presentness couched in spatial metaphors such as On the spot' or 'from the inside'. After a brief account of the habitual reading of the French imparfait, the chapter closes with a note on iterative aspect. Chapter 6 is what might be called an introduction to Aktionsart, a much neglected verbal category in narrative-theoretical scholarship. Besides providing syntactic criteria for aspectual class membership and (a)telic coercion principles for overcoming aspectual multivalence, there is a large spectrum of formal semantic approaches to the subject: from Dowty's Aspect-calculus built upon an improved first-order logic to Verkuyl's thesis on the compositional nature of aspect formation, and from multi-indexed tense-logical analyses of 'gradual becoming' to Krifka's most influential mereological approach to event semantics, not to mention more recent two-dimensional dynamic temporal logic. The chapter is meant to make the book self-contained.5 Chapter 7 represents the culmination point of a long process, showing how Aktions art-dependent the showing-telling distinction is; packaging tense information in ordered triples of overlapping points is a necessary but not sufficient condition for certain narrative effects, and viewpoint aspect is ultimately amenable to situation aspect forms. -STAT eventualities (achievements, accomplishments and activities) are thoroughly examined in the light of narrative mediation processes, paving the way for the necessary phenomenology of event perception; the results achieved within the paradigm of motional changes of state are then generalized to cover nonmotional patterns of change as well. +STAT eventualities split into stage- and individual-level predicates, improving on the phenomenology of state perception of Chapter 5. The chapter is brought to its term by taking into consideration how empirical knowledge of the way situations do transpire in the world interferes with Aktionsart typologies based on logico-linguistic criteria. The conclusion naturally collects the data amassed throughout the body of the book and assesses its main findings in the light of whatever competing theories are currently available in the literature. The first of the two appendices is a glossary of narratological terms; the second, a notation index.

12

Introduction

5. Some final notes on other matters of interest are also in order here. Parenthetical abbreviations as in, say, Futebol Clube do Porto (FCP) are rather indicated by capitals in boldface or outside the parenthesis, i.e. either as Futebol Clube do Porto or F(utebol) C(lube) do P(orto). Whereas sentences prefixed with an asterisk are ungrammatical, the symbol \\ indicates non-standard readings. In a book-length interdisciplinary research study like the present, it is normal that a good deal of information be relegated to the notes, lest the reader may definitely lose the main thread of the argument. However, this is not to imply that they are unimportant. Although I am not much of a translator, quite on the contrary, not all English translations of German texts are referred to a source; in these cases they are of course my own.

Part 1 Prolegomena

Chapter 1 The linguistic structure of narrative transmission 1.1. Introduction

To the extent that the main thread of argument can be easily lost in a book-length piece of interdisciplinary research at the interface of narrative theory, linguistics, philosophy and neighbouring disciplines, this chapter has been designed to put the pieces together in a nutshell. One of the texts quoted from the outset is an instance of marked uses of narrative technique suggesting 'how to show things with words', but the associated quantitative indicators as well as the comparison with scenic models are hardly convincing tentative approximations; only a modal conception of narrativity broad enough to encompass both epistemic and evidential modalities can help forward the claim that narratees can be said to be witnesses to the events recounted. In fact, there is a host of languages across the world endowed with complex inflectional systems of linguistic (en)coding of epistemology, whereby it becomes relatively straightforward to determine whether the kind of knowledge transpiring along narrative discourse structures was arrived at on the basis of direct or indirect sources of experience: perception vs. reported- or inference-based knowledge and even belief. Though English and other familiar European and classical languages do not use to signal the evidentiary basis of epistemic commitment, it turns out that so-called intuitive and signitive objectivating acts, which are roughly equivalent to direct and indirect evidence markers, are inferable from the temporal-perspectival organization of discourse. Insofar as these psychic phenomena are manifested by indication signs, either the narrator is taken to have seen the situations he talks about, in which case he can then be said to be in a position to show them, or else to have been told of, or rather inferred, those same situations, in which case it is then assumed that he cannot but simply tell what he will not have managed to make out by direct sensory experience.

16

The linguistic structure of narrative transmission

In particular, tenses with overlapping event and reference points are the conditio sine qua non of direct forms of narrative transmission, lest those perceptual idioms with prima facie non-epistemic readings be inevitably subject to evidential-skewing (re)interpretive processes. Predictions like these, however, are nonmonotonic forms of reasoning which turn out to be defeasible by quantificational-aspect distinctions and inherent meaning considerations based on the cognitive processes associated with eventuality-description types. 1.2. The showing-telling distinction 1.2.1. Quantitative indicators of scene and summary A quarter of century ago, narrative transmission was defined as "the class of kinds of narrative presentation which includes as its two subclasses showing and telling (always remembering that narrative showing is different from, say, theatrical showing)." (Chatman 1975: 215). To say that showing is characterized by its emphasis on the scene, while telling is characterized by its emphasis on pure narration or summary just begs the question; for what is it that distinguishes scene from summary and, in the last analysis, showing from telling? In order to throw light on this intricate narrative-theoretical issue, consider the following text: (8) Entre les charges de cavalerie, des escouades de sergents de ville survenaient, pour faire refluer le monde dans les rues. Mais, sur les marches de Tortoni, un komme, - Dussardier, - remarquable de loin a sa haute taille, restart sans plus bouger qu'une cariatide. Un des agents qui marchait en tete, le tricorne sur les yeux, le menaga de son epee. L'autre alors, s'avangant d'un pas, se mit a crier: - Vive la Republique! II tomba sur le dos, les bras en croix. Un hurlement d'horreur s'eleva de la foule. L'agent fit un cercle autour de lui avec son regard; et Frederic, beant, reconnut Senecal. (Flaubert 1983: 489)

The showing-telling distinction 17 'Between the calvary charges squads of policemen came up to drive the crowds back into the side streets. But on the steps of Tortoni's a man stood firm, as motionless as a caryatid, and conspicuous from afar on account of his tall stature. It was Dussardier. One of the policemen, who was marching in front of his squad, with his three-cornered hat pulled down over his eyes, threatened him with his sword. Then Dussardier took a step forward and started shouting: "Long live the Republic!" He fell on his back, with his arms spread out. A cry of horror rose from the crowd. The policeman looked all around him, and Frederic, open-mouthed, recognized Senecal.' (Flaubert 1964: 411) X

Text (8) is the famous street-riot scene from L'Education Sentimentale in which Frederic Moreau sees his former friend Dussardier being shot by a policeman who turns out to be another former friend, Senecal. The scene is supposed to give the reader the feeling of witnessing the described course of events as something happening under his eyes, but the fact is that whatever sort of 'dramatic illusion' he might have experienced upon reading these last lines of Chapter V of Part III of Flaubert's novel is inevitably shattered at the turn of the page. Chapter VI indeed departs from the adagio-like cadence of that scene, offering a brief summary of Frederic's subsequent deeds: (9) II voyagea. II connut la melancolie des paquebots, les froids reveils sous la tente, l'etourdissement des paysages et des mines, l'amertume des sympathies interrompues. II revint. R frequenta le monde, et U eut d'autres amours encore. Mais le souvenir continuel du premier les lui renaa.it insipides; et puis la vehemence du desir, la fleur meme de la sensation etait perdue. Ses ambitions d} esprit avaient egalement diminue. Des anne.es passerent; et U supportait le desoeuvrement de son intelligence et l'inertie de son coeur. (Flaubert 1983: 491)

18

The linguistic structure of narrative transmission

travelled. He came to know the melancholy of the steamboat, the cold awakening in the tent, the tedium of landscapes and ruins, the bitterness of interrupted friendships. He returned. He went into society, and he had other loves. But the ever-present memory of the first made them insipid; and besides, the violence of desire, the very flower of feeling, had gone. His intellectual ambitions had also dwindled. Years went by; and he endured the idleness of his mind and the inertia of his heart.' (Flaubert 1964: 411) As a first approach to the disparity in effects showing up in these texts, it is common practice in narrative theory to resort to quantitative indicators such as the ratio between story-time and discourse-time, (the Erzählzeit of Müller 1947 or the temps-papier of Barthes 1967) defined as the amount of text-space, measured in printed lines/pages, alloted to the different subperiods or time-sections into which the erzählte Zeit ('story-time') spanned by the entire novel can be divided. Thus, whereas text (9) compresses some sixteen years of diegetic time in scarcely ten lines, text (8) devotes sensibly the same amount of discourse-time to cover a series of physical actions of short duration, i.e. the kind that do not take much longer to perform than to relate. The whole episode is developed in the cadence of a slow movement, phase by phase, in the entire development of its concrete fullness. Admittedly, there is by necessity a detail that will have been omitted in this kind of minute rendering of what took place on the occasion; story-time is isomorphic to the set of real numbers and cannot be exhausted in its flowing continuity by a finite number of sentences. Unlike a passage of quoted speech, which is supposed to report every word pronounced in a dialogue,6 it is only by convention that there is a sort of equality between discourse- and story-time in a text like (8). As soon as the page is turned, however, the speed limit of the average narrative rhythm of the novel taken as a whole is definitely broken. Text (9) is no longer the slow exposition of a single entire episode, but a generalized account or report of a series of recurrent situations covering a considerably lengthy period of time and a variety of locales.

The showing-telling distinction 19

The same amount of discourse-time alloted to a very brief time-span such as that of text (8) becomes obviously too meager to mention, even cursorily, each of the denumerably many incidents crowded on a sixteen-year long story-time, let alone to render them at any length. As a consequence, the narrator is compelled, not only to ignore the concrete course and multifarious detailing of the individual events, but also to group them together into equivalence classes of occurrence, whose recurrent features are summarily drawn in broad free strokes. In addition to its drastically foreshortened eventive or actional part, text (9) also contains stative information about certain dispostions like the idleness of mind and the inertia of heart that have become Frederic's habitual mood all over the time stretch under consideration. Furthermore, the time-ratio of discourse-time to story-time stands in direct proportion to contextual relevance: while the most important passages in the overall economy of the plot are presented with detail, the interim periods, on the contrary, are rather rapidly summarized; to achieve certain unexpected effects,7 however, the detailed rendering of trivial incidents is a subtle technique to conceal the centrality of the dramatic high-points, which are given a perfunctory reference, if any. However interesting, this toying with the plot can be dispensed with; only those variations in time-ratios that can be said to have a bearing on the structure of narrative transmission are of concern to us here. 1.2.2. Markedness in narrative pace: the dramatic method Although much confusion has arisen about what it exactly means, 'markedness in oppositions' is a concept essentially characterized by: (i) the existence of alternative marked and unmarked constructions; (ii) the presence vs. the absence or nonpertinence of some feature re, which is just the mark of the opposition; (iii) the inverse proportion relation between the marked form and the frequency of its occurrence; (iv) the 'expressive' communicative value intended with the marked construction type, that is, the effects or implicatures of markedness. Since the concept of markedness is inversely correlated with frequency of occurrence, to say that "summary narrative ... seems to be the normal untutored mode of storytelling" (Friedman [1955] 1967: 119-120) is to imply that scene forms the marked member of the opposition.

20

The linguistic structure of narrative transmission

As a corollary, the full-fledged discrimination of scenes is naturally supposed to occur less frequently than its unmarked counterpart, which runs against the fact that we have become so accustomed to the minute-detailing technique of the scenic model that we have great difficulty experiencing the broad and generalized strokes of summary found in Butler, Dickens, Fielding, Hardy, Thackeray and Trollope. Rather than an apparent contradiction, what really happened is that the modern novel has overexploited its characteristic propensity for scene in such a way that made it lose much of its markedness value. Note, however, that this is a devaluation of the mark through a sort of inflation process and not the markedness shift suggested in Table 1. For this to be the case, the marked member of the opposition would necessarily have to become 'demarked' and the unmarked one obsolete in order to make room for the appearance of a new marked category.8 Table 1. Pseudomarkedness shift in narrative discourse rhythmic oppositions Marked

Unmarked

Obsolete

Stage 1

Scene

Summary

—

Stage 2

—

Scene

(Summary)

Stage 3

?

Scene

—

In fact, as there is no new narrative form in the marked-member slot and as it cannot be predicted whether such a form will ever be realized, this phenomenon is hardly eligible as an instance of markedness-shift. Rather than implying that it got 'demarked', the historical process of devaluation of scene through an increase of frequency means only that this kind of detailing method is no longer a construction type (of narrative modelling) with the extra attentional-commanding value experienced after the first massive onslaughts on omniscient narration. The scene is still marked by the presence of some feature χ to achieve specific communicative effects that are beyond the reach of summary, so that the task we next come to grips with is to determine what kind of feature χ might be and for what purpose it will have been created.

The showing-telling distinction

21

As a rule, there are large stretches of the world (re)presented in fiction which are necessary to the plot but not worth being dwelt upon lest they become so exceedingly tedious as to dismay novel readers; these stretches are normally interspersed in the course of narration to fill the temporal gap between two scenes and betray the adoption of an Olympian vantage point of perspective beyond time and place, from the height of which they are surveyed with a sweeping glance. This method is thus clearly incompatible with the spatio-temporal constraints which discriminate the hie et nunc of particular occasions like Senecal's shooting of Dussardier at the sight of Frederic Moureau. Summaries like text (9) are meant to convey a piece of information about a tract of past time which they traverse as quickly as possible, but mere knowledge is more akin to history than to narrative fiction. "In history our object is to know. In fiction, our object is to feel and appreciate, and what counts is not the number of facts but the degree to which we have been made to live with them." (Beach 1932: 188). As reported in text (9), the events are too remote 'to be lived with'. Contrary to the characteristic looseness of default summary practices, the scene is governed by two closely related constraints by which it comes up as the marked construction type of narrative transmission. The first and foremost is undoubtedly the time-limitation constraint, which coerces the story-time into spanning a relatively short period and contributes thereby to the dense, strict partial order(ing) of events which is the hallmark of the scenic model of narrative presentation. Most narratives do not provide explicit markers of duration and a text like (8) is no exception. By everyday world-knowledge criteria, however, it is by no means unreasonable to assume that the streetriot scene does not considerably outlast the time it takes to read it. The second constraint is nothing but a mere corollary of the first one: it is the space-limitation as exemplified by the fact that the actional stage of text (8) is confined to a street around the steps of Tortoni's. Both space- and time-limitation are narrative-extraneous principles borrowed from dramatic representation, where the amount of action enacted has to be adjusted to the spatio-temporal frame of the stage. The scene represents therefore the incorporation of dramatic methods into narrative fiction, as a means of increasing the expressive power of an essentially symbolic and distorting medium like natural language.

22

The linguistic structure of narrative transmission

The following five texts have been concocted in such a way that each one but the first takes its cue, as it were, from that which precedes it, adding up to a chronologically ordered, self-enclosed account of the dramatic method and its problems directly based on historical sources: Alle mimischen Mittel, durch die der Dramatiker seinen Vorgang vor das äußere Auge und Ohr des Zuschauers bringt, gelangen zur Anwendung, um den Leser zu einer Art Zuschauer und Zuhörer zu machen, der - mittels des inneren Sinnes - Gestalten sieht und ihre Reden hört ... Ganz wie auf der Bühne. [All mimetic means by which the playwright places his scene before the external eye and ear of the onlooker are put to good use for turning the reader into a sort of spectator and hearer who sees characters and listens to their speech in his mind's eye ... Just like on the stage.] (Walzel 1926: 183) On the stage everything is present because everything is enacted directly for our eyes and ears. In a story, we have the psychological equivalent of the dramatic present whenever we have a vividly "constituted scene", as Henry James calls it, a selected, a "discriminated occasion". The dramatic method is the method of direct presentation, and aims to give the reader the sense of being present, here and now, in the scene of action. (Beach 1932: 148; 181) The dramatic mode ... enables the reader to merge himself into the fictive present and fictional time of the book, and creates in him the illusion of being present at the action, in both meanings of the word 'present', as he is when seeing a play on the stage. It achieves these effects by the direct presentation of scenes, where the fullness of detail, the limitation of time and the exclusion of extraneous comment or explanation, all help to give the feeling of what Beach calls 'the continuous dramatic present'. (Mendilow [1952] 1972: 111) Naturally, in all but the scenic arts - like drama and the ballet - pure mimesis, that is, direct witnessing, is an illusion. The question, then, is how this illusion is achieved, by what convention does a reader, for example, accept the idea that it is 'as if he were personally on the scene, though the fact is that he comes to it by turning pages and reading words. (Chatman 1975:215) This crucial question has never been satisfactorily answered by the proillusionists. (Sternberg 1978: 22)

The showing-telling distinction

23

In all these passages it is unequivocally assumed from the very outset that dramatic representation has an advantage over narrative fiction. To bridge the gap between the two genres, Walzel and Beach argue that narrative rendering should be modelled after the enacted scene, if not to achieve the immediacy in which that advantage consists of, at least to try to induce the reader to yield to an illusion of the sort. Mendilow does not fail to notice the temporal determination of scene (the limitation of time and the consequent fullness of detail is just the distinctive feature underlying that marked member of the opposition) and for Chatman the illusion of direct witnessing is called mimesis. The term designates the capacity of literature to represent reality, but had already been opposed to diegesis in the works of Plato and Aristotle in order to distinguish direct from indirect (re)presentation. The opposition, however, did not really come into play until it surged forth in Anglo-American criticism, rebaptized as showing and telling. Ever since then, dyads like mimesis vs. diegesis, showing vs. telling, dramatic vs. undramatic and, metonymically, scene vs. summary have become roughly equivalent forms to couch the same opposition. Chatman does not raise any objections against the illusionist tenet,9 but it is clear that on implying that there is quite a difference between the enacted scene proper and what might be called the paper-scene, then the question which immediately arises has certainly to do with how that illusion is achievable in verbal reports of non-verbal events. Indeed, it should come as no surprise that this question has never been satisfactorily answered by literary criticism or by narrative theory. On the one hand, literary critics hardly ever addressed the question of narrative transmission without both partisanship and preconception, out of a wish not to pass judgements but just to describe and explain. On the other hand, those narrative theorists who have tried to account for the technical devices by which novelists are supposed to preserve the illusion were not prepared to correlate narrative categories with discourse structures in a systematic and linguistically informed way, since they had only old-grammar rags to dress their naked intuition. Some of these attempts/efforts will be briefly reviewed in Chapter 2, where it can be seen that there has been very little agreement as to the linguistic structure underlying the polarity of narrative presentation. The next sections outline the guidelines of a solution to this problem.

24

The linguistic structure of narrative transmission

1.3. Narrative transmission as cognitive distance: from evidential modalities to indication signs 1.3.1. The problem of mimesis In the last few decades at least, the neo-Aristotelian valuing of showing as being the quintessence of storytelling has been strongly criticized, on the grounds that narrative fiction is a fact of natural language and that natural languages are known to signify without imitating. For Genette, "... aucun recit ne peut "montrer" ou "imiter" I'histoire qu'il raconte. II ne peut que la raconter de fagon detaillee, precise, "vivante", et donner par la plus ou moins Villusion de mimesis ..." (1972: 185) [... no narrative can "show" or "imitate" the story it tells. All it can do is is tell it in a manner which is detailed, precise, "alive", and in that way give more or less the illusion of mimesis (1980: 164)]. Applied to the theory and practice of the dramatic method, however, this sort of criticism is clearly misguided, all the more so since it is Genette who equates the idea of showing with the concept of imitation; in none of the texts quoted above is it implied that narration possesses the aptitude of showing, in the (pseudo) imitative sense of the word, the smile of the lover, the frown of the tyrant, the grimace of the clown. Literary theorists and critics need not read Saussure or be acquainted with the semiotics of Peirce to see what is obvious even to the layman, namely that the link between the signifier and the signified is arbitrary. Similarly, Genette's claim that the craft of showing ultimately boils down to an illusion of mimesis only achievable by different kinds of telling (i.e. more or less detailed, precise and 'alive') sounds as if he were proposing something unheard of, when the fact is that he is just repeating Lubbock, Beach, Stanzel, Mendilow and many other authors who have been blamed for not having said what they actually did and for having allegedly said what will never have crossed their minds. What is more, Genette did not quite succeed in his attempt to account for narrative rendering in terms of the modal category of distance because he failed to see that this is a question of the kind of evidence that narrators have about the diegetic world they proceed to recount. This section is meant to show how distance can be measured on an evidentiary scale between direct and indirect data-source information.

From evidential modalities to indication signs 25 1.3.2. The linguistic coding of epistemology All natural-language systems provide their users with the means for making statements, asking questions, issuing directives and the like. Which specific attitude a speech act expresses towards prepositional content is called its illocutionary force and in most languages these attitudinal differences are cast in the grammatical category of mood. The main function of narrative fiction, however, is just to tell a story (to recount whatever events certain dramatis personae participate in) by making statements instead of asking questions or expressing wishes — the illocutionary force of interrogative and imperative sentences consubstantiates the author's intrusions of omniscient narration types but such garrulous excurses are clearly narrative-extraneous elements. Now, the particular kind of attitude which defines the illocutionary force of statement-making sentences is called epistemic commitment, and the grammatical mood that is used for the purpose of expressing unqualified epistemic commitment is just the indicative (Lyons 1995). Just as the illocutionary force of statements is not the same as that of questions and directives, there are also quite distinct ways by which speakers and/or narrators can qualify their epistemic commitment, depending on what sort of evidence is adduced to support their claims. "On peut en effet raconter plus ou moins ce que Γόη raconte, et le raconter selon tel ou tel point de vue ... "Distance" et "perspective", ainsi provisoirement denommees et definies, sont les deux modalites essentielles de cette regulation de l'information narrative qu'est le mode." (Genette 1972: 183-84) [Indeed, one can tell more or tell less what one tells, and can tell it according to one point of view or another ... "Distance" and "perspective", thus provisionally designated and defined, are the two chief modalities of that regulation of narrative information that is mood. (1980: 161-62. Genette's own emphasis)]. Unfortunately, this definition is too vague and does take into account the evidential categories required by a modal foundation of distance. The discourse of narrative can, indeed, provide more or fewer details and thus keep at a lesser or a greater distance from what it recounts, so that the quantity of information is in inverse proportion to distance: just as maximal information implies a minimum of distance (showing), a minimum of information corresponds to maximal distance (telling).

26

The linguistic structure of narrative transmission

Ultimately, what this amounts to is that the fictional world presented is built around some 'deictic centre' defined by cognitive parameters: whereas some States of Affairs come closer to that deictic centre, carrying the bulk of detailed, action-centered, assertive information, others lie further away and are therefore much more difficult to access. Conceived of in this strictly cognitive sense, distance varies according to the reported SoAs are cognitively more or less close to the I-Origo, that is, according to the epistemic strength of the information source, so that narrative discourse develops along a modal axis representing transitional states from what is known to what is less so and vice versa. Consider a short passage from an autobiographical story in Tsez, (a Nakh-Daghestanian language spoken in the Daghestan Republic) where the narrator assigns a task to someone he is acquainted with but who fails to carry it out: (10)

t' αςίζί-η

0-oq-no

O-oxi-n

idur-no

disappear-and

I-become-ANTCVB

I-run-ANTCVB

home-and

0-ik' i-n I-go-ANTCVB

paraq'at relax

kec-xo

zow-no. ...

sleep-SIMCVB

be-PSTUNWIT

ς

dahamaq' aw usyno

0- uX'-er-si.

little

I-fear-CAUS-PSTWIT

also

'He disappeared and ran off home and slept peacefully. ... I frightened him a little.' (Comrie 2000: 7)10 The self-evident part of (10) is marked in the witnessed past (PSTWIT) -s(i) and the bulk of the story in the unwitnessed past (PSTUNW) -n(o), which seems to point out that direct-evidence signalling morphemes correlate with showing as do indirect-evidence markers with telling. The evidential parameter expressed in Tsez is basically that of mere witnessed vs. unwitnessed experience reported in the past tense and has been accordingly called a two-term evidential system of type Al, but there are languages with finer-grained models of evidential coding such as the four-term systems of type Cl (Aikhenvald 2003, 2004). Thus, in Tariana, a North Arawak language spoken in Amazonia, whereas both direct visual and auditory evidence is marked by verb suffixes like ka and mahka in the examples (lla) and (lib) below, indirect data-sources such as second-hand information and inference,

From evidential modalities to indication signs 27 either from physical perception evidence (results) or mere reasoning, are coded by the system of verbal inflections in (lie), (lid) and (lie) (Aikhenvald 2003a: 134-135): (11) a. Ceci Cecilia

tjinu-nuku

du-kwisa-ka.

dog-TOP.NON.A/S 3SGF-scold-REC.P.VIS

'Cecilia scolded the dog.' (I saw it: VISUAL) b. Ceci tjinu-nuku du-kwisa-mahka. Cecilia

dog-TOP.NON.A/S 3SGF-scold-REC.P.NONVIS

'Cecilia scolded the dog.' (I heard it: NONVISUAL) c. Ceci tjinu-nuku du-kwisa-pidaka. Cecilia

dog-TOP.NON.A/S 3SGF-scold-REC.P.REP

'Cecilia scolded the dog.' (I was told: REPORTED) d. Ceci tjinu-nuku du-kwisa-nihka. Cecilia

dog-TOP.NON.A/S

3SGF-scold-SPEC.INFR.REC.P.

'Cecilia scolded the dog.' (I inferred it: SPECIFIC INFERENCE) e. Ceci tjinu-nuku du-kwisa-sika. Cecilia

dog-TOP.NON.A/S

3SGF-scold-REC.P.INFR

'Cecilia scolded the dog.' (I assume it: GENERAL INFERENCE) The sentences in (12) differ from those in (11) in prepositional content, but the evidential paradigm is just the same (Aikhenvald 2004: 2-3): (12)

a. Juse irida di-manika-ka. 'Jose played football (I saw him play).' b. Juse

irida

di-manika-mahka.

'Jose played football (I heard him play).' c. Juse

irida

di-manika-pidaka.

'Jose d. Juse 'Jose e. Juse 'Jose

played football (Someone else told me).' irida di-manika-nihka. played football (I see evidence for it).' irida di-manika-sika played football (It is reasonable to assume so)'

Tariana developed under the areal influence of Tucanoan languages and the following examples show its evidential parallel with Tuyuca, a Tucanoan language spoken in Brazil and in Colombia (Barnes 1984):

28

The linguistic structure of narrative transmission

(13)

a. diiga ape-wi. 'He played soccer b. diiga ape-ti. 'He played soccer c. diiga ape-yigi. 'He played soccer d. diiga ape-yi. 'He played soccer e. diiga ape-hiyi. 'He played soccer

(I saw him play).' (I heard, but did not see, him play).' (Someone else told me).' (I see evidence for it).' (It is reasonable to assume so).'

These data clearly indicate that in some languages an ±EXPERIENTIAL parameter subdivides into attested, reported and inferred evidence, each one further subcategorized as Figure 1 indicates (Willett 1988).n

{

Visual (a) Auditory (b) Other senses

Evidentials -EXPERIENTIAL

Inference

_. . (e) / \ Reasoning

Figure 1. Nested structure of evidential modalities according to Willett

12

In order to pick up the thread of evidentially in narrative rendering, consider now the contrasts in the verb inflection(al) system of Sherpa, a language from the Tibeto-Burman family closely related to Tibetan: (14)

a. daa rice

saap-p eat-NOMNLZR

mi

ti

man he

yembur-laa

de -ki-nok.

Katmandu-DAT

stay-HE

'The man who is eating rice lives in Katmandu.' (I see or I have seen that ...) b. daa saap-p mi U yembur-laa de -ki-wi. rice

eat-NOMNLZR

man he

Katmandu-DAT stay-GN

'The man who is eating rice lives in Katmandu.' (It is known that ...) (Woodbury 1986: 191)

From evidential modalities to indication signs 29

In the present-tense system, direct sensory evidence is marked by the HABITUAL EXPERIENTIAL form no(k) and unspecified indirect evidence, be it quotative and/or inferential, by the GNOMic-signalling suffix -wi, but this evidential pair can be skewed in its meaning and distribution (Woodbury 1986: 193): (15)

a. 'jon-ki John-ERG

'ti

'kursw)q

the

chair

'ti it

dzo -βυ,η. build-PE

'John built the chair.' (I saw that ...) b. 'jon-ki 'ti 'kurs^q 'ti dzo -nok. John-ERG

the

chair

it

build-Pi

'John built the chair.' (I inferred/was told that ...) In the past-tense system, PAST EXPERIENTIAL value is marked by -swj in (15a) and -no(k) takes the place of -wi in (15b) as PAST INFERENTIAL. (The process by which a certain evidential marker in one tense takes on a different meaning in another tense is called evidential skewing.13) Now consider the Life of the Buddha told by a Lama (Givon 1982: 34): (16)

sangye cumden-dye thungup-ki nyima din zur-la, Buddha Sakya-muni

birth-GEN

day

seven after-DAT

Seven days after the birth of Buddha Sakyamuni, sangye cumden-dye-ki yum gcnj-pa-zoo-nok; Buddha Sakya-muni-GEN mother think-finish-PI

his mother died; tema zur-la, then

after-DAT

so after that, sangye cumden-dye-ki saη-yum zen-yi se-nok; Buddha Sakya-muni-GEN father-ERG secret-mother another-ACC marry-PI

Buddha Sakyamuni's father took another wife; tema zur-la, then

after-D AT

so after that, sangye cumde-dye-ki yap-taη uru-la see-cik swj-nok. ... Buddha Sakya-muni-GEN father- and step-mother-DAT son-one be-PI

Buddha Sakyamuni's father and mother had a son. ...

14

30

The linguistic structure of narrative transmission

Since the Lama will not have witnessed what was just passed down to him via a long chain of oral transmission (Givon 1982; Palmer 1986), the bulk of (16) is reported in the PAST INFERENTIAL marker -no(k), which is scant evidential justification for a 'true' story like Buddha's and thus casts doubt on the view that subjective epistemic certainty is necessarily based on the topmost categories of the evidentiary scale. The suffix -8ηη occurs only twice as part of direct-quoted speech forms: (17) a. acu-ti-ki-kii:

'nyal-i-l^ba Μαη-Μιαη ίοη-δυ,η?' si^-nok

brother-his-GEN-ERG

hell-Gen-land how-many

see-PE

say-Pi

his brother asked him: 'How many things did you see on hell?' b.'nye nyal-i-lt^ba-la οιυα-ΐαη-updaasa maa-mu ίοη-βυ,η' si-kyaa-nok ... I-ERG hell-GEN-land-DAT fear-rise-INF place many see-PE say-AUX-PI

Ί saw many fearful places in hell', the brother told Buddha ... The WH-question in (17a) and the clause in (17b) within which the -βυη form occurs are both embedded in -no(fc)-marked verba dicendi, stemming thereby from the same source of oral transmission as (16); the difference is that while summary (16) is by nature a low-ranked construct on the scale of evidentiality (Chung and Timberlake 1985), speech transaction (17) is of a kind that one reports as z/overheard, as it can be seen in the following Salar narrative (Dwyer 2000: 48):15 (18) a. nenosur var a. bowusur varar a. ... indzi awucuxnigi abasi vursen adziunigi helli o eile apparmis. 'There was an old woman and an old man. ... So the boy's father borrowed some of Uncle Vursen's money.' b. idzanigi jahtuguni gun ana appardzi. mother-GEN

pillow-3POSS.ACC

sun girl

take-go-ANT.DIR

'The sun maidens took your mother's pillow.' Directly perceived events are marked in Salar with the verb suffix -dzi and indirectly experienced ones with the suffix -mis or the clitic -a. Because fictional narrative is taken to recount -REALIS occurrences, both indirect!ve-signalling forms -mis and -a appear in the opening series of events (18a) but not in the -ck^-marked quoted speech (18b).

From evidential modalities to indication signs 31 If to package chunks of experience into linearly ordered sets of events seems to be indeed a distance-inducing cognitive process linguistically coded by indirective markers like -no(k] in (16) and -mis or -a in (18a), to describe is, on the other hand, a perspectivally situated activity. "Man beschreibt das, was man vor sich sieht" (Luk cs [1936] 1974:56) [We describe what we see], so that it should come as no surprise that evidentiality is marked by visual-perception signalling forms like erken in this paradigmatic descriptive passage from an Afghan Uzbek text (Jarring, quoted in Johanson 2000: 78-9): (19) Usu

toxtagan

dzajini

kordi

ki

that

stop-PART

place-POSS3SG-ACC

see-PAST3SG

that

He saw the place where he was staying (and established:) besr masdziidni b&r ujide turar ek'an. one

mosque-GEN

one

house-POSS3SG-LOG stand-PRES

ERKEN

He was (as he saw) in a house belonging to a mosque. ujini taginda beer jertola ba:r ek'an, house-POSS-GEN

bottom-POSS-LOC

one cellar

existing ERKEN

Under the mosque there was (as he saw) a cellar, usa a:lip ba:ryan rfersclemi usa jerge qojar that take-CONV go-PART thing-ACC

ek/εη. ...

that place put-PRES ERKEN

where the thief put (as he saw) the things he had taken. ... Erte bilen ... usa crfri turgan dzajg6 bardi. morning

with

that

thief

stay-PART

place-DAT go-PAST3SG

In the morning he ... went to the place where was the thief. Qarasa-ki usa o'yri usa masdzitni mullasi ek'an. Look-COND-that that thief

that

mosque-GEN

mullah-POSS ERKEN

Looking (he saw that) the thief was the mullah of the mosque. where the sole eveni-denoting sentence is unmarked for directivity. The coding of epistemology is a context-sensitive process determined by discourse-pragmatic factors such as the social intentions of the speech act participants and genre-bound conventions (Dwyer 2000). In particular, there seems to be evidence that evidential marking in narrative is based on world knowledge of the evidence-getting forms inherent in, say, direct speech and ±STAT eventuality descriptions, irrespective of empirically available data-source information, if any.16

32

The linguistic structure of narrative transmission

Further evidence for the cognitive basis of these mismatches between source of information and evidential coding is provided by so-called epistemological stance adoption in the process of narrative retelling. Macedonian has a morphological paradigm of evidential oppositions such that the definite confirmative past-tense form is +EXPERIENTIAL and the nonconfirmative past with the ί-participle codes reportedness. However, a narrative retelling fragment such as (Mushin 2001: 183): (20)

a. Mnogu se very

godd_

lebot

kaj majka mu.

REFL turn.out:L bread:DEF LOG mother

3SG.DAT

'The bread often turned out well at her mother's house.' b. Sega ama ovoj pat mu se zgreschi lebot. now but

DEM time 3SG.DAT REFL be.wrong:SP bread:DEF

'Now, but this time the bread had turned out wrong.' brings out that the reportive epistemological stance adopted in (20a) then shifts to a definite past marked +EXPERIENTIAL strategy in (20b) to vouch for information about a single occurrence of the event type. Epistemological stance shiftings in English can be illustrated with the Mouse soup story also from Mushin's corpus-based study on retelling. The teller is a middle-age woman who was once in the living room with all the family when a 'plopping' sound came from the kitchen. Her mother then sent the father in and he found a mouse in the soup. This is part of one of the narrative retellings (Mushin 2001: 132; 138): (21)

so she sends the husband in he's looking around and he notices that the cat's on the stove and apparently the cat never gets on the stove so he goes over to the cat and he looks into the pot ...

As the recount shows no marks of a reportive epistemological stance, (21) is quite reminiscent of current reports of direct perceived events, except when it comes to the background information about the cat. The reteller then has to resort to the reportive code by means of an adverbial of prepositional attitude which betrays that this is just the kind of knowledge s/he could not have made out by direct experience.

From evidential modalities to indication signs

33

The same goes for a second retelling of the mouse soup narrative, where the single clause under the scope of the reportive framing device drawing attention to the evidentiary source of the following excerpt: (22) and and when the father went into the kitchen he saw the cat on the top of the stove and she[previous teiier]said that the cat never used to do that right when the father saw that he~ you know he shushed it away and figured that the cat was trying to get some meat right

happens to be just the one modified by the reportive adverbial in (21). In fact, to the extent that habitual meaning is clearly inconsistent with singulative eventuality descriptions and threatens discourse coherence, epistemological stance adoption strategies redress the balance between the source of information and the pragmatics of discourse production.

Actual source of information (Invariant)

—>

Speaker assessment of actual source Adoption of epistemological / stance

Pragmatic interactional setting

/ Speaker / —* assessment of interaction

Linguistic choice of evident iality —

Figure 2. Model of epistemological stance adoption according to Mushin (2001)

Epistemological stance adoption depends on genre-bound conventions, rhetorical purposes and world-knowledge of eventuality descriptions; evidential coding in Figure 2 is just the linguistic output ofthat choice.

34

The linguistic structure of narrative transmission

1.3.3. Evidential retrieval and perception verb complements (PVCs)17 1.3.3.1. The use of the term 'evidential'18 is clearly synecdochical; it refers to evidence proper as well as to all data-sources which are downgraded modifications ofthat Urform ('proto-form') of knowledge. Evidentials are also a subsystem of epistemic modalities lato sensu, in the sense that the (de)gradation of subjective certainty in which qualified epistemic commitment consists of is a necessary by-product of the evidentiary strength of the available data-source information. It seems thus in order to draw the distinctions depicted in Figure 3.

Modalities

' Experience Evidential < Inference Hearsay I Belief ' Certain Epistemic Possible Probable >. Doubtful

Figure 3. Evidential and epistemic modalities

Now, although English has a rich repertoire of these modal categories (adverbs like maybe, modal verbs may/might and perceptual idioms), evidential specification is not as frequent as it is in those languages which explicitly mark it with coherent sets of verb affixes and clitics. For instance, the following excerpt from Hemingway's "The Killers": (23)

Nick opened the door and went into the room. Ole Andreson was lying on the bed with all his clothes on. He had been a heavyweight prize-fighter and he was too long for the bed. (Hemingway 1984: 374)

does not specify whether the whole situation has been experienced rather than deduced or merely inferred from circumstancial evidence, or observed by someone who might have reported it to the speaker. On the face of it, the problem is to look for what has been called an evidence evaluation procedure (Givon 1982) with which to retrieve the epistemological imprint of the text from deep and surface information.

From evidential modalities to indication signs

35

One way to carry this out is just take the sentences of which (23) is composed as clausal complements of high deep-syntactic structures containing perception verbs, modal auxiliaries and quotative phrases. The idea is to test the propositions in the subordinate clauses against different kinds of evidential categories for epistemological consistency, but it should be borne in mind that the interpretation of perception verbs varies according to the complement type that they occur with, as shown by the following examples: (24)

a. b. c. d.

i SAW Nick open the door and go into the room. ι SAW Nick opening the door and going into the room. *i SAW Nick to open the door and to go into the room. I SAW that Nick opened the door and went into the room.

(25)

a. b. c. d.

i SAW Ole Andreson lie on the bed with all his clothes on. ι SAW Ole Andreson lying on the bed with all his clothes on. *i SAW Ole Andreson to lie on the bed with all his clothes on. ι SAW that 0. A. was lying on the bed with all his clothes on.

(26)

a. b. c. d.

*i SAW him *i SAW him ?i SAW him [}i SAW that

(27)

a. b. c. d.

i SAW him be too long for the bed/*intelligent. ι SAW him being too long for the bed/*intelligent. ι SAW him to be too long for the bed/intelligent. ι SAW that he was too long for the bed/intelligent.

have been a heavyweight prize-fighter. having been a heavyweight prize-fighter. to have been a heavyweight prize-fighter. he had been a heavyweight prize-fighter.

There are four types of Perception Verb Complement (Felser 1999): bare Infinitive Perception Verb Complements, Participial Perception Verb Complements, full infinitives or to-PVGs and finite that-PVCs. Except for their failure to occur with non-stative predicates in (24c), full infinitival clauses pattern with that-PCVs rather than with Bis. IPVCs and PPVCs, in turn, are small or reduced clausal constituents with no tense markers and subject to a set of restrictions on auxiliaries and predicate type that are different only in aspectual specification.

36

The linguistic structure of narrative transmission

That the meaning of perception verbs depends on these two groupings is shown by the auxiliary in (26) and the predicate-type level in (27). IPVCs and PPVCs have non-epistemic or direct perception readings and sentences (26a) and (26b) are considered ungrammatical because the act of perceiving is not simultaneous with the situation perceived; yet the state described in (26d) obtains before the reference time of the matrix clause and so that-PVCs are not perceptual reports/idioms.20 The situation is identical with ίο-infinitives like (26c) (Felser 1999).21 IPVCs and PPVCs are subject to stage-level predicate constraints, but the admissibility of individual-level predicates in (27c) and (27d) suggests that see has the epistemic sense οι find in to- and that-PVCs. Figure 4 relates in/direct perception to the relevant features of PVCs. PERCEPTION-VERB COMPLEMENTS

[-Tense]

[-Prog]

I IPVC

[+Tense]

[+Prog]

[-Finite]

[+Finite]

I

I

I

PPVC

to-PVC

that-PVC

direct perception

indirect perception

Figure 4· The classification of clausal perception verb complement types

In order to retrieve the evidential information lurking beneath (23), however, ί/ιαί-constructions not only have the advantage of preserving the inflectional surface structure of the embedded clausal complement, as they are also compatible with all evidence-signalling verbal forms; evidentials other than perception verbs do not admit BI complements, and even perceivable physical properties are ruled out in IVPCs like */ saw Ole Andreson (be) tall (Higginbotham 1983; Chierchia 1995). Furthermore, depending on whether or not the temporal reference of the main clause overlaps with that of its complement (see 1.4.1 below), that-PVCs allow the whole gamut of evidential value assignments on a scale of evidentiality between direct and indirect sources of experience.

From evidential modalities to indication signs

37

Except for (26d), most of text (23) is based on direct experience. This choice can be provisionally motivated by the fact that there is cross-linguistic evidence from what Bhat (1999) suggestively called 'mood-prominent languages' that +EXPERIENTIAL and -EXPERIENTIAL oppose to each other as unmarked vs. marked evidential categories, which strongly suggests that decoding the data-source information of, say, English sentences is a default process that first assigns them an +EXPERIENTIAL value in preference to its -EXPERIENTIAL counterpart, unless that evidential value becomes 'defeated' by further information disconfirming the (default) assumptions one might have started with. In fact, (26d) shows that when the complement is epistemologically inconsistent with the kind of evidentially marked in the higher clause, that is, perceptual-based knowledge, there is no alternative but to choose between indirect data-sources like hearsay, induction or belief: (28)

a. i WAS TOLD THAT Ole Andreson had been a heavyweight prize-fighter. b. Ole Andreson MUST HAVE BEEN a heavyweight prize-fighter. c. IT SEEMS THAT Ole Andreson has been a heavyweight prizefighter. d. ι THOUGHT/SUPPOSED THAT Ole Andreson had been a heavyweight prize-fighter.

A state that obtains prior to the current reference time is not open to direct perception but can nevertheless be indirectly apprehended, either through verbal report or perception of results ((28a) and (28b)); in the latter case, say, the narrator might have taken perceptual data like the photograph of a heavyweight fight hanging on the wall or a dusty pair of boxing gloves as circumstancial evidence for inferring what he could not possibly have made out by simple direct experience. Example (28c) encompasses the quotative and inferential readings of both (28a) and (28b) as different types of non perceptual knowledge; other than its primary function of marking inference or induction, seems is also pressed into service, on a par with hearsay devices such as I was told that of signalling knowledge acquired through language. Belief in (28d) is an evidential modality in which the epistemic warrant for the assertion is not as reliable as in the other sources (Chafe 1986).

38

The linguistic structure of narrative transmission

1.3.3.2. The fact that quite often we are rather certain of the actuality of what happens just before your eyes does not mean that, on occasion, direct sensory evidence may not be treated as less than fully reliable: (29) a. i SAW Nick open the door and go into the room. b. IT LOOKS LIKE he opened the door and went into the room. (30) a. Vänia FELT that she was growing fat. b. The fact is that she wasn't. Both (b)-sentences cast doubt on the validity of the (a)-sentences pair, but as (29a) represents the highest degree of certainty about SoAs, calling direct perception into question is less likely a phenomenon than disconfirming other verba sentiendi (Kirsner and Thompson 1976); evidence is not supposed to be downgraded by epistemic modalities.22 Table 2. Interaction between evidential and epistemic modalities

MODALITIES EVIDENTIAL

EPISTEMIC

Evidence

Certain

Induction

Probable Possible Doubtful

Hearsay Belief

Data-sources which do not have this prima facie tangibility, however, move up and down the scale of reliability at the centre of Table 2, which cannot in any case be conceived of as a function that takes each evidential in the left-hand column as argument and yields the epistemic modality it lines up with in the right-hand column as value. The relation remains functional all the same (for instance, there can be no inference process that is certain and doubtful at the same time), but evidentials can be freely combined with the epistemic modalities, so that their degree of reliability increases or decreases in accordance.

From evidential modalities to indication signs 39

For if data-source markers carry implications as to degree of reliability, these implications should nonetheless not be thought of ne varietur, considering that evidential and epistemic modalities can be combined in order to bring out the forms of qualifying epistemic commitment. Consider the following examples: (31)

a. I WAS TOLD THAT Ole Andreson seems to have been a heavyweight prize-fighter. b. Ole Andreson MUST certainly HAVE BEEN a heavyweight prize-fighter. c. IT SEEMS THAT Ole Andreson must have been a heavyweight prize-fighter. d. ι THOUGHT/SUPPOSED THAT Ole Andreson perhaps had been a heavyweight prize-fighter.

Sentences (31a)-(31d) have in common that they not fully commit, to begin with, but they certainly do not do so to the same extent: while (3la) indicates that the epistemic warrant for what it asserts is less good than it might be, (31b) signals a higher degree of reliability; epistemic authority in (31c,d) is also downgraded to the level of (31a). In the second place, the particular arrangement of evidential forms in the left-hand side of Table 2 does not reflect a fixed hierarchical order; as already noted, these evidentials move up and down a continuum from less to more reliable knowledge according to epistemic marking. The fact that certainly has a higher degree of reliability than seems and perhaps means only that induction ranks over hearsay and belief in the specific cases of (31b) vis-a-vis (31a) and (31b) vis-a-vis (31d); the inferential reading of (31c) is on a par with both (31a) and (31d), not to mention that it expresses a less compelling inference than (31b). One last point that is worth mentioning in the present context is that modal auxiliaries either function as evidential markers proper or as the epistemic modalities which make it explicit how (un)reliable they are. Thus, seems plays the role of epistemic indicator of reliability in (31a), but there is no doubt that it provides data-source information in (31c). Conversely, while must [have] is clearly an evidential marker in (31b), its form in sentence (31c) becomes instead an epistemic modality.23

40

The linguistic structure of narrative transmission

1.3.4· Evidentials as indexicals: the origins of showing

1.3.4.1. Much of the recent interest in evidentiality was aroused by the study of Indian languages, whose verbal systems possess inflectional markers to indicate the sources on which the information is based. In most Indo-European languages, however, verbal systems are not inflected for evidential categories and there is little use of the modal adverbs, auxiliaries and idiomatic phrases they are equipped with.24 Our approach to the self-expression of epistemology is thus Husserl's, who developed Marty and Twardowski's work on linguistic functions to take essentially into account the indicative resources of language; for words are meaning-animated expressions (sinnbelebten Ausdrücke] as well as indication signs (Anzeichen) of acts of perception, cognition, emotion and will. Indication is the process of inferring from something actually present another (real) thing which happens not to be present. Texts like (23) do not need a system of inflectional markers so that evidentiality can be recovered from test-batteries (24)-(27) and (28); surface linguistic structure and deep-synctatic evidential information are related by association or causality in such a way that the presence of the indication sign points out to the existence of what is indicated. This approach is, of course, too coarse-grained if compared with the examination of coherent sets of verb suffixes and so cannot take into account the whole spectrum of evidential marking outlined in 1.3.2. For instance, the direct perception reading assigned to the higher evidential clauses of IVPCs and PPVCs (24a,b), (25a,b) and (27a,b) is excluded from (26a,b) in view of the grammatical structure of (23), but which of the remaining (quotative, inferential and/or belief-based) evidentials is actually the epistemological source of (26c,d) in (28) is something that clearly falls outside the indicative power of language. Fortunately, this is inconsequential to the showing-telling distinction. Considering that So As either happen before one's very eyes or do not, it is quite immaterial whether the information is based on hearsay, induction or belief; no matter how reliable, none of these sources of knowledge can cut down narrative distance as direct perception does. Besides, the dichotomy of (unmarked) direct perception vs. indirect data-sources is grosso modo the same as that of intuitive vs. signitive objectivating acts in the sense of the phenomenology of constitution.

From evidential modalities to indication signs

41

Signitive acts are a sort of catch-all class that encompasses induction, deduction and the like, but they will not be resumed until Chapter 3. For the time being, it suffices to observe that HusserPs distinction between two types of intentional acts is reminiscent of the way Kant splits up ideas into intuitions (Anschauungen) and concepts (Begriffe), which is more familiar philosophical parlance. 1.3.4.2. Verbal narratives consist of linearly ordered sets of sentences describing whatever relations certain dramatis personae participate in, constituting thereby the sequences of events that plots are made of. Other than this strictly referential content, narrative statements also convey information about the illocutionary force of the speech acts through which the narrator addresses his narratees, on the one hand, and the psychic acts that he will have experienced in order to arrive at the knowledge of the eventualities of his recountal, on the other. As mentioned earlier, this knowledge is either intuitive or conceptual. Intuitive knowledge essentially consists of acts of perception that seize objects 'in the flesh' and constitute accordingly the source of such derivative forms of direct experience as recollection and imagination. Conceptual or non-intuitive knowledge, in turn, comprehends all sorts of indirect or second-hand experience that cannot apprehend objects (and SoAs) except in the empty forms of imagery-void consciousness. In the last analysis, it depends on the linguistic structure of discourse for the kind of knowledge indicated to turn out intuitive rather than conceptual or vice versa. In other words, the grammatical parts of speech not only signal which mental acts the narrator is carrying out; they also signal to the narratee to execute those same acts or activities. Thus, if a sentence like Lucy was reading 'Principia Mathematica' while sunbathing in her yacht permits the inference that these things will have been within the visual field of whoever has made the report, the same cannot be said of Lucy worked a lot and got rich. Then she started learning mathematical logic and read 'Principia Mathematica'. In fact, while 'be reading' or 'be sunbathing' are observable states and, as such, can be caught a glimpse of at any time of the period for which they obtain, 'work a lot', 'get rich' and 'read' Russell's three-volume set inquiry into the logical foundations of Mathematics are so extended in time that cannot have been the object of the speaker's perception.

42

The linguistic structure of narrative transmission

The first sentence is an indicium of an act of perception and the one who reads it tunes himself in to the same 'epistemological wavelength', as it were, yielding to the illusion that he might as well have perceived Lucy's state(s) if he happened to have been on the spot at the time, in such a way as to vicariously re-live the perceptual experience itself. The second sentence, quite on the contrary, is clearly the 'expression' of the opposite modus cognoscendi and urges the reader to switch over to the corresponding conceptual wavelength.25 1.3.4.3. Just as conceptual knowledge is often used as a catachrestic cover-term for indirect evidenciary sources, also intuitive knowledge is a bundle of perceptual, imaginative and reminiscent phenomena. Perception, however, is that Urmodus of intuitive awareness by which objects are given/presented in person (zur Selbstgegebenheit kommen); memory and imagination are just its secondary or derivative modes. To the extent that visual perception is the most common evidential, the other four channels of sensory evidence can also be dispensed with, so that to trace the indication signs of the epistemological structure of narrative boils down to decide whether or not the narrator will have seen the SoAs that are part of his report; showing is only a short way. For although 'see' and 'show' are morphologically unrelated verbs, the semantic relationship between the two is indeed so close that the lexical entry for the latter has been accounted for by taking the former as the sentential object of the abstract predicator CAUSE: (32)

DO (χ,Φ) CAUSE [BECOME see' (t/,C)]

The embedded bivalent structure in (32) refers to a situation brought about by the (unspecified) activity which is the subject of CAUSE, whereby there is a clear connection between causativity and valency: the verb 'show' is causative and has augmented valency in relation to the transitive verb 'see'. (Lyons 1977; Van Valin and LaPolla 1997). As it stands, however, LF (32) hardly accounts for narrative showing. For Φ has, at least in what seems to be its most natural interpretation, the valency schema MOVE (ENTITY, SOURCE, GOAL) of locomotive verbs, which is not quite the sort of activity implied in narrative transactions.

From evidential modalities to indication signs 43

This schema is really appropriate for, say, / showed Vienna to Sophia, where both first-order entities do change location as indicated below: (33)

MOVE (ENTITY, SOURCE, GOAL ) CAUSE [BECOME see' (y, / 2 )]

I x,y

I k

I h

but on generalizing localistic terms from the paradigm case of concrete locomotion to more abstract kinds thereof like narrative utterances, the valency schema of 'journey' can be applied to this activity as well. For the sake of illustration, (re)consider part of (23), repeated as (34): (34)

Nick opened the door and went into the room.

"These words, and the sentence as a whole, can be understood as signals. They are signals of activities, and signals in two senses ... The same word is a signal o/and a signal £o." (Sokolowski 1978: 100). Thus, (34) is not only a signal of the act of perception explicitly referred to in the high deep-syntactic evidential constituent of (24d); it is also a signal to pass something other than a first-order entity from some (abstract) location X to another (abstract) location Y. As implied above, the entity we are referring to is not an individual, but rather the second-order entity see'(χ, ζ) object of the indication. Second-order entities do not travel by taking a particular path/route, but somehow get transferred from SOURCE to GOAL or RECIPIENT; the journey of see'(a:, ζ) from narrative (in)stance χ to its recipient y is thus a clear process of epistemological transfer induced by the text: (35)

TRANSFER (ENTITY, SOURCE, RECIPIENT) CAUSE

I

see'(.τ, ζ)

I

χ

I

y

[BECOME (3u)(see'(x,C)*(v) Λ see'(y,ν}}}26 Schema (35) is a first attempt to provide a sound linguistic basis to the narrative category of showing. In particular, it reflects the idea that readers vicariously go through the same (perceptual) experience ascribed to the speaker or to the narrator. (Dowty 1986; Smith 1999).

44

The linguistic structure of narrative transmission

Assuming that telling is a simple question of substituting ->see'(x,£) for see'(χ, ζ] in the ENTITY slot of (35), then the whole structure of narrative transmission can be accounted for as a generalized process of epistemological transfer based on the localistic concept of journey. It is worth considering the following text: A passage that presents objects and events as seen, perceived, or conceptualized from a specific focus-1 will, naturally and automatically, invoke a reader's adoption of (or transposition to) this point of view and open a window denned by the perceptual, evaluative, and affective parameters that characterize the agent providing the focus-1. If the focalization window is anchored in a narratorial origo, the usual reading assumption will be that the narrator is talking about what he or she imaginatively perceives. (Jahn 1996: 256)

Actually, this is just the linguistic footing of showing at its weakest, for the epistemological transfer induced by symbolic representations by no means entails that an actual act of perception will ever occur. The difference can be brought out by a tentative (re) formulation of (35) along the following lines: (36) TRANSFER (ENTITY, SOURCE, RECIPIENT) CAUSE

8€β'(χ,ζ)

χ

y

[D [BECOME see'(y,v)} Λ Ο [BECOME see' Where V still names [see'(χ, ζ)]*, there are two distinct situations which are brought about by the activity of uttering the narrative text: the process of epistemological transfer and an act of perception proper. These are, however, under the scope of opposite alethic operators: while the epistemological transfer is a necessary consequence of the Kundgabe/Kundnahme, so-called meaning-fulfilling intentions are just a simple possibility which obviously may or may not be actualized. This is not the place to go into the phenomenology of objectivation, but the predicate constant 'see' being presented in boldface or in italics is already a sort of iconic device to mirror the meaning-fulfilling role played by intuitive acts with regard to their signitive counterparts. The reader is referred to Chapter 3 for further details on the subject.

The role of tense, aspect and Aktionsart 45 1.4. The role of tense, aspect and 'Aktionsart' 1.4-1- Tense and evidential skewing Evidentiality was assumed to be inferable from S-discourse structure, but so far little has been said about the linguistic categories involved in the evidential value assigned to the sentential constituents of (23). To pave the way for further developments, this section points out the linguistic indication of epistemology underlying narrative rendering. Example (26d) was given one of the evidential values in (28a)-(28d) on account of the indirect-perception reading repeated here as (37): (37)

''i SAW THAT Ole Andreson had been a heavyweight prize-fighter.

but the evidential meaning assigned to (25d), also repeated as (38), is by no means incompatible with a non-epistemic perception reading: (38)

ι SAW THAT he was lying on the bed with all his clothes on.

Because in that-PVCs each clause carries a tense marker of its own, the S(equence) o(f) T(enses) may not have the same time reference so that evidential value assignment becomes a SoT-sensitive process laid down in these two principles of tense and evidential interaction: (39)

If the tense marking of the higher evidential clause is different from that of the proposition with which it occurs, the resulting evidential value is necessarily -EXPERIENTIAL.

(40)

If the tense marking of the higher evidential clause is the same as that of the embedded complement clause, then the resulting evidential value is, or may be, +EXPERIENTIAL.

In (37), the matrix and its clausal complement are not simultaneous: the SoA described in the embedded proposition is in the past perfect and occurs before the evidential marker, which is in the simple past and cannot be assigned a +EXPERIENTIAL value as predicted by (39).

46

The linguistic structure of narrative transmission

The unacceptability of (37) in its direct-perception interpretation stems from a clear incongruity between tense and evidential meaning: ρτϊ,πια facie perceptual idioms cannot possibly report accomplished eventuality descriptions without being subject to evidential skewing. Unlike (37), the time reference of the higher evidential clause in (38) overlaps with that of the propositional ί/ιαί-complement it occurs with; in the light of (40), then, the evidential value may be +EXPERIENTIAL. The matrix clause in itself has no bearing upon evidential assignment, which is clearly SoT-based in that-PVCs like: (38') a.^i SEE THAT he was lying on the bed with all his clothes on. b. ι SEE THAT he is lying on the bed with all his clothes on. In fact, note that (38'b), which is just a present tense variant of (38), has a higher clause with the same tense marking as that of (38'a), but while (38'a) is -EXPERIENTIAL by evidential skewing principle (39), (38) and (38'b) become instead +EXPERIENTIAL according to (40). Evidential value assignment is thus a SoT-dependent phenomenon, regardless of whether the tense of the matrix clause is present or past, but first and foremost complement-type sensitive as seen in 1.3.3.1. Consider the following pair of sentences: (41)

a. i SEE him lying on the bed with all his clothes on. b. ι SAW him lying on the bed with all his clothes on.

Since (41a) and (41b) are PPVCs with no complementizers or CPs and the tense opposition is neutralized in 0 complementizer types, (41a) and (41b) are then vacuously +EXPERIENTIAL in view of (40). Evidential value is thus a function of complement type and SoT-rules, never of the tense form of evidential markers alone (Woodbury 1986). Now, monotonic property Γ Η 0 — > · Γ υ Δ Η 0 does not hold of (40) because adding new premises may lead to withdrawal of its conclusion, in which case the (nonmonotonic) inference is Γ h φ and ΓυΔ | « φ.27 (See Krifka et al. 1995; Thomason 1997; Pelletier and Ascher 1997). In fact, it will be shown that the consequent of (40) is defeasible by taking into account the interplay of tense with aspect and Aktionsart.

The role of tense, aspect and Aktionsart

47

1.4-2. Episodic and generic sentences

Non-overlapping time reference between PVCs' clausal constituents indicates second-hand information, but their simultaneity does not necessarily imply that the data-source is direct perceptual experience. Text (42) contains one of the linguistic categories that may affect (40): (42) a. Job rose early in the morning and offered burnt offerings according to the number of them all; for Job said, It may be that my sons have sinned, and cursed God in their hearts. b. Thus did Job continually. (The Book of Job, 1: 5) In fact, (42a) consists of particular sentences referring to specific SoAs while the (b)-sentence S-structure contains an adverb of quantification that transforms the episodic meaning of (42a) into a habitual one. This clausal type of genericity of so-called 'characterizing sentences' (Krifka et al. 1995) can also be marked by used-to constructions like: (43)

Job used to rise early in the morning and offer burnt offerings according to the number of them all; for Job used to say, It may be that my sons have sinned, and cursed God in their hearts.

or by the luou/d+Infinitive form, which is what comes nearest to a future-in-the-past tense in English, but is also used on a par with the auxiliary used to in order to indicate habituality (Palmer [1965] 1987): (43')

Job would rise early in the morning and offer burnt offerings according to the number of them all; for Job would say, It may be that my sons have sinned, and cursed God in their hearts.

The verbal predicates in (43) and (43') are morphologically related to the episodic ones in (42a) and thus generalize over patterns of events, but so-called 'lexical', as opposed to 'habitual', generic sentences like Philosophers know German do not possess any episodic counterparts and do generalize instead over characterizing properties of individuals.

48

The linguistic structure of narrative transmission

Now, the adverbial form in (42b) and the morphological relatedness of habitual to particular sentences makes it possible to pick out, say, the opening sentences of (42a), (43) and (43'): (44)

a. b. c. a.

Job rose early in the morning. Job continually rose early in the morning. Job used to rise early in the morning. Job would rise early in the morning.

to test them for evidential value assignment against deep-syntactic higher clauses containing a verb of perception: (45)

a. i SAW THAT b. ι SAW THAT c.^i SAW THAT dA SAW THAT

Job rose early in the morning. Job continually rose early in the morning. Job used to rise early in the morning. Job would rise early in the morning.

Sentence (45a) then assigns a +EXPERIENTIAL value to (44a) by (40), but not to (44b) and (44c), as predicted by the very same principle. By the same token, (45d) becomes -EXPERIENTIAL irrespective of (39), for (44d) does not have the future-in-the-past sense of: (46)

Sophia would become even more beautiful than she already was at the time when I first realized that she was no longer a kid.

In (39) and (40), evidential value is predicted on the basis that tenses are not frequency indicators. The tense forms of both (44c) and (44d), however, contain what has been called 'un multiplicateur ou coefficient de frequence' (Imbs 1960) and so carry the aspectual marker +HAB, which makes tense-based evidential predictions such as (40) defeasible. In fact, (42a) reports events which took place in a single morning, whereby the story-teller might have been right on the spot sometime in the morning of Job's performing certain acts early in that morning, and the epistemological source of episodic meaning is then perceptual. Texts (42b), (43) and (43'), on the contrary, report generic events referring to all instances of Job's doing something every morning; the events may even have been seen one by one, but never all at once.

The role of tense, aspect and Aktionsart

49

Characterizing sentences are thus inductive generalizations based on a past series of observed instances that must have accumulated enough times so that the generic form can be truly asserted (Carlson 1995); generalizations are conceptual rather than prepredicative experience. 1.4.3. ±STAT eventuality descriptions 1.4.3.1. As pointed out in 1.4.2, tense-based evidential predictions become defeasible on account of quantificational-aspect distinctions. It is time to see whether other linguistic categories so far unconsidered also interfere with the kind of epistemological indication brought out by tense and the aspect forms of quantifying over occurrences of So As. Consider the following texts: (47) a. Elizabeth jumped out; and, after giving each of them an hasty kiss, hurried into the vestibule, where Jane ... met her. b. They travelled as expeditiously as possible; and sleeping one night on the road, reached Longbourn by dinner-time the next day. (Austen [1813] 1984: 251-2) In (47a) and (47b), the simple past tense of the first clause is resumed after the occurrence of a participial form in the second clause and so the texts should be given the +EXPERIENTIAL value predicted by (40). Besides, since that there are also no surface markers of +HAB aspect, the episodic reading confirms the tense-based evidential assignment. Compare, however, the following that-PVCs: (48) a. i SAW THAT Elizabeth jumped out and hurried into the vestibule, where Jane met her. b. *Ί SAW THAT they travelled and reached Longbourn by dinnertime the next day. Tenses with overlapping event and reference points, on the one hand, and aspect forms of existential quantification over events, on the other, are far from being a sufficient condition for evidential determination. In particular, (48b) makes it clear that +EXPERIENTIAL predictions based on tense and aspect are defeasible according to the relevant properties of the aspectual class that the verb predication belongs to.

50

The linguistic structure of narrative transmission

For it seems that achievements are indication signs of direct evidence while durative situations are signals of conceptual forms of knowledge. Until further notice, it suffices to point out that text (47a) is a typical instance of the incorporation of the dramatic method into narrative; it cannot reproduce events as they would be presented on the stage, but signals that the evidential source they stem from is perception. A signal of becomes thus a signal to adopt this perceptual parameter and the reader yields to the illusion of being in the scene of action. The sequence of events in (47b), on the contrary, cannot be enacted, which blocks any kind of illusion-preserving epistemological transfers. 1.4.3.2. The following text(s) refer(s) to states rather than to events, opposing the simple description of a character to what is called its characterization proper, namely: (49) a. [Harriet Smith] was a very pretty girl... She was short, plump and fair, with a fine bloom, blue eyes, light hair, regular features, and a look of great sweetness. b. Emma Woodhouse, handsome, clever, and rich, with a comfortable home and happy disposition, seemed to unite some of the best blessings of existence; and had lived nearly twenty -one years in the world with very little to distress or vex her. (Austen [1816] 1994: 5; 18. Italics supplied) What strikes the most in (49b) is that the data-source information is overtly represented by seemed instead of being simply indicated. This is an evidential modality rather than an expression of reliability, but whether it turns out to be a marker of induction or hearsay is immaterial because in either event nothing happens before our eyes, as embodied in the comment: Statements like 'Emma Woodhouse, handsome, clever, and rich, with a comfortable home and and happy disposition ...' show prior 'knowledge' of the character on the part of the narrator who can therefore identify the former to the reader at the very beginning of the text. Such statements also imply an assumption that the narratee-reader does not share this knowledge ... Austen's narrator goes beyond identification to provide a whole characterization of the heroine. (Rimmon-Kenan 1983: 97)

Concluding remarks

51

In fact, text (49b) is already considered a matter of past knowledge, regardless of either the evidential marker or the past perfect clause. The question, then, is what it is that distinguishes a text like (49a), where Harriet Smith seems to stand directly before us as on the stage, from the intoductory section of (49b): (49')

b. Emma Woodhouse, handsome, clever, and rich, with a comfortable home and happy disposition ...

where the telescopic view of the character cannot be brought on stage. This disparity in effects cannot be accounted for in terms of tense, aspect and Aktionsart, since every single clause is in the simple past, has singulative aspect and also denotes an individual-level predicate. Not all these i-level predicates, however, are alike in (49a) and (49'b): for whereas so-called private stative predicates such as to be clever, rich or happy all refer to unobservable dispositions and properties, adjectival predicates such as to be short, plump, fair, blue-eyed or light-haired are rather non private and can be observed at any time. The opposition of non private and private predicates (Smith 1991) thus turns out to emerge as the opposition of direct and immediate vs. indirect and second-hand sources of epistemology underpinning the corresponding forms of narrative transmission exemplified in the particular cases of (49a) and (49'b). 1.5. Concluding remarks

The linguistics of evidentiality is only sparsely treated in what follows, assuming that it has really contributed to shed light on how Husserl's phenomenology of constitution can bear upon narrative transmission; for opposing signitive to intuitive objectivating acts of consciousness is not just a theory-constrained distinction with no linguistic import, but a philosophically sound way to accommodate information source as coded in languages having obligatory evidential marking systems, especially if it is borne in mind that some evidential stance strategies adopted in those languages with epistemological coding paradigms yield perspectival modulations akin to those of a showing-telling type. It is easy to find instances of this parallelism in the relevant literature:

52

The linguistic structure of narrative transmission

traditional Tsez tales longer than (10), as already observed in note 16, usually begin with the unwitnessed-evidence signalling past form -s(i) and then switch to the opposite marker -n(o) as the story progresses, which is an eyewitnessing strategy for native speakers (Comrie 2000). The narrative-specific evidential marker in Macedonian is the /-past, with the definite form used for the sake of vividness (Friedman 2003) (a case in point is the evidential stance adoption exemplified in (20)). Yukaghir -EXPERIENTIAL stories also switch to +EXPERIENTIAL forms, at least for those few episodes presented with a great amount of detail where the narrator plays again the role of a witness (Maslova 2003), not to mention other language-specific -EXPERIENTIAL default markers of fictional genres punctuated by +EXPERIENTIAL passages of dialogue (cf. both (17) and (18) for the cases of Sherpa and Salar, respectively). Interestingly enough, Turkish indirectives like -mis, -ibdir and -gan are plot-advancing or discourse-propulsive functors (Johanson 2003), which means that not all passages propel narrative time forward. As the structure of narrative transmission obeys the very same logic of contracting into economical summary to expand into vivid detail, it should accordingly be also accounted for as shifting constraints on evidential-like dyads as the signitive and intuitive objectivating acts. The problem is that the languages we are particularly interested in clearly disallow that the showing-telling distinction be read off from (non-existent) traces of grammaticalization processes of epistemology, which suggests that one way out of this predicament might go through Nichols's (1986) universal-covariance hypothesis:28 (50)

UNIVERSAL COVARIANCE

aspect: evidentiality:

'perfective' -EXPERIENTIAL

'imperfective' +EXPERIENTIAL

to explain the tense-aspect based distribution of evidential meaning in Aikhenvald's Α-type languages with minimal two-term paradigms, where direct experience, visual or auditory, opposes to everything else. Languages like these are supposed to group ±EXPERIENTIAL meanings together into a single evidential parameter and it is up to tense-aspect morphology to predict the particular meaning intended in each case,

Concluding remarks

53

and though Sherpa is not a prototypical instance of Α-type systems, evidential skewing of -no(k) is among the distributional phenomena taken as manifestations of the universal covariance principle in (50). The terms 'perfective' and 'imperfective' only label kinds of meaning, irrespective of whether they are inflectional categories or verb classes; whereas perfective has the Aktionsart sense of culminated processes, encompassing accomplishments and achievements but not activities, imperfective subsumes meanings like durative, habitual and iterative. Perfectivity then correlates with -EXPERIENTIAL evidential meaning because situations seen as a whole cannot be immediately perceived and imperfective with +EXPERIENTIAL inasmuch as ongoing ones can. Figure 5 below is meant to improve on Nichols's universal covariance between tense-aspect morphology and evidentials cross-linguistically: Aspect PROGRESSIVE

HABITUAL

+EXPERIENTIAL/INTUITIVE STATES

ACHIEVEMENTS

ITERATIVE

-EXPERIENTIAL/SIGNITIVE ACCOMPLISHMENTS

ACTIVITIES >

V

Aktionsart Figure 5. Correlation of aspect and Aktionsart with evidential stance adoption

Let ongoing processes or situations be treated as states of change. Since every bit of a state is exactly the same as every other bit of it, homogeneity down to single instants is what makes them perceivable, but then the same holds vacuously of near-instantaneous eventualities; characterized by the fact of being over as soon as they have begun, achievements are stage-free and can be perceived whenever they occur. The progressive, in its turn, is a state-forming intrasentential modifier and so must be kept apart from other shades of imperfective meaning. Now, just as ±EXPERIENTIAL meanings in the evidential paradigm of Α-type languages (and very particularly of Chinese pidgin Russian) are tense-aspect based contextual variants of a single form or category, there are indicative functions in languages with no evidential coding

54

The linguistic structure of narrative transmission

enabling us to decide which objectivating acts, signitive or intuitive, are inferable from the kind of linguistic structure displayed in Figure 5; in particular, it depends on whether the eventualities recounted turn out to be punctual rather than durative to have direct and immediate, as opposed to indirect and mediate, forms of narrative transmission. This is definitely as close as we can get to an evidential-like strategy in English, French, German or any other related language providing the corpus for the narrative-theoretical study undertaken in this book, though in the end it will turn out that evidential retrieval strategies are not mere surrogates for grammaticalized systems of evidentiality: the indication of epistemology is as effective as its linguistic encoding.

Chapter 2 Linguistics in narratology: A critical historical survey 2.1. Introduction

Narrative transmission has been defined in 1.2.1 as the two different modes of storytelling known as showing and telling; "this distinction is often subsumed under the term 'point of view', but it is clear that that term can be seriously misleading ..." (Chatman 1975: 215). Indeed, most extant viewpoint typologies still conflate the distance between the narrative instance and the fictional world represented, which is a cognitive epiphenomenon of time-dependent parameters, with the subjective rendering of a reflector-character's inner life or the objective recount of the deeds he performs or is the patient of, which is a matter of focal choice based on empathy (Einfühlung).29 Yet no distinction between temporal and psychological points of view (Uspensky 1973) is required in order to steer clear of the polysemy; owing to Genette's proposal to term psychologische Standortsbegriffe ('concepts concerning the psychological perspective') as focalization, internal and external viewpoints can safely be taken as opposite forms of immediate and direct vs. second-hand and indirect transmission. Therefore, the expressions 'narrative transmission' and 'point of view' are used henceforth to denote the showing-telling distinction as well. As pointed out in 1.2.2, the question of point of view was already addressed in the works of Plato, who is credited with having first made the distinction between mimesis and diegesis, and Aristotle, who then merged the latter into the former as a mere variant thereof. Point of view did not, however, become the main concern of narrative theory until the prefaces to Henry James's novels paved the way for Beach (1918, 1932) and Lubbock's (1921) first theoretical construals, although Otto Ludwig's dyad eigentliche- ('true') szenische-Erzählung ('scenic narration') had already initiated the relevant discussion in the German-speaking countries by the turn of the nineteenth century.

56

Linguistics in narratology: A critical historical survey

Ever since then, there have hardly been any study or handbooks on narrative without their own definition of point of view, but eager to establish exclusive validity for one of the kinds of narrative rendering; while Ludwig, Spielhagen, Lubbock, Walzel and Beach have claimed the superiority of showing over telling, Friedemann, Forster, Petsch, Kayser and Booth are known for overstating their plea to the contrary. In either event, these opposite normative claims were put forward on the basic assumption that storytelling is an illusion-directed process, in the sense referred to in 1.2.2, i.e. who is engrossed in reading fiction is supposed to yield to the illusion of witnessing the course of action. This being the case, then "... d'oü procede au juste cette aptitude des recits de fiction, sinon a reproduire reellement les evenements qu'ils narrent, du moins a susciter chez le lecteur 1'illusion que ces evenements se deroulent sous ses yeux?" (Vuillaume 1990: 109).31 [... this aptitude of narrative fiction, if not to reproduce really the narrated events, at least to prompt the reader to the illusion that these events happen under his eyes, where exactly does it come from?] Although this question did not receive any satisfactory answer so far, some scholars have suggested interesting ways for further research. This chapter is a chronologically ordered survey of their writings, in the hope of gaining the clues to set up a unified theory of point of view based on the temporal-perspectival organization of discourse. The exposition is built around the names and topics shown in Figure 6, where 'aspect' refers to the way eventualities may be focused upon, i.e. whether they are seen as oingoing processes or as a single whole, and where the old Germanic term Aktionsart denotes verb semantics in the sense of the internal contours of Vendler's 'time schemata'; aspect is but a cover term for both viewpoint and situation aspects.32 Looking at Figure 6, however, it is apparent that there has been little, if any, agreement as to the role played by the tense-aspect categories of the verb in the linguistic coding of point of view in narrative fiction. There are authors who take into account either tense or aspect but fail to do the same in respect of aspectual class phenomena and vice versa; in no circumstance are all the linguistic categories brought together. Two apparent exceptions are Uspensky and, in particular, Weinrich, who yet takes the progressive vs. the nonprogressive form as a tense, rather than aspectual, opposition, let alone as a matter of Aktionsart.

Introduction

57

Both authors seem to take into consideration two linguistic categories, but they assign the same function to different sets of tense oppositions. In a similar fashion, those who are particularly attentive to the role played by Aktionsart also seem to disagree on which specific features of verb predications pair up with which modes of narrative rendering: Müller, Chatman and Prince claim that punctual verbs are especially cut out for showing while Fleischman goes just the other way around. Point of View (Linguistic encoding of) TENSE

P. Perf. «—> S. Past S. Pres. (Weinrich) | (Uspensky) I P. Prog. (Weinrich) ASPECT

ASPECT

PERFECTIVE


·

IMPERFECTIVE

| | S. Past

Past Prog. (Uspensky, Hopper, Kamp)

AKTIONSART

PUNCTUAL


·

DURATIVE

(Müller, Chatman, Prince)

QO

Figure 6. Binary oppositions underlying point of view found in the literature

Caenepeel has finally opposed perspectivally situated sentences and perspectivally non-situated sentences on a stative/non-stative basis. Weighing up the pros and cons of the several theses in Figure 6, it is our purpose to take stock of what they will have, or have not, accomplished in order to point out the guidelines of the research to be carried out in the technical chapters on tense, aspect and Aktionsart.

58

Linguistics in narratology: A critical historical survey

2.2. Ingarden, Stanzel, Hamburger: the neutralization of the 'episches Präteritum' as a past-tense form 2.2.1. Ingarden's famous Das literarische Kunstwerk is an outgrowth of Husserl's theory of meaning as further developed in Pfänder's Logik. Though this piece of interdisciplinary research study reveals a lifelong concern with major philosophical issues such as idealism and realism, Ingarden's main purpose, however, is to provide a thorough analysis of the multi-layered structure and mode of existence of the literary work. At the end of § 36 on "represented time and the time perspectives", for instance, there is a passage that might have laid the foundations of the contemporary view that the deictic time value of past-tense forms in normal speech does not hold in the context of narrative fiction. That passage is important enough to deserve being quoted at length: Erst wo eine Szene in ihrer konkreten Fülle und in ihrer vollen zeitlichen Ausbreitung zur Schau gestellt wird, haben wir wiederum mit der qualitativ bestimmten, dargestellten Zeit zu tun. Oder anders gesagt: Nur in dem letzteren Falle werden konkrete Zeitphasen in ihrer Individualität dar- und evtl. zur Schau gestellt. In den anderen Fällen dagegen wird die zur Darstellung gelangende Zeit nur ihrer allgemeinen Struktur nach als Zeit - oder als Zeitphase in einer bestimmten Lage in dem Zeitkontinuum - dargestellt, nicht aber als ein schlechthinniges Individuum in ihrer Individualität. ... Bei der bloß "berichtenden" Erzählungsweise werden die dabei dargestellten Zeitperioden immer als vergangen, von einem "späteren", übrigens sonst unbestimmten, Zeitmoment aus aufgefaßt. Eine charakteristische Zeitferne tritt deutlich zutage. Dagegen kann die in ihrer schlechthinnigen Individualität, Phase für Phase in ihrem Verlauf dargestellte Zeit zwar auch als vergangen aufgefaßt werden, sie tritt aber in einer eigentümlichen Nähe auf: der Nullpunkt der zeitlichen Orientierung wird da in jenen vergangenen Zeitmoment hineinversetzt, wo die darzustellende Szene anfängt, und verschiebt sich dann mit der Abwicklung der Geschehnisse stetig in dem zugehörigen Abschnitt des Zeitkontinuums bis zu dem "letzten" Moment dieser Szene. Die vergangenen Zeitphasen werden dadurch in eigentümlicher Weise - nacheinander - "gegenwärtig gemacht", als ob wir Leser Zeugen der betreffenden Ereignisse wären und "damals" genauer: in den "damaligen" "Jetzt" - lebten. Wird das Ganze im Modus der Gegenwart entworfen, dann haben wir eine besondere Darstellungsweise, die für die "dramatischen" Werke charakteristisch ist. (Ingarden [1931] 1973:256)

The neutralization of the episches Präteritum 59 [Only when a scene is shown in its concrete fullness and in its entire temporal extension are we again dealing with qualitatively determined, represented time. Or, to put it differently: only in the latter case are the concrete time phases represented, or exhibited, in their individuality. In the other cases, however, represented time is represented as time - or as a time phase in a determinate position in the time continuum - only in its general structure and not as a simple individual in its individuality. ... In the simple "informational" narrative mode, the time periods represented are always conceived from a "later, though in other respects indeterminate, time moment as something that is in the past. Thus, there clearly appears a characteristic temporal distance. On the other hand, time represented in its simple individuality, running its course phase by phase, can, in fact, also be conceived as being in the past. Nevertheless, it appears in characteristic proximity: the zero point of temporal orientation is transposed to that moment in the past where the represented scene begins and then, as the events develop, constantly shifts in each segment of the temporal continuum up to the "last" moment of this scene. Thus, in a characteristic manner, past-time phases are - successively - made "present", and we, the readers, seem to become witnesses to the given events and to live "then" or, more exactly, in the "erstwhile" "now". If the whole work is projected in the mode of the present, we are dealing with a manner of representation that is characteristic of "dramatic" works.] (Ingarden 1973: 241-42)

Following a trend of long standing in German Literaturwissenschaft, Ingarden's analysis is not only linguistic-sensitive, in the sense that narrative categories cannot be kept apart from grammatical forms, as it is also influenced by Dilthey's epistemic-based view of poetics and its focus on the standpoint-governed subject-object relationship. As can be seen, the structure of narrative transmission no longer depends on the tense form of the verb but on the particular rhythmic alternation between a slow, pace for pace, cadence and speeding up. The fact that the orientation of temporal perspective is conceived of as a function of the Erzählweisen has definitely paved the way for Stanzel's insistent claim that the epic preterite has or has not pasttime reference according as the narrative mode is summary or scene. Whether Ingarden's stand on (a)temporal reference will also have laid the foundations of Hamburger's theory of the atemporality of fiction is something to be (re)considered after Stanzel's narrative situations.

60

Linguistics in narratology: A critical historical survey

2.2.2. Stanzel's influential theory of narrative situations is based on the assumption that mediacy (Mittelbarkeit) of presentation is the distinctive feature of narrative fiction as opposed to poetry and drama. Mediacy, in turn, depends on a 3-tuple system of binary oppositions: person (the narrator belongs or does not belong to the fictional world), mode (adoption of a retrospective or synchronic temporal position) and perspective (inside view of characters or behavioural description). Permutation of the binary oppositions resulting from these categories then yields the different forms of mediacy called narrative situations, representable according to a chart slightly adapted from Cohn (1981): Table 3. Two-dimensional display of Stanzel's typology of narrative situations

Narrative Situation

PERSPECT. Centre of MODE PERSON (VOICE) (DISTANCE) (FOCALIZ.) Orientat.

A) Author. Third

Summary

Epic Preterite

Narrator

+PAST

B) Figural

Third

Scene

Internal

Character

-PAST

C) Neutral

Third

Scene

External

Observer

-PAST

D) Ich-ES.

First

Summary

Narrator-I

+PAST

ΡΛ L·;

First

Scene

Internal

Charact.-I

PA O